Communication method and apparatus

Abstract

A communication method and apparatus. The method comprises: a terminal device receiving first information, which is used for instructing the terminal device to execute a first task; and the terminal device executing the first task, wherein when a first model or function is in a first state, compared with the situation where the first model or function is in a second state, at least one of the following corresponding to the first task is different: a latency requirement, a usage time period of the first model or function, a memory cell occupancy time period of the first model or function, and a computing unit occupancy time period of the first model or function, and the first model or function corresponds to the first task. The requirement corresponding to a first task can match the state of a first model or function corresponding to the first task, such that a terminal device can smoothly execute the first task.

Description

Communication method and apparatus

[0001] This application claims priority to the Chinese Patent Application No. 202410408519.7, filed on April 3, 2024, entitled "A communication method and apparatus", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the field of communication technology, and in particular to a communication method and apparatus. BACKGROUND

[0003] Artificial intelligence (AI) has been widely applied to many application scenarios of air interface technology, for example: channel state information (CSI) feedback scenarios, CSI prediction scenarios, beam management scenarios, positioning scenarios, etc.

[0004] In various AI application scenarios, when a model is deployed on a terminal device side, a network device can not be able to know the state of the model in the terminal device, thereby causing a task to be unable to be smoothly executed. For example, a model 1 deployed in a terminal device is used to predict CSI, a network device instructs the terminal device to activate the model 1 and perform 1 time of CSI reporting, thereby the terminal device activates the model 1, and then uses the model 1 to perform CSI calculation and CSI reporting; after a period of time, the terminal device deactivates the model 1, however, the network device considers that the model 1 has been activated, and when instructing the terminal device to perform CSI reporting again, does not reserve time for the terminal device to activate the model 1, thereby the terminal device is unable to complete CSI calculation at a specified reporting time, and further unable to complete this time of CSI reporting. SUMMARY

[0005] Embodiments of the present application provide a communication method and apparatus, which are beneficial to matching of a demand corresponding to a first task with a state of a first model or function corresponding to the first task, thereby being beneficial to smooth execution of the first task by a terminal device.

[0006] In a first aspect, an embodiment of the present application provides a communication method, which can be executed by a terminal device. The terminal device can refer to the terminal device itself, or a processor, a module, a chip, or a chip system, etc. in the terminal device that implements the method. In the method, the terminal device receives first information, the first information being used to instruct the terminal device to execute a first task; and the terminal device executes the first task. Wherein, the first model or function is in a first state, which is different from the first model or function being in a second state, in at least one of the following corresponding to the first task: a time delay requirement, a use time period of the first model or function, a storage unit occupation time period of the first model or function, a computing unit occupation time period of the first model or function, and the first model or function corresponding to the first task.

[0007] It can be seen that, in the embodiment of the present application, the first model or function corresponding to the first task is in the first state, which is different from the first model or function being in the second state, in at least one of the following corresponding to the first task: the time delay requirement, the use time period of the first model or function, the storage unit occupation time period of the first model or function, the computing unit occupation time period of the first model or function. This way is beneficial for the terminal device to determine the state of the first model or function after receiving the first information, and to determine at least one of the following corresponding to the first task according to the state of the first model or function: the time delay requirement, the use time period of the first model or function, the storage unit occupation time period of the first model or function, the computing unit occupation time period of the first model or function, and then to execute the first task according to the determined at least one, which is beneficial for the requirement corresponding to the first task to match the state of the first model or function corresponding to the first task, and further beneficial for the terminal device to execute the first task smoothly.

[0008] Optionally, the "instruction" of the first information for instructing the terminal device to execute the first task can be understood as "triggering" or "configuring". For example, the first information is used to trigger the terminal device to execute the first task, and for another example, the first information is used to configure the terminal device to execute the first task. Optionally, the storage unit can be replaced by "storage resource", and the computing unit can be replaced by "computing resource". Therefore, the storage unit occupation time period of the first model or function can be replaced by "storage resource occupation time period of the first model or function", and the computing unit occupation time period of the first model or function can be replaced by "computing resource occupation time period of the first model or function".

[0009] Optionally, the first model or function corresponding to the first task can be one of the following: the terminal device executes the first task by using the first model or function, the first task is executed by using the first model or function, the first task is a task supported by the first function, the first task is one of one or more tasks supported by the first function, the first task is a task executed by using the first model or function, and the first task is one of one or more tasks executed by using the first model or function.

[0010] In an optional implementation, the first state of the second model or function is expected to last from the first time point to the second time point, the second model or function corresponds to a second task, and the start time of the second task is earlier than the start time of the first task. The second time point is later than the first time point. That is, during the time period from the first time point to the second time point, the second model or function corresponding to the second task is expected to be in the first state.

[0011] In an optional implementation, the first state of the second model or function is expected to last from the first time point to the second time point, the second model or function corresponds to a second task, and the start time of the second task is earlier than the start time of the first task. The second time point is later than the first time point. That is, during the time period from the first time point to the second time point, the second model or function corresponding to the second task is expected to be in the first state.

[0012] In an optional implementation, the first time point corresponds to one of the following: a start time of the first downlink signal, an end time of the first downlink signal, a start time of the first uplink signal, an end time of the first uplink signal, the first offset time point, and the second offset time point. The first downlink signal and the first uplink signal are both related to the second task.

[0013] In an optional implementation, the first offset time point and the second offset time point correspond to a first time offset and a second time offset, respectively. The first time offset corresponds to a time interval between the start time or the end time of the first downlink signal and the first offset time point. The second time offset corresponds to a time interval between the start time or the end time of the first uplink signal and the second offset time point. The first downlink signal can be a downlink signal or a downlink channel. The first uplink signal can be an uplink channel. The value of the first time offset is positive or negative. The value of the second time offset is positive or negative.

[0014] In an optional implementation, the first time offset and / or the second time offset are reported by the terminal device, preconfigured by a protocol, or configured or indicated by the network device.

[0015] In addition, the first time point corresponds to the start time of the first downlink signal, which can be that the first time point is the start time of the first downlink signal, or that the first time point is determined based on the start time of the first downlink signal. When the first time point is determined based on the start time of the first downlink signal, the first time point can be the start time or the end time of the time unit in which the start time of the first downlink signal is located, or the first time point can be the start time of the first time unit after the time unit in which the start time of the first downlink signal is located. Similarly, when the first time point corresponds to other time points, the same meaning applies, and thus is not repeated here.

[0016] Optionally, the first time point corresponds to one of the following: a start time point of a usage time period of the second model or function corresponding to the second task, a start time point of a storage unit occupation time period of the second model or function, a start time point of a computing unit occupation time period of the second model or function, an end time point of the usage time period of the second model or function corresponding to the second task, an end time point of the storage unit occupation time period of the second model or function, and an end time point of the computing unit occupation time period of the second model or function.

[0017] In an optional embodiment, the time interval between the first time point and the second time point is a first time length. Alternatively, the second time point can be determined based on the first time point and the first time length. The first time length is configured or indicated by the network device or preset by a protocol.

[0018] In the embodiment in which the time interval between the first time point and the second time point is the first time length, the second task belongs to a periodic task or a semi-persistent task. The second task belonging to the periodic task or the semi-persistent task can be understood as that the second task is a task in a certain period of a plurality of periods of the periodic task or the semi-persistent task to which the second task belongs.

[0019] Alternatively, the terminal device can determine the time characteristic of the second task, and when the terminal device determines that the second task belongs to the periodic task or the semi-persistent task, the second time point is determined based on the first time point and the first time length, and the second time point is the duration time point of the first state of the second model or function corresponding to the second task. That is, the first state of the second model or function can last until the second time point.

[0020] Optionally, in the embodiment in which the time interval between the first time point and the second time point is the first time length, the second task can also belong to a dynamically scheduled task. That is, when the terminal device determines that the second task belongs to the dynamically scheduled task, the second time point can also be determined based on the first time point and the first time length. The dynamically scheduled task can also be referred to as a non-periodic task.

[0021] In an optional embodiment, the first time point is determined based on a sixth time point and a sixth time length. The embodiments of the sixth time point and the sixth time length are similar to those of the first time point and the first time length, and are not described herein again.

[0022] In an optional implementation, when the second task belongs to a periodic task or a semi-persistent task, the first time point corresponds to the start time point of the first period in the N periods, and the second time point corresponds to the end time point of the Nth period in the N periods. The N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1. Alternatively, when the second task belongs to a periodic task or a semi-persistent task, the first state of the second model or function lasts from the start time point of the first period to the end time point of the Nth period in the N periods to which the second task belongs, that is, the second model or function is in the first state in the N periods to which the second task belongs.

[0023] In an optional implementation, when the second task belongs to a periodic task or a semi-persistent task, the first time point corresponds to the start time point of the first period in the N periods, and the second time point corresponds to the end time point of the Nth period in the N periods. The N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1. In the N periods to which the second task belongs, the second model or function periodically switches between the first state and the third state. In each period except the Nth period in the N periods, the second model or function switches from the third state to the first state at the end time point of the period, and switches from the first state to the third state at the start time point of the period. In particular, at the start time point of the first period in the N periods, the second model or function switches from the first state to the third state, or the second model or function switches from the second state to the third state. In particular, at the end time point of the Nth period in the N periods, the terminal device switches the third state of the second model or function to the second state, that is, after the end time point of the Nth period in the N periods, the second model or function is in the second state, or the second state of the second model or function starts from the end time point of the Nth period in the N periods. The third state is a state in which the model or function is in an activated state and is used, that is, the time period in which the second model or function is in the third state is a time period in which the terminal device performs model inference using the second model or function, or when the second model or function is in the third state, the terminal device is performing model inference using the second model or function.

[0024] In an optional implementation, when the second task belongs to a periodic task or a semi-persistent task, if the second task is not the last one in the N periods, the first time corresponds to the end time of the period in which the second task is located, and the second time corresponds to the start time of the next period of the period in which the second task is located; if the second task is the last one in the N periods, the first time corresponds to the end time of the period in which the second task is located, and the second time corresponds to the end time of the period in which the second task is located. Wherein, the N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1. Alternatively, when the second task belongs to a periodic task or a semi-persistent task, the first state of the second model or function is periodically from the end time of each period to the start time of the next period. That is, in the N periods to which the second task belongs, the second model or function is periodically in the first state.

[0025] Alternatively, the third state of the second model or function starts from the second time, or in other words, the second time is the time at which the terminal device switches the first state of the second model or function to the third state. The third state is a state in which the model or function is in an activated state and is used, or in other words, the time period in which the second model or function is in the third state is a time period in which the terminal device performs model inference using the second model or function, or in other words, when the second model or function is in the third state, the terminal device is performing model inference using the second model or function. When the second model or function is in the third state, tasks other than the second task cannot use the second model or function to perform model inference.

[0026] Alternatively, the first time is the time at which the terminal device switches the third state of the second model or function to the first state, or in other words, the second model or function is in the third state for a period of time before the first time.

[0027] Alternatively, the second model or function is in the second state after the end time of the Nth period in the N periods, or in other words, the end time of the Nth period in the N periods is the time at which the terminal device switches the third state of the second model or function to the second state, or in other words, the second state of the second model or function starts from the end time of the Nth period in the N periods.

[0028] In an optional implementation, when the time interval between the first time and the second time is the first time length, the second task belongs to a task that does not need to report information.

[0029] Or, the terminal device determines the second time based on the first time and the first time length when the terminal device determines that the second task does not need to report information. Wherein, the second task belongs to a task that does not need to report information, or the second task does not need to report information, which means that the result obtained by the terminal device executing the second task does not need to be directly reported through the uplink channel or the uplink signal. The result obtained by the terminal device executing the second task can be applied to other tasks, and the result obtained by the terminal device executing other tasks can be directly reported through the uplink channel or the uplink signal. Alternatively, the second task belongs to a task that does not need to report information, and the second task does not need to report information, which means that the network device does not configure the terminal device with an uplink channel or an uplink signal for the second task for result feedback.

[0030] In another optional implementation, the second time corresponds to one of: a starting time or an ending time of a channel carrying report information of the second task; a starting time or an ending time of a time unit in which the starting time or the ending time of the channel carrying the report information of the second task is located; a starting time of a first time unit after the time unit in which the starting time or the ending time of the channel carrying the report information of the second task is located.

[0031] Wherein, the second time corresponds to the starting time or the ending time of the channel carrying the report information of the second task, which can be: the second time is the starting time or the ending time of the channel carrying the report information of the second task, or can be: the second time is determined based on the starting time or the ending time of the channel carrying the report information of the second task. Similarly, the second time corresponds to other times, which have the same meaning and will not be repeated.

[0032] Optionally, the second time corresponds to one of: a starting time or an ending time of a channel carrying report information of the second task; a starting time or an ending time of a time unit in which the starting time or the ending time of the channel carrying the report information of the second task is located; a starting time of a first time unit after the time unit in which the starting time or the ending time of the channel carrying the report information of the second task is located, which is applicable to a task scenario in which the second task belongs to dynamic scheduling.

[0033] Or, the terminal device can determine the time characteristic of the second task, and when the terminal device determines that the second task belongs to dynamic scheduling, it is determined that the second time corresponds to one of: a starting time or an ending time of a channel carrying report information of the second task; a starting time or an ending time of a time unit in which the starting time or the ending time of the channel carrying the report information of the second task is located; a starting time of a first time unit after the time unit in which the starting time or the ending time of the channel carrying the report information of the second task is located.

[0034] In an optional implementation, when the second function or model corresponding to the second task is in the second state from the first time point to the second time point, the first model or function corresponding to the first task is in the first state, and the start time point of the first task is between the first time point and the second time point.

[0035] In other words, the terminal device can determine the state of the first model or function corresponding to the first task according to the first time point, the second time point, and the start time point of the first task; and when the start time point of the first task is between the first time point and the second time point, the terminal device determines that the first model or function is in the first state.

[0036] In an optional implementation, when the first model or function corresponding to the first task is in the first state, the first state of the first model or function lasts to a third time point, and the third time point corresponds to the first task and is later than the second time point. The third time point corresponding to the first task can be understood as that the third time point is a time point at which the first state of the first model or function corresponding to the first task lasts, that is, the first state of the first model or function can last to the third time point.

[0037] It can be seen that when the first model or function corresponding to the first task is in the first state and the third time point is later than the second time point, the first state of the first model or function can last to the third time point, that is, the first state of the first model or function corresponding to the first task lasts from the start time point of the first task to the third time point. In a possible case, the third time point is an end time point of the first model or function, or in other words, the third time point is a time point at which the terminal device switches the first state of the first model or function to a second state.

[0038] In an optional implementation, the third time point acts on at least one of the following corresponding to a third task: a time delay requirement, a use time period of a third model or function, a storage unit occupation time period of the third model or function, and a calculation unit occupation time period of the third model or function. The third model or function corresponds to the third task, and a start time point of the third task is later than the start time point of the first task.

[0039] In other words, the third time point can be used to determine at least one of the following corresponding to the third task: a time delay requirement, a use time period of a third model or function, a storage unit occupation time period of the third model or function, and a calculation unit occupation time period of the third model or function. Specifically, the third time point is used to determine a state of the third model or function corresponding to the third task, and the state of the third model or function is used to determine at least one of the following corresponding to the third task: a time delay requirement, a use time period of the third model or function, a storage unit occupation time period of the third model or function, and a calculation unit occupation time period of the third model or function.

[0040] In an optional implementation, when the first function or model corresponding to the first task lasts from the starting moment of the first task to the third moment, the third model or function corresponding to the third task is in the first state, including that the starting moment of the third task is between the starting moment of the first task and the third moment.

[0041] In other words, the terminal device can determine the state of the third model or function corresponding to the third task according to the starting moment of the first task, the third moment, and the starting moment of the third task; and when the starting moment of the third task is between the starting moment of the first task and the third moment, the terminal device determines that the third model or function is in the first state.

[0042] In an optional implementation, when the first model or function is in the first state, the first state of the first model or function lasts to the second moment, or in other words, the first state of the first model or function ends at the second moment, or in other words, the second moment is the moment when the terminal device switches the first state of the first model or function to the second state.

[0043] In an optional implementation, when the first state of the first model or function lasts to the second moment, the second state of the first model or function starts from the second moment, or the time period corresponding to the second state of the first model or function is included in a time outside the time period corresponding to the first state.

[0044] As can be seen, when the first state of the first model or function lasts to the second moment, the terminal device can complete the switching of the first model or function from the first state to the second state at the second moment, so that the second state of the first model or function starts from the second moment.

[0045] Optionally, when the first state of the first model or function lasts to the second moment, the terminal device completes the switching of the first model or function from the first state to the second state within a certain time period after the second moment, and then the time period corresponding to the second state of the first model or function is a time outside the time period corresponding to the first state.

[0046] In an optional implementation, when the first state of the second model or function lasts to the second moment, the first model or function is in the second state, including that the first model or function is not in the first state, or the starting moment of the first task is after the second moment.

[0047] As can be seen, when the first state of the second model or function lasts to the second moment, if the first model or function is not in the first state, the terminal device determines that the first model or function is in the second state. Alternatively, when the first state of the second model or function lasts to the second moment, if the starting moment of the first task is after the second moment, the terminal device determines that the first model or function is in the second state.

[0048] In an optional implementation, when the first model or function corresponding to the first task is in the first state, the first state of the first model or function lasts until the seventh moment, or in other words, the first state of the first model or function ends at the seventh moment, or in other words, the seventh moment is the moment when the terminal device switches the first state of the first model or function to the third state. The seventh moment corresponds to the first task. Optionally, the seventh moment corresponds to the starting moment of the first task. The seventh moment corresponding to the starting moment of the first task can be that the seventh moment is the starting moment of the first task, or can be that the seventh moment is determined based on the starting moment of the first task.

[0049] Optionally, after the seventh moment, the first state of the first model or function starts again at the eighth moment, or in other words, the eighth moment is the moment when the terminal device switches the third state of the first model or function to the first state. The eighth moment corresponds to the first task. Optionally, the eighth moment corresponds to the ending moment of the first task. The eighth moment corresponding to the ending moment of the first task can be that the eighth moment is the ending moment of the first task, or can be that the eighth moment is determined based on the ending moment of the first task.

[0050] Optionally, when the first model or function corresponding to the first task is in the first state, the first state of the first model or function starts at the ninth moment, or in other words, the ninth moment is the moment when the first model or function switches from the third state to the first state. Optionally, the ninth moment corresponds to the ending moment of the second task. The ninth moment corresponds to the ending moment of the second task, which can be that the ninth moment is the ending moment of the second task, or can be that the ninth moment is determined based on the ending moment of the second task.

[0051] Optionally, the third state of the first model or function lasts until the eighth moment, or in other words, the third state of the first model or function ends at the eighth moment, or in other words, the eighth moment is the moment when the terminal device switches the third state of the first model or function to the second state. The eighth moment corresponds to the first task, or the eighth moment is the second moment. Optionally, the eighth moment corresponds to the ending moment of the first task. The eighth moment corresponding to the ending moment of the first task can be that the eighth moment is the ending moment of the first task, or can be that the eighth moment is determined based on the ending moment of the first task.

[0052] In an optional implementation, the starting moment of the first task corresponds to one of the following: the starting moment of the second downlink signal, the ending moment of the second downlink signal, the third time offset, the starting moment of the usage time period of the first model or function, the starting moment of the storage unit occupation time period of the first model or function, and the starting moment of the calculation unit occupation time period of the first model or function. The time interval between the third time offset and the starting moment or the ending moment of the second downlink signal corresponds to the third time offset. The second downlink signal is related to the first task, and the second downlink signal can be a downlink signal or a downlink channel. The third time offset is a positive number or a negative number.

[0053] The starting moment of the first task corresponds to the starting moment of the second downlink signal, which can be that the starting moment of the first task is the starting moment of the second downlink signal, or that the starting moment of the first task is determined based on the starting moment of the second downlink signal. The starting moment of the first task is determined based on the starting moment of the second downlink signal, which can be that the starting moment of the first task is the starting moment or the ending moment of the time unit in which the starting moment of the second downlink signal is located, or that the starting moment of the first task is the starting moment of the first time unit after the time unit in which the starting moment of the second downlink signal is located. Similarly, the starting moment of the first task corresponds to other moments, and has similar meanings, which will not be repeated.

[0054] In an optional implementation, the third time offset is reported by the terminal device, or is preset, or is configured or indicated by the network device.

[0055] In an optional implementation, the second model or function is the same as the first model or function. The second model or function is the same as the first model or function, which can be that the second model or function is used to implement the same function as the first model or function, such as both are used to implement CSI prediction, or the physical model corresponding to the second model or function is the same as the first model or function. In addition, the structure and all parameters of the second model or function are the same as those of the first model or function, or the structure of the second model or function is the same as that of the first model or function and part of the parameters are different, such as the number of layers, the width, and the inter-layer connection relationship of the neural network, and the weight value and the bias of the neural network.

[0056] In an optional implementation, the latency requirement is one of the following: a time interval requirement between the second downlink signal and the second uplink signal; a time interval requirement between the second downlink signal and an end moment of a time period for using the first model or function; a time interval requirement between the second downlink signal and an end moment of a time period for occupying a storage unit of the first model or function; and a time interval requirement between the second downlink signal and an end moment of a time period for occupying a calculation unit of the first model or function. It can be understood that the time interval requirement between the second downlink signal and the second uplink signal means that the time interval between the second downlink signal and the second uplink signal needs to be greater than or equal to a time length, or in other words, the terminal device expects the time interval between the second downlink signal and the second uplink signal to be greater than or equal to a time length. When the time interval between the second downlink signal and the second uplink signal is less than the time length, the latency requirement is not met, and the terminal device can ignore the first task. Ignoring the task can also be understood as not executing the task, or not reporting the reporting information corresponding to the task, or not updating the reporting information corresponding to the task, or stopping executing the task. Similarly, not ignoring the task can also be understood as executing the task, or reporting the reporting information corresponding to the task, or updating the reporting information corresponding to the task, or continuing to execute the task. Optionally, "ignoring" can be replaced by "not retaining", and "not ignoring" can be replaced by "retaining". The meaning of the latency requirement in other requirements is similar.

[0057] In an optional implementation, when the first model or function is in the first state, the latency requirement corresponding to the first task is smaller than when the first model or function is in the second state. When the first model or function is in the first state, the latency requirement corresponding to the first task includes a time requirement for model inference, that is, a time required by the terminal device for model inference using the first model or function; and when the first model or function is in the second state, the latency requirement corresponding to the first task includes a time requirement for model activation and model inference, that is, a time required by the terminal device for activating the first model or function and a time required by the terminal device for inference using the first model or function. Therefore, when the first model or function is in the first state, the latency requirement corresponding to the first task is smaller than when the first model or function is in the second state. In this way, the terminal device can determine different latency requirements based on the state of the first model or function when executing the first task, and can successfully execute the first task.

[0058] In an optional implementation, when the first model or function is in the first state, the use time period of the first model or function corresponding to the first task starts from the start time or the end time of the second downlink signal corresponding to the first task; when the first model or function is in the second state, the start time of the use time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task. In this way, the terminal device can execute the first task based on the state of the first model or function, and the terminal device can successfully execute the first task.

[0059] In an optional implementation, when the first model or function is in the first state, the use time period of the first model or function corresponding to the first task starts from the start time or the end time of the second downlink signal corresponding to the first task; when the first model or function is in the second state, the start time of the use time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task. In this way, the terminal device can execute the first task based on the state of the first model or function, and the terminal device can successfully execute the first task.

[0060] In an optional implementation, when the first model or function corresponding to the first task is in the first state, the storage unit occupation time period of the first model or function corresponding to the first task starts from the start time or the end time of the second downlink signal corresponding to the first task, and the second downlink signal is a downlink signal or a downlink channel. When the first model or function corresponding to the first task is in the first state, the terminal device does not need to activate the first model or function, and thus the storage unit occupation time period of the first model or function corresponding to the first task can start from the start time or the end time of the second downlink signal corresponding to the first task.

[0061] In an optional implementation, when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second downlink signal corresponding to the first task, the second downlink signal being a downlink signal. When the terminal device successfully completes the first task in the resource scheduled by the network device, the first model or function is activated from the time earlier than the start time or the end time of the second downlink signal corresponding to the first task, so that the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second downlink signal corresponding to the first task.

[0062] In another optional implementation, when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task, the second downlink signal being a downlink channel.

[0063] In still another optional implementation, when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second uplink signal corresponding to the first task, and the start time of the storage unit occupation time period of the first model or function corresponding to the first task is determined by the start time or the end time of the second uplink signal corresponding to the first task, the second uplink signal being an uplink channel.

[0064] When the first model or function corresponding to the first task is in the second state, the terminal device successfully completes the first task before the start time or the end time of the second uplink signal scheduled by the network device, the first model or function is activated from the time earlier than the start time or the end time of the second uplink signal, so that the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second uplink signal corresponding to the first task. Further, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second uplink signal corresponding to the first task, and the time interval between the start time of the storage unit occupation time period of the first model or function corresponding to the first task and the start time or the end time of the second uplink signal corresponding to the first task needs to be greater than or equal to the sum of the model activation time and the time for which the first task uses the model for inference.

[0065] Optionally, when the first model or function corresponding to the first task is in the second state, the start time of the usage time period of the first model or function is earlier than the time when the first model or function is in the first state, and the start time of the computation unit occupation time period of the first model or function is earlier than the time when the first model or function is in the first state. In other words, the start time of the usage time period of the first model or function is earlier than the start time or the end time of the second uplink signal corresponding to the first task, and the start time of the computation unit occupation time period of the first model or function is earlier than the start time or the end time of the second uplink signal corresponding to the first task.

[0066] In an optional implementation, the first state is an active state, and the second state is a deactivated state.

[0067] In another optional implementation, the first state is an active and idle state, and the second state is a deactivated state. The first state is an active and idle state, which means that the model or function is in an active state and is not used. In this implementation, the state of the model or function further includes a third state.

[0068] In an optional implementation, the first time is configured by the network device or is preset.

[0069] In an optional implementation, the first time length is determined according to at least one of the following: an index or an identifier corresponding to the first time length; 1 / M times of a reporting period of the second task, where M is a positive integer; and a time characteristic of the second task. The time characteristic of the second task can be that the second task belongs to a periodic task, or the second task belongs to a semi-persistent task, or the second task belongs to a dynamically scheduled task.

[0070] In a second aspect, the embodiments of the present application further provide a communication method, which can be executed by a terminal device. The terminal device can refer to the terminal device itself, or a processor, a module, a chip, or a chip system, etc. in the terminal device that implement the method. In the method, the terminal device receives first information, and the first information is used to instruct the terminal device to execute a first task. The terminal device executes the first task. The computation unit occupation time period of a first model or function corresponding to the first task starts at a fourth time, and the fourth time is later than the start time of a second downlink signal corresponding to the first task, or the fourth time is earlier than the start time of the second downlink signal corresponding to the first task. The second downlink signal can be a downlink signal or a downlink channel.

[0071] It can be seen that, in the embodiment of the application, the start time of the time period occupied by the calculation unit of the first model or function corresponding to the first task is later than the start time of the second downlink signal corresponding to the first task, compared with the start time of the time period occupied by the calculation unit of the first model or function starting from the start time of the second downlink signal corresponding to the first task, the model or function calculation unit occupation time period can be saved, thereby facilitating the terminal device to perform other tasks using the redundant model or function calculation unit, and the resource utilization rate can be improved. Alternatively, the fourth time is earlier than the start time of the second downlink signal corresponding to the first task, compared with the start time of the time period occupied by the calculation unit of the first model or function starting from the start time of the second downlink signal corresponding to the first task, the terminal device can activate the first model or function using sufficient resources when the first model or function corresponding to the first task is in the second state, thereby facilitating the smooth execution of the first task.

[0072] In an optional implementation, the time interval between the fourth time and the start time of the second downlink signal is a second time length. The second time length is reported by the terminal device or preset.

[0073] In an optional implementation, the fourth time is earlier than the start time or end time of the second uplink signal corresponding to the first task, and the time interval between the fourth time and the start time or end time of the second uplink signal is a third time length. The third time length is reported by the terminal device or preset. The second uplink signal is an uplink channel.

[0074] In an optional implementation, the time interval between the fourth time and the start time or end time of the second uplink signal is a fourth time length. The fourth time length is reported by the terminal device or preset.

[0075] In an optional implementation, the time interval between the fourth time and the start time or end time of the second uplink signal is a fourth time length. The fourth time length is reported by the terminal device or preset.

[0076] In an optional implementation, the time interval between the fourth time and the start time or end time of the second uplink signal is a fourth time length. The fourth time length is reported by the terminal device or preset.

[0077] In a third aspect, the embodiments of the present application further provide a communication apparatus. The communication apparatus has some or all of the functions of the terminal device in the first aspect, or some or all of the functions of the terminal device in the second aspect. For example, the communication apparatus can have the functions of some or all of the terminal devices in the embodiments of the first aspect, or the functions of any one of the embodiments of the first aspect alone. The functions can be implemented by hardware, or by hardware running corresponding software. The hardware or software includes one or more units or modules corresponding to the above functions.

[0078] In a possible design, the communication apparatus can include a processing unit and a communication unit. The processing unit is configured to support the communication apparatus to perform the corresponding functions in the above method. The communication unit is configured to support the communication between the communication apparatus and another communication apparatus. The communication apparatus can further include a storage unit configured to be coupled to the processing unit and the communication unit, and store the necessary program instructions and data of the communication apparatus.

[0079] In an embodiment, the communication apparatus includes a processing unit and a communication unit, and the apparatus is applied to a terminal device.

[0080] The communication unit is configured to receive first information, where the first information is used to instruct the terminal device to perform a first task.

[0081] The processing unit is configured to perform the first task.

[0082] The first model or function is in a first state, and at least one of the following is different between the first state and a second state of the first model or function corresponding to the first task: a latency requirement, a time period of using the first model or function, a time period of occupying a storage unit of the first model or function, and a time period of occupying a computing unit of the first model or function.

[0083] In addition, in this aspect, other optional embodiments of the communication apparatus can refer to the related content of the first aspect, which will not be repeated here.

[0084] In another embodiment, the communication apparatus includes a processing unit and a communication unit, and the apparatus is applied to a terminal device.

[0085] The communication unit is configured to receive first information, where the first information is used to instruct the terminal device to perform a first task.

[0086] The processing unit is configured to perform the first task.

[0087] The first model or function corresponding to the first task occupies a time period of a fourth time point; the fourth time point is later than a starting time point of a second downlink signal corresponding to the first task, or the fourth time point is earlier than the starting time point of the second downlink signal corresponding to the first task.

[0088] In addition, in this aspect, other optional embodiments of the communication device can refer to the related content of the second aspect described above, which will not be described in detail here.

[0089] As an example, the communication unit can be a transceiver or a communication interface, the storage unit can be a memory, and the processing unit can be a processor.

[0090] In an embodiment, the communication device includes a processor and a transceiver, and the device is applied to a terminal device.

[0091] The transceiver is configured to receive first information, the first information being used to instruct the terminal device to perform a first task.

[0092] The processor is configured to perform the first task.

[0093] The first model or function in the first state is different from the first model or function in the second state in at least one of the following: a time delay requirement, a usage time period of the first model or function, a storage unit occupation time period of the first model or function, and a calculation unit occupation time period of the first model or function; the first model or function corresponds to the first task.

[0094] In addition, in this aspect, other optional embodiments of the communication device can refer to the related content of the first aspect described above, which will not be described in detail here.

[0095] In another embodiment, the communication device includes a processor and a transceiver, and the device is applied to a terminal device.

[0096] The transceiver is configured to receive first information, the first information being used to instruct the terminal device to perform a first task.

[0097] The processor is configured to perform the first task.

[0098] The first model or function corresponding to the first task occupies a time period of a fourth time point; the fourth time point is later than a starting time point of a second downlink signal corresponding to the first task, or the fourth time point is earlier than the starting time point of the second downlink signal corresponding to the first task.

[0099] In addition, in this aspect, other optional embodiments of the communication device can refer to the related content of the second aspect described above, which will not be described in detail here.

[0100] In another implementation, the communication device is a chip or a chip system. The processing unit can also be implemented as a processing circuit or a logic circuit; the communication unit can be an input / output interface, an interface circuit, an output circuit, an input circuit, a pin, or related circuit on the chip or chip system, etc.

[0101] In implementation, the processor can be configured to perform, for example but not limited to, baseband related processing, and the transceiver can be configured to perform, for example but not limited to, radio frequency transceiving. The above-mentioned devices can be respectively arranged on independent chips, or at least partially or entirely arranged on the same chip. For example, the processor can be further divided into an analog baseband processor and a digital baseband processor. The analog baseband processor can be integrated on the same chip as the transceiver, and the digital baseband processor can be arranged on an independent chip. With the continuous development of integrated circuit technology, more and more devices can be integrated on the same chip. For example, the digital baseband processor can be integrated on the same chip as various application processors (such as but not limited to a graphics processor, a multimedia processor, etc.). Such a chip can be referred to as a system on a chip (SoC). Whether to arrange the devices independently on different chips or to integrate them on one or more chips often depends on the needs of product design. The embodiments of the present application do not limit the implementation form of the above-mentioned devices.

[0102] In a fourth aspect, the embodiments of the present application further provide a processor for performing the above-mentioned various methods. In the process of performing these methods, the processes of transmitting and receiving the above-mentioned information in the above-mentioned methods can be understood as the processes of outputting the above-mentioned information by the processor and the processes of receiving the inputted above-mentioned information by the processor. When outputting the above-mentioned information, the processor outputs the above-mentioned information to the transceiver so as to be transmitted by the transceiver. After being outputted by the processor, the above-mentioned information can still need to be processed before reaching the transceiver. Similarly, when the processor receives the inputted above-mentioned information, the transceiver receives the above-mentioned information and inputs it to the processor. Furthermore, after being received by the transceiver, the above-mentioned information can still need to be processed before being inputted to the processor.

[0103] For the transmission and reception operations of the processor, if there is no special description, or if it does not contradict the actual role or inherent logic in the related description, it can be more generally understood as the output and input operations of the processor, rather than the transmission and reception operations directly performed by the radio frequency circuit and the antenna.

[0104] In implementations, the processor can be a processor specially configured to execute the methods, or can be a processor that executes computer instructions in a memory to execute the methods, such as a general-purpose processor. The memory can be a non-transitory memory, such as a read only memory (ROM), which can be integrated on the same chip as the processor, or can be separately arranged on different chips. The embodiments of the present application do not limit the type of the memory and the arrangement of the memory and the processor.

[0105] In a fifth aspect, the embodiments of the present application further provide a communication system, which includes a terminal device and a network device. In another possible design, the system can further include other devices / functions that interact with the terminal device and the network device.

[0106] In a sixth aspect, the embodiments of the present application provide a computer readable storage medium, which stores instructions, when the instructions are executed by a computer, implement the method in the first aspect or the second aspect.

[0107] In a seventh aspect, the embodiments of the present application further provide a computer program product including instructions, when the instructions are executed by a computer, implement the method in the first aspect or the second aspect.

[0108] In an eighth aspect, the embodiments of the present application provide a chip system, which includes a processor and an interface. The interface is configured to obtain a program or instructions. The processor is configured to invoke the program or instructions to implement or support the terminal device to implement the functions in the first aspect, or to implement or support the terminal device to implement the functions in the second aspect. For example, the processor is configured to determine or process at least one of the data and the information in the method. In a possible design, the chip system further includes a memory. The memory is configured to store necessary program instructions and data of the terminal. The chip system can be composed of a chip, or can include a chip and other discrete devices.

[0109] In a ninth aspect, the embodiments of the present application provide a communication apparatus, which includes a processor configured to execute computer programs or executable instructions stored in a memory. When the computer programs or executable instructions are executed, the apparatus performs the method in each possible implementation of the first aspect or the second aspect.

[0110] In a possible implementation, the processor and the memory are integrated together.

[0111] In another possible implementation, the memory is located outside the communication apparatus.

[0112] The beneficial effects of the third to ninth aspects can refer to the beneficial effects of the first or second aspects, which will not be repeated here. BRIEF DESCRIPTION OF DRAWINGS

[0113] FIG. 1 is a schematic diagram of an application framework;

[0114] FIG. 2 is a schematic diagram of another application framework;

[0115] FIG. 3 is a schematic diagram of a communication system;

[0116] FIG. 4 is a schematic diagram of another communication system;

[0117] FIG. 5 is a schematic diagram of CSI reporting;

[0118] FIG. 6 is a schematic diagram of a communication method provided by an embodiment of the present application;

[0119] FIG. 7 is a schematic diagram of periodic CSI reporting provided by an embodiment of the present application;

[0120] FIG. 8 is a schematic diagram of aperiodic CSI reporting provided by an embodiment of the present application;

[0121] FIG. 9 is a schematic diagram of another periodic CSI reporting provided by an embodiment of the present application;

[0122] FIG. 10 is a schematic diagram of yet another periodic CSI reporting provided by an embodiment of the present application;

[0123] FIG. 11a is a schematic diagram of another CSI reporting provided by an embodiment of the present application;

[0124] FIG. 11b is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0125] FIG. 11c is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0126] FIG. 11d is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0127] FIG. 11e is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0128] FIG. 12 is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0129] FIG. 13 is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0130] FIG. 14a is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0131] FIG. 14b is a schematic diagram of yet another CSI reporting provided by an embodiment of the present application;

[0132] FIG. 15a is a schematic diagram of another CSI reporting according to an embodiment of the present application;

[0133] FIG. 15b is a schematic diagram of another CSI reporting according to an embodiment of the present application;

[0134] FIG. 16 is a schematic diagram of another CSI reporting according to an embodiment of the present application;

[0135] FIG. 17 is a schematic diagram of interaction between a network device and a terminal device according to an embodiment of the present application;

[0136] FIG. 18 is a schematic diagram of interaction between a network device and a terminal device according to another embodiment of the present application;

[0137] FIG. 19 is a schematic diagram of another CSI reporting according to an embodiment of the present application;

[0138] FIG. 20 is a schematic diagram of another CSI reporting according to an embodiment of the present application;

[0139] FIG. 21 is a schematic diagram of a communication apparatus according to an embodiment of the present application;

[0140] FIG. 22 is a schematic diagram of another communication apparatus according to an embodiment of the present application. DETAILED DESCRIPTION

[0141] The technical solutions in the present application will be described below with reference to the accompanying drawings.

[0142] The technical solutions provided in the present application can be applied to various communication systems, for example: a 5th generation (5G) or new radio (NR) system, a long term evolution (LTE) system, an LTE frequency division duplex (FDD) system, an LTE time division duplex (TDD) system, a wireless local area network (WLAN) system, a satellite communication system, a future communication system such as a 6th generation (6G) mobile communication system, or a converged system of multiple systems, and the like. The technical solutions provided in the present application can also be applied to device to device (D2D) communication, vehicle-to-everything (V2X) communication, machine to machine (M2M) communication, machine type communication (MTC), and an internet of things (IoT) communication system or other communication systems.

[0143] A network element in a communication system can send a signal to another network element or receive a signal from another network element. The signal can include information, signaling, or data, etc. The network element can also be replaced by an entity, a network entity, a device, a communication device, a communication module, a node, a communication node, etc. The network element is taken as an example for description in the present disclosure. For example, the communication system can include at least one terminal device and at least one network device. The network device can send a downlink signal to the terminal device, and / or the terminal device can send an uplink signal to the network device. It can be understood that the terminal device in the present disclosure can be replaced by a first network element, and the network device can be replaced by a second network element, both of which perform the corresponding communication method in the present disclosure.

[0144] In the embodiments of the present application, the terminal device can also be referred to as a user equipment (UE), an access terminal, a user unit, a user station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a user terminal, a terminal, a wireless communication device, a user agent, or a user apparatus.

[0145] The terminal device can be a device providing voice / data, for example, a handheld device with wireless connection function, a vehicle-mounted device, etc. At present, some examples of terminals are: mobile phone, tablet computer, notebook computer, palm computer, mobile internet device (MID), wearable device, virtual reality (VR) device, augmented reality (AR) device, wireless terminal in industrial control, wireless terminal in self driving, wireless terminal in remote medical surgery, wireless terminal in smart grid, wireless terminal in transportation safety, wireless terminal in smart city, wireless terminal in smart home, cellular phone, cordless phone, session initiation protocol (SIP) phone, wireless local loop (WLL) station, personal digital assistant (PDA), handheld device with wireless communication function, computing device or other processing device connected to a wireless modem, wearable device, terminal device in a 5G network, or terminal device in a future evolved public land mobile network (PLMN), etc. The embodiments of the present application are not limited thereto.

[0146] By way of example and not limitation, in the embodiments of the present application, the terminal device can also be a wearable device. The wearable device can also be referred to as a wearable smart device, which is a general term for devices that are designed and developed by applying wearable technology to daily wear, such as glasses, gloves, watches, clothing, and shoes. The wearable device is a portable device that is directly worn on the body or integrated into the user's clothes or accessories. The wearable device is not only a hardware device, but also a device that realizes powerful functions through software support and data interaction and cloud interaction. The general wearable smart device includes devices with full functions, large size, and the ability to realize complete or partial functions without relying on a smart phone, such as smart watches or smart glasses, and devices that focus on a certain application function and need to be used in cooperation with other devices, such as smart phones, such as various smart wristbands and smart jewelry for monitoring vital signs.

[0147] In the embodiments of the present application, the apparatus for implementing the function of the terminal device can be a terminal device, or an apparatus capable of supporting the terminal device to implement the function, for example, a chip system, which can be installed in the terminal device or used in matching with the terminal device. In the embodiments of the present application, the chip system can be composed of a chip, or can include the chip and other discrete devices. In the embodiments of the present application, only the apparatus for implementing the function of the terminal device is taken as an example for description, and the scheme of the embodiments of the present application is not limited in this way.

[0148] The network device in the embodiments of the present application can be a device for communicating with a terminal device, and the network device can also be referred to as an access network device or a radio access network device, for example, the network device can be a base station. The network device in the embodiments of the present application can refer to a radio access network (RAN) node (or device) for accessing a terminal device to a wireless network. The base station can broadly cover various names in the following or be replaced by the following names, such as: Node B (NodeB), evolved Node B (eNB), next generation Node B (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), primary station, secondary station, motor slide retainer (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), radio unit (RU), positioning node, etc. The base station can be a macro base station, a micro base station, a relay node, a donor node or the like, or a combination thereof. The base station can also refer to a communication module, modem or chip used in the aforementioned device or apparatus. The base station can also be a mobile switching center and a device assuming a base station function in D2D, V2X, M2M communication, a network side device in 6G network, a device assuming a base station function in future communication system, etc. The base station can support networks of the same or different access technologies. Optionally, the RAN node can also be a server, a wearable device, a vehicle or a vehicle-mounted device, etc. For example, the access network device in vehicle to everything (V2X) technology can be a road side unit (RSU). The embodiments of the present application do not limit the specific technology and specific device form of the network device.

[0149] The base station can be fixed or mobile. For example, a helicopter or a drone can be configured to act as a mobile base station, and one or more cells can move according to the location of the mobile base station. In other examples, the helicopter or the drone can be configured to serve as a device communicating with another base station.

[0150] In some deployments, the network device mentioned by embodiments of the present application can be a device including a CU, or a DU, or a device including a CU and a DU, or a control plane CU node (central unit-control plane (CU-CP)) and a user plane CU node (central unit-user plane (CU-UP)), and a DU node. For example, the network device can include a gNB-CU-CP, a gNB-CU-UP, and a gNB-DU.

[0151] In some deployments, wireless access is assisted by multiple RAN nodes cooperating to assist a terminal, and different RAN nodes respectively implement part of the functions of a base station. For example, the RAN node can be a CU, a DU, a CU-CP, a CU-UP, or an RU, etc. The CU and the DU can be separately arranged, or can also be included in the same network element, such as a BBU. The RU can be included in a radio frequency device or a radio frequency unit, such as an RRU, an AAU, or an RRH.

[0152] The RAN node can support one or more types of fronthaul interfaces, different fronthaul interfaces respectively corresponding to DUs and RUs having different functions. If the fronthaul interface between the DU and the RU is a common public radio interface (CPRI), the DU is configured to implement one or more of baseband functions, and the RU is configured to implement one or more of radio frequency functions. If the fronthaul interface between the DU and the RU is another interface, which, relative to the CPRI, moves one or more of partial baseband functions of the downlink and / or uplink, such as, for the downlink, one or more of precoding, digital beamforming (BF), or inverse fast Fourier transform (IFFT) / add cyclic prefix (CP), from the DU to the RU for implementation, and for the uplink, one or more of digital beamforming (BF), or fast Fourier transform (FFT) / remove cyclic prefix (CP), from the DU to the RU for implementation. In a possible implementation, the interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the splitting manner between the DU and the RU is different, corresponding to different categories (Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, and F.

[0153] Taking eCPRI Cat A as an example, for downlink transmission, with layer mapping as the cut, the DU is configured to implement layer mapping and one or more functions (i.e., one or more of encoding, rate matching, scrambling, modulation, layer mapping) before layer mapping, while other functions (e.g., one or more of resource element (RE) mapping, digital beamforming (BF), or inverse fast Fourier transform (IFFT) / adding cyclic prefix (CP)) after layer mapping are implemented in the RU. For uplink transmission, with de-RE mapping as the cut, the DU is configured to implement de-mapping and one or more functions (i.e., one or more of decoding, de-rate matching, de-scrambling, de-modulation, inverse discrete Fourier transform (IDFT), channel equalization, de-RE mapping) before de-mapping, while other functions (e.g., one or more of digital BF or fast Fourier transform (FFT) / CP removal) after de-mapping are implemented in the RU. It can be understood that the function description of the DU and the RU corresponding to various types of eCPRI can refer to the eCPRI protocol, which is not described here.

[0154] In a possible design, the processing unit in the BBU for implementing baseband functions is referred to as a base band high (BBH) unit, and the processing unit in the RRU / AAU / RRH for implementing baseband functions is referred to as a base band low (BBL) unit.

[0155] In different systems, the CU (or CU-CP and CU-UP), DU or RU can also have different names, but those skilled in the art can understand their meanings. For example, in the ORAN system, the CU can also be referred to as an O-CU (open CU), the DU can also be referred to as an O-DU, the CU-CP can also be referred to as an O-CU-CP, the CU-UP can also be referred to as an O-CU-UP, and the RU can also be referred to as an O-RU. Any of the CU (or CU-CP, CU-UP), DU and RU in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0156] In the embodiments of the present application, the apparatus for implementing the function of the network device can be a network device, or can be an apparatus capable of supporting the network device to implement the function, such as a chip system, a hardware circuit, a software module, or a hardware circuit plus a software module. The apparatus can be installed in the network device or used in combination with the network device. In the embodiments of the present application, only the apparatus for implementing the function of the network device is taken as an example for description, and the scheme of the embodiments of the present application is not limited in this way.

[0157] The network device and / or the terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; can also be deployed on water surface; and can also be deployed on airplanes, balloons and satellites in the air. The scenarios in which the network device and the terminal device are located are not limited in the embodiments of the present application. In addition, the terminal device and the network device can be hardware devices, or software functions running on special hardware, general hardware, such as virtualized functions instantiated on a platform (for example, a cloud platform), or entities including special or general hardware devices and software functions. The specific forms of the terminal device and the network device are not limited in the present application.

[0158] In a wireless communication network, for example, in a mobile communication network, the services supported by the network are increasingly diverse, and therefore the needs to be met are increasingly diverse. For example, the network needs to be able to support ultra-high rate, ultra-low latency, and / or ultra-large connection. This feature makes network planning, network configuration, and / or resource scheduling increasingly complex. In addition, as the functions of the network become increasingly powerful, for example, supporting increasingly high spectrum, supporting high-order multiple input multiple output (MIMO) technology, supporting beamforming, and / or supporting new technologies such as beam management, network energy saving has become a hot research topic. These new needs, new scenarios and new features have brought unprecedented challenges to network planning, operation and efficient operation. In order to meet this challenge, artificial intelligence technology can be introduced into the wireless communication network, thereby realizing network intelligentization.

[0159] In order to support AI technology in the wireless network, an AI node can also be introduced into the network.

[0160] Optionally, the AI node can be deployed in one or more of the following positions in the communication system: an access network device, a terminal device, or a core network device, etc., or the AI node can also be deployed separately, for example, in a position other than any of the above devices, such as a host or a cloud server of an over the top (OTT) system. The AI node can communicate with other devices in the communication system, which can be one or more of the following: a network device, a terminal device, or a network element of a core network, etc.

[0161] It can be understood that the number of AI nodes is not limited in the present application. For example, when there are multiple AI nodes, the multiple AI nodes can be divided based on functions, such as different AI nodes being responsible for different functions.

[0162] It can also be understood that the AI nodes can be independent devices, can be integrated into the same device to implement different functions, or can be network elements in a hardware device, can be software functions running on a dedicated hardware, or can be virtualized functions instantiated on a platform (for example, a cloud platform), and the specific form of the AI nodes is not limited in the present application.

[0163] The AI node can be an AI network element or an AI module.

[0164] FIG. 1 is a schematic diagram of an application framework. As shown in FIG. 1, network elements in a communication system are connected through interfaces (for example, NG, Xn) or air interfaces. One or more AI modules (only one is shown in FIG. 1 for clarity) are arranged in one or more of the network element nodes, such as a core network device, an access network node or device (RAN node or device), a terminal, or an OAM. The access network node can be a separate RAN node or can include multiple RAN nodes, for example, including a CU and a DU. The CU and / or the DU can also be provided with one or more AI modules. Optionally, the CU can also be split into a CU-CP and a CU-UP. One or more AI models are arranged in the CU-CP and / or the CU-UP.

[0165] The AI module is used to implement a corresponding AI function. The AI modules deployed in different network elements can be the same or different. The AI module can implement different functions according to different parameter configurations of the model of the AI module. The model of the AI module can be configured based on one or more of the following parameters: a structural parameter (for example, at least one of the number of neural network layers, the width of the neural network, the connection relationship between layers, the weight of a neuron, the activation function of a neuron, or the bias in the activation function), an input parameter (for example, the type of the input parameter and / or the dimension of the input parameter), or an output parameter (for example, the type of the output parameter and / or the dimension of the output parameter). The bias in the activation function can also be referred to as the bias of the neural network.

[0166] One AI module can have one or more models. One model can infer an output including one parameter or multiple parameters. The learning process, the training process, or the inference process of different models can be deployed in different nodes or devices, or can be deployed in the same node or device.

[0167] Figure 2 is a schematic diagram of another application framework. As shown in Figure 2, a RAN intelligent controller (RIC) is included in the communication system. The RIC can be the AI module shown in Figure 1, for example, to implement AI-related functions. The RIC includes a near-real time RIC (near-RT RIC) and a non-real time RIC (Non-RT RIC). The non-real time RIC mainly processes non-real time information, such as data that is not sensitive to latency, which can be on the order of seconds. The real-time RIC mainly processes near-real time information, such as data that is relatively sensitive to latency, which can be on the order of tens of milliseconds.

[0168] The near-real time RIC is used for model training and inference. For example, to train an AI model, and to use the AI model for inference. The near-real time RIC can obtain network-side and / or terminal-side information from RAN nodes (such as CUs, CU-CPs, CU-UPs, DUs, and / or RUs) and / or terminals. This information can be used as training data or inference data. Optionally, the near-real time RIC can deliver inference results to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CUs and DUs, and / or between DUs and RUs. For example, the near-real time RIC delivers inference results to a DU, which then delivers them to an RU.

[0169] The non-real time RIC is also used for model training and inference. For example, to train an AI model, and to use the AI model for inference. The non-real time RIC can obtain network-side and / or terminal-side information from RAN nodes (such as CUs, CU-CPs, CU-UPs, DUs, and / or RUs) and / or terminals. This information can be used as training data or inference data, and inference results can be delivered to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CUs and DUs, and / or between DUs and RUs. For example, the non-real time RIC delivers inference results to a DU, which then delivers them to an RU.

[0170] The near-real time RIC and the non-real time RIC can also be separately provided as a network element. Alternatively, the near-real time RIC and the non-real time RIC can also be part of other devices. For example, the near-real time RIC can be provided in a RAN node (such as a CU or a DU), while the non-real time RIC can be provided in an OAM, a cloud server, a core network device, or another network device.

[0171] FIG. 3 is a schematic diagram of a communication system. As shown in FIG. 3, the communication system 100 can include at least one network device, such as the network device 110 shown in FIG. 3, and can include at least one terminal device, such as the terminal device 120 and the terminal device 130 shown in FIG. 3. The network device 110 and the terminal devices (such as the terminal device 120 and the terminal device 130) can communicate with each other through wireless links. The communication devices in the communication system, such as the network device 110 and the terminal device 120, can communicate with each other through multi-antenna technology.

[0172] FIG. 4 is a schematic diagram of another communication system. Compared with the communication system 100 shown in FIG. 3, the communication system 200 shown in FIG. 4 further includes an AI network element 140. The AI network element 140 is configured to perform AI-related operations, such as constructing a training data set or training an AI model.

[0173] In a possible implementation, the network device 110 can send data related to training of an AI model to the AI network element 140, and the AI network element 140 can construct a training data set and train the AI model. For example, the data related to training of the AI model can include data reported by the terminal device. The AI network element 140 can send a result of an AI model-related operation to the network device 110, and the network device 110 can forward the result to the terminal device. For example, the result of the AI model-related operation can include at least one of the following: a trained AI model, an evaluation result or a test result of the model, and the like. For example, part of the trained AI model can be deployed on the network device 110, and the other part can be deployed on the terminal device. Alternatively, the trained AI model can be deployed on the network device 110. Alternatively, the trained AI model can be deployed on the terminal device.

[0174] It should be understood that FIG. 4 is used as an example in which the AI network element 140 is directly connected to the network device 110, and in other scenarios, the AI network element 140 can also be connected to the terminal device. Alternatively, the AI network element 140 can be connected to both the network device 110 and the terminal device. Alternatively, the AI network element 140 can also be connected to the network device 110 through a third-party network element. The embodiments of the present application do not limit the connection relationship between the AI network element and other network elements.

[0175] The AI network element 140 can also be configured as a module in the network device and / or the terminal device, such as the network device 110 or the terminal device shown in FIG. 3.

[0176] It should be noted that FIG. 3 and FIG. 4 are merely simplified schematic diagrams for ease of understanding, and for example, other devices such as wireless relay devices and / or wireless backhaul devices can also be included in the communication system, which are not shown in FIG. 3 and FIG. 4. In actual application, the communication system can include multiple network devices and / or multiple terminal devices. The number of network devices and terminal devices included in the communication system is not limited by the embodiments of the present application.

[0177] Embodiments disclosed in the present application will present various aspects, embodiments or features of the present application around systems including a plurality of devices, components, modules, etc. It should be understood and appreciated that each system can include additional devices, components, modules, etc., and / or can not include all of the devices, components, modules, etc. discussed in conjunction with the accompanying drawings. In addition, combinations of these schemes can also be used.

[0178] In order to facilitate the understanding of the schemes of the embodiments of the present application, the terms that can be involved in the embodiments of the present application are explained below.

[0179] (1) AI, machine learning, model:

[0180] AI refers to the ability of a machine to have human intelligence, for example, the ability of a machine to apply computer hardware and software to simulate certain intelligent behaviors of humans. Machine learning (ML) is an important technical approach to realize AI, and machine learning can be divided into supervised learning, unsupervised learning and reinforcement learning.

[0181] In the machine learning method, the machine learns (or trains) the model by using the training data, and the model represents the mapping between the input and the output. The model can be used for inference (or prediction), that is, the output corresponding to a given input can be predicted by using the model. The output can also be referred to as an inference result (or a prediction result). The model can also be referred to as an AI model, an ML model, a rule, a function or other names. The AI model can be considered as a specific method to realize a certain AI function, and the model and the function represent the same meaning, and the model and the function can be replaced. The type of AI model can be a neural network, a linear regression model, a decision tree model, a support vector machine (SVM), a Bayesian network, a Q-learning model or other machine learning models.

[0182] (2) Activation state of the model or function, deactivation state of the model or function:

[0183] The active state of the model or function refers to at least one of the following: the terminal device supports using the model or function, the terminal device can execute or run the model or function, the terminal device supports executing or running the model or function, the terminal device has the capability of executing or running the model or function, the terminal device supports execution (or running or reasoning) of the model or function, the model or function in the terminal device is in a state of being able to be used or executed (or run or reasoned), and the like. The model or function in the terminal device is in the active state, for example, including that the terminal device has locally loaded a configuration file or a model file of the model or function, or the model or function in the terminal device can be immediately executed (or run or reasoned).

[0184] Correspondingly, the deactivation state of the model or function refers to at least one of the following: the terminal device does not support using the model or function, the terminal device cannot execute or run the model or function, the terminal device does not support executing or running the model or function, the terminal device does not have the capability of executing the model or function, the terminal device does not support execution (or running or reasoning) of the model or function, the model or function in the terminal device is not in a state of being able to be used or executed (or run or reasoned), the model or function in the terminal device is in a state of not being able to be used or executed (or run or reasoned), and the like. The model or function in the terminal device is in the deactivation state, for example, including that the terminal device has not locally loaded a configuration file or a model file of the model or function, or the model or function in the terminal device cannot be immediately executed (or run or reasoned), or the model or function in the terminal device needs a certain preparation time before being executed (or run or reasoned).

[0185] When the model or function is in the active state, the terminal device can switch the active state of the model or function to the deactivation state. Similarly, when the model or function is in the deactivation state, the terminal device can switch the deactivation state of the model or function to the active state.

[0186] When the terminal device switches the active state of a certain model or function to the deactivation state, there is a time delay or there is no time delay. If there is a time delay when the terminal device switches the active state of a certain model or function to the deactivation state, the terminal device can switch the active state of the model or function to the deactivation state after a period of time; if there is no time delay when the terminal device switches the active state of a certain model or function to the deactivation state, the terminal device can directly switch the active state of the model or function to the deactivation state at a certain time.

[0187] Similarly, when the terminal device switches the deactivated state of a certain model or function to the activated state, there is a time delay or there is no time delay. If there is a time delay when the terminal device switches the deactivated state of a certain model or function to the activated state, the terminal device can switch the activated state of the model or function to the activated state after a period of time; if there is no time delay when the terminal device switches the deactivated state of a certain model or function to the activated state, the terminal device can directly switch the deactivated state of the model or function to the activated state at a certain moment.

[0188] If there is a time delay in the process of switching the activated state of a certain model or function to the deactivated state by the terminal device, the terminal device does not process the activation or configuration related to the model or function in the time period corresponding to the time delay; if there is a time delay in the process of switching the deactivated state of a certain model or function to the activated state by the terminal device, the terminal device does not process the deactivation related to the model or function.

[0189] (3) Channel state information (CSI), CSI feedback:

[0190] CSI is a kind of channel information, which is an information capable of reflecting channel characteristics and channel quality. In a communication system (for example, an LTE communication system or an NR communication system), a network device needs to determine the configuration of a resource of a downlink data channel of a terminal device, a modulation and coding scheme (MCS), and precoding based on CSI.

[0191] In a time division duplex (TDD) system, since there is reciprocity between uplink and downlink channels, the network device can obtain accurate downlink CSI by measuring uplink CSI, for example, using uplink CSI as downlink CSI. In a frequency division duplex (FDD) system, the uplink and downlink reciprocity cannot be guaranteed, and the downlink CSI is obtained by the terminal device measuring downlink reference signals, such as channel state information reference signals (CSI RS) or synchronizing signal / physical broadcast channel blocks (SSB). Therefore, the terminal device needs to generate a CSI report according to the protocol predefinition or network device configuration, and feed back the downlink CSI to the network device through the CSI report, that is, the terminal device needs to feed back the downlink CSI to the network device.

[0192] In the NR, the configuration and reporting process of the downlink CSI includes: the network device sends a CSI reporting configuration (CSI-ReportConfig) to the terminal device, the CSI reporting configuration is used to specify the reporting type (reportConfigType), the reporting amount (reportQuantity), etc., wherein the reporting type can be periodic reporting, semi-persistent reporting or aperiodic reporting, and the reporting amount can be the rank indicator (RI), the channel quality indicator (CQI) and the precoding matrix indicator (PMI), the reference signal received power (RSRP), etc.; the network device sends the CSI-RS to the terminal device; the terminal device performs channel measurement and interference measurement according to the CSI-RS, and obtains the measurement result; the terminal device determines the configuration reporting amount according to the measurement result, and reports the downlink CSI to the network device, wherein the downlink CSI includes the RI, the CQI, the PMI and the RSRP measured by the terminal device.

[0193] Wherein, the RI is used to indicate the number of layers of the downlink transmission suggested by the terminal device, the CQI is used to indicate the modulation and coding scheme supported by the current channel condition judged by the terminal device, and the PMI is used to indicate the precoding suggested by the terminal device. The number of layers of the precoding indicated by the PMI corresponds to the RI.

[0194] It should be understood that the RI, the CQI and the PMI, etc. indicated by the above-mentioned CSI report are only the suggested values of the terminal device, and the network device can perform downlink transmission according to part or all of the information indicated by the CSI report. Alternatively, the network device can also not perform downlink transmission according to the information indicated by the CSI report.

[0195] In addition, if the reporting type in the CSI reporting configuration is periodic, the terminal device performs periodic reporting according to the period specified in the radio resource control (RRC) signaling, and the network device does not need to trigger the terminal device to report each time signaling is sent; if the reporting type in the CSI reporting configuration is semi-persistent, the initial reporting of the terminal device needs to be triggered by signaling, and once the terminal device is triggered, it performs periodic reporting according to the specified period; if the reporting type in the CSI reporting configuration is aperiodic, the network device needs to trigger reporting through downlink control information (DCI). In the semi-static CSI reporting trigger, when the CSI is carried on the physical uplink control channel (PUCCH) for reporting, the network device can trigger the terminal device to perform CSI reporting through media access control layer control element (MAC-CE) signaling; when the CSI is carried on the physical uplink shared channel (PUSCH) for reporting, the network device can trigger the terminal device to perform CSI reporting through DCI.

[0196] (4) Number of CSI processing units (CPU), CSI calculation time:

[0197] Number of CSI processing units available for terminal device to report CSI CPU , representing that the terminal device simultaneously supports N CPU CSI report calculations, N CPU is a positive integer. On a certain symbol, if the CSI report calculation occupies L CPUs, the terminal device has N CPU -L CPUs that are not occupied, and L is a positive integer less than N CPU . For a certain symbol, if N CPU -L CPUs are not occupied, if there are N CSI reports that need to occupy their respective CPUs from this symbol, and each CSI report corresponds to a number of CPUs , the terminal device does not need to update N-M CSI reports with the lowest priority, where 0≤M≤N, and M is a number that satisfies , N is a positive integer. That is, when the unoccupied CPU is not enough for the terminal device to process all the CSI reports, the terminal device can not process part of the CSI reports according to the priority.

[0198] , where the number of CPUs occupied by each CSI report processing is related to the configured reporting quantity and the number of reference signal resources used for channel measurement. For example, when the reporting quantity is RSRP, the number of CPUs occupied by each CSI report processing is When the reporting quantity is PMI, and the number of currently occupied CPUs is not 0, the number of CPUs occupied by each CSI report processing is , where K s is the number of CSI RS resources in the CSI RS resource set used for channel measurement. In addition, for the processing of each CSI report, the CPU will continuously occupy a number of symbols.

[0199] When the reporting type (reportConfigType) is not set to 'none', the number of CPU occupied symbols is determined according to the following rules: rule a, the time of CPU occupation for periodic CSI reporting or semi-persistent CSI reporting (excluding the initial semi-persistent CSI report on PUSCH after PDCCH triggered reporting) is: from the first symbol of the earliest reference signal resource used for channel measurement, to the last symbol of the PUSCH / PUCCH carrying the report. Wherein each latest measurement resource is not later than the corresponding CSI reference resource. Rule b, the time of CPU occupation for aperiodic CSI reporting is: from the first symbol after the PDCCH triggering the CSI report to the last symbol of the PUSCH carrying the CSI report. Rule c, the time of CPU occupation for the initial semi-persistent CSI report on PUSCH after PDCCH triggered CSI reporting is: from the first symbol after the PDCCH to the last symbol of the PUSCH carrying the CSI report.

[0200] In addition, when triggering CSI reporting, the network device needs to reserve sufficient time for the terminal device to perform CSI calculation, that is, to reserve CSI calculation time. For CSI reporting on PUSCH triggered by DCI, it is stipulated that the terminal device will report valid CSI reports only when the following two conditions are met:

[0201] Condition 1: The first uplink symbol carrying the corresponding CSI report (including the effect of timing advance) starts no earlier than symbol Z ref ;

[0202] Condition 2: The first uplink symbol carrying the nth CSI report (including the effect of timing advance) starts no earlier than symbol Z'ref (n), n is a positive integer.

[0203] wherein, Z ref is an uplink symbol, the interval between the start time of the cyclic prefix (CP) of the uplink symbol and the end time of the last symbol of the PDCCH triggering the CSI report is greater than or equal to T proc,CSI = (Z)(2048+144)·κ2 -μ ·T C , and is the earliest uplink symbol satisfying the condition. Z is a value obtained according to a table in a protocol; T C is a basic time unit in NR, T c = 1 / (Δf max ·N f ), wherein Δf max = 480·10 3 Hz, N f = 4096; κ = 64; μ = min(μ PDCCH , μ CSI-RS , μ UL ), μ PDCCH is the subcarrier spacing configuration of the PDCCH transmitting the DCI, μ CSI-RS is the minimum subcarrier spacing configuration of the CSI-RS, and μ UL is the subcarrier spacing configuration of the PUSCH transmitting the CSI report.

[0204] When aperiodic CSI RS is used for channel measurement of the n-th triggered CSI report, Z' ref (n) is defined as an uplink symbol, the interval between the start time of the CP of the uplink symbol and the end time of the last symbol of the latest ending resource in the resource used for measurement is greater than or equal to T' proc,CSI = (Z')(2048+144)·κ2 -μ ·T C , and is the earliest uplink symbol satisfying the condition. The values of Z and Z' are determined according to a table and principle given in a protocol.

[0205] Please refer to FIG. 5, which is a schematic diagram of CSI reporting. As shown in FIG. 5, the time interval between the first symbol of the PUSCH carrying the CSI report and the end time of the last symbol of the PDCCH triggering the CSI report is greater than or equal to a specified time parameter T proc,CSI , and the time interval between the first symbol of the PUSCH carrying the CSI report and the end time of all reference resources used for channel measurement is greater than or equal to a specified time parameter T' proc,CSI .

[0206] When the condition 1 and the condition 2 are not satisfied, the terminal device does not need to update the reported CSI. In addition, for non-DCI triggered reporting (i.e., periodic reporting and semi-persistent reporting), the network device can limit the CSI calculation time by defining a CSI reference resource to ensure that the terminal device needs to update the reported CSI only when there is sufficient calculation time. The CSI reference resource is defined as a block of time-frequency resources, which in the frequency domain, the CSI reference resource is defined by a set of downlink physical resource blocks corresponding to the frequency band related to the calculation of CSI; in the time domain, the CSI reference resource is defined as one valid time slot before the uplink time slot of CSI reporting, and the number of symbols between the time slot where the CSI reference resource is located and the CSI reporting time slot needs to be greater than a specified value. The CSI RS used to calculate the CSI report cannot be later than the CSI reference resource, and if there is no valid downlink time slot corresponding to a certain CSI reporting configuration, the terminal device can not report the CSI. That is, the time interval between the CSI RS used to calculate the CSI report and the CSI reporting time slot is greater than or equal to a specified time parameter.

[0207] (5) Air interface AI.

[0208] AI has been widely applied to many application scenarios of air interface technology, such as: CSI feedback scenarios, CSI prediction scenarios, beam management scenarios, positioning scenarios, etc. Illustratively, when AI is applied in the CSI feedback scenario, an autoencoder architecture can be used for CSI feedback, which generally includes an AI encoder and an AI decoder. Among them, the AI encoder can be deployed in the terminal device, and the AI decoder can be deployed in the network device. Compared with traditional CSI feedback technology, the CSI feedback based on the AI model can reduce the feedback overhead of the air interface and the calculation complexity of the terminal device under the same CSI feedback performance, and has greater application prospects. Illustratively, when AI is applied in the CSI prediction scenario, the terminal device or the network device can use a prediction model to predict the CSI at a future time using historical CSI and feed it back to the network device. The AI model can be located in the terminal device or in the network device. By accurately predicting the CSI at a future time, the problem of inaccurate CSI feedback information due to channel time variation can be solved. Illustratively, when AI is applied in the beam management scenario, the terminal device or the network device can use an AI model to efficiently and accurately identify the best beam. This AI model can be located only in the terminal device or only in the network device. Illustratively, when an AI model is applied in the positioning scenario, a three-point positioning method can be used for positioning. The terminal device obtains the position information of the surrounding three network devices and inputs it into the corresponding AI model. Then, the AI model obtains the position of the terminal device according to the distance, direction, and channel information between the terminal device and the three network devices.

[0209] The life cycle of an AI model in use involves the following aspects: data collection, model training (or model learning), model information release, model activation / deactivation, model inference (or model reasoning, reasoning, or prediction, etc.), model monitoring or model verification, model updating, or inference result release, etc. The life cycle management (LCM) of an air interface AI model can be based on model identification (model ID) or based on functionality. The model identification is an identification assigned in some way to identify the model. In the model ID-based LCM, the model ID is used to indicate operations on the model, such as activation / deactivation / selection / fallback / switching, etc. The functionality refers to an AI feature related to a configuration, or the functionality corresponds to a specific configuration under an AI feature, wherein the configuration is supported based on a terminal device capability indication. The AI feature refers to a feature that can use AI, for example: AI-based CSI feedback, AI-based beam management, etc. In the functionality-based LCM, the network device can indicate the terminal device to operate the AI functionality, such as activation / deactivation / selection / fallback / switching, etc., through 3rd generation partnership project (3GPP) signaling (for example: RRC, media access control control element (MAC-CE), DCI).

[0210] Specifically, one AI feature can include one or more functionalities. For one functionality, there can be one model or multiple models to implement the functionality. In addition, one model can also be used to implement multiple functionalities. For example, one functionality corresponds to one AI feature, for example, functionality 1 is AI-based time domain beam prediction, and functionality 2 is AI-based spatial domain beam prediction.

[0211] For example, one functionality corresponds to one AI feature + a specific set of RRC configurations, for example: functionality 1 is AI-based time domain beam prediction under configuration 1, and functionality 2 is AI-based time domain beam prediction under configuration 2, wherein configuration 1 and configuration 2 include at least one of the following: CSI RS resource configuration, CSI reporting configuration, beam set configuration, or prediction window configuration, etc.

[0212] For example, one functionality corresponds to one AI feature + a specific set of RRC configurations + a scenario / site identifier, e.g., functionality 1 is AI-based time-domain beam prediction under configuration 1 and scenario 1, and functionality 2 is AI-based time-domain beam prediction under configuration 1 and scenario 2, where configuration 1 includes at least one of the following: CSI RS resource configuration, CSI reporting configuration, beam set configuration, or prediction window configuration, and scenario 2 can be at least one of the following: urban, suburban, urban macrocell (UMa), urban microcell (UMi), indoor hotspot cell (InH), or highway, etc.

[0213] When model parallel running is supported, one model or functionality can have multiple processes working in parallel, i.e., for the same model or functionality, multiple processes can run in parallel, and each process performs inference independently. In this case, the "model or functionality" in this application can be replaced by "one process of the model or functionality".

[0214] When the model is deployed on the terminal device side, the network device can not know the specific model used and the model state. In the LCM based on functionality, the network device can instruct the terminal device to activate / deactivate the AI functionality through 3GPP signaling. However, there can be multiple models corresponding to one functionality, and multiple functionalities can also be implemented by one model. Therefore, the network device cannot accurately know the activation / deactivation state of the model for a certain functionality, which can cause the task to be unable to be smoothly performed.

[0215] In addition, even if the terminal device can normally perform the task, it can cause waste of resources. For example, if the model 1 is implemented in the terminal device corresponding to the functionality 1 and the functionality 2, the network device first activates the functionality 1 and instructs the terminal to perform CSI reporting for one time of beam prediction, and then the terminal device activates the model 1 to perform inference. After a period of time, the network device activates the functionality 2 and instructs the terminal device to perform CSI reporting for one time of beam prediction, and then the terminal device can directly use the model 1 to perform inference without activating the model 1, while the network device considers that the model corresponding to the functionality 2 needs to be activated for a period of time, and thus reserves the model activation time for the second CSI reporting, which can cause waste of CSI calculation resources.

[0216] It can be seen that if the network device and the terminal device are not aligned with the activation / deactivation state of a certain model, it may cause the terminal device to fail to successfully perform a task when using the model again, or cause resource waste during the execution of the task.

[0217] It should be understood that the above only takes the CSI reporting of the AI model for beam prediction as an example for description. In the CSI feedback, the AI model can also be used in other scenarios. For example, the AI model can be used for CSI prediction. The specific use of the AI model in the CSI feedback scenario is not limited in the embodiments of the present application.

[0218] It should be understood that in the present application, indication includes direct indication (also known as explicit indication) and implicit indication. Among them, the direct indication information A means including the information A; the implicit indication information A means indicating the information A through the corresponding relationship between the information A and the information B and the direct indication information B. Among them, the corresponding relationship between the information A and the information B can be pre-defined, pre-stored, pre-burned, or pre-configured.

[0219] It should be understood that in the present application, information C is used for the determination of information D, which includes that information D is determined based on information C only, and also includes that information D is determined based on information C and other information. In addition, information C used for the determination of information D can also include the case of indirect determination, such as the case that information D is determined based on information E, and information E is determined based on information C.

[0220] In addition, in the embodiments of the present application, "network element A sends information A to network element B" can be understood as that the destination of the information A or the intermediate network element in the transmission path between the destination is network element B, which can include direct or indirect sending of information to network element B. "Network element B receives information A from network element A" can be understood as that the source of the information A or the intermediate network element in the transmission path between the source is network element A, which can include direct or indirect receiving of information from network element A. The information can be processed as necessary between the source and the destination of the information transmission, such as format change, etc., but the destination can understand the valid information from the source. Similar expressions in the present application can be understood similarly, and will not be described here.

[0221] In the embodiments of the present application, the terminal device can be a terminal device with one or more models or functions. For example, the terminal device can be the terminal device shown in FIG. 1.

[0222] In an embodiment of the present application, the first task belongs to one of a periodic task, a semi-persistent task, and a dynamically scheduled task. The second task also belongs to one of a periodic task, a semi-persistent task, and a dynamically scheduled task. The third task also belongs to one of a periodic task, a semi-persistent task, and a dynamically scheduled task. The embodiments of the present application do not limit the time characteristics of the first task, the second task, and the third task. For example, the first task belongs to a dynamically scheduled task, the second task belongs to a periodic task, and the third task belongs to a dynamically scheduled task. For another example, the first task, the second task, and the third task all belong to periodic tasks.

[0223] In an embodiment of the present application, the first downlink signal and the first uplink signal are signals related to the second task. For example, the first downlink signal is a downlink signal for triggering the second task. For another example, the first uplink signal is an uplink signal carrying reported information corresponding to the second task. The first downlink signal can be a downlink signal or a downlink channel, and the first uplink signal is an uplink channel.

[0224] In an embodiment of the present application, the second downlink signal and the second uplink signal are signals related to the first task. For example, the second downlink signal is a downlink signal for triggering the first task. For another example, the second uplink signal is an uplink signal carrying reported information corresponding to the first task. The second downlink signal can be a downlink signal or a downlink channel, and the first uplink signal is an uplink channel.

[0225] The downlink signal can be a reference signal or a synchronization signal and physical broadcast channel block (SSB). The reference signal can be one of a CSI RS, a tracking reference signal (TRS), a phase-tracking reference signal (PTRS), and a positioning reference signal (PRS). The downlink channel can be a PDCCH, and the uplink channel can be a PUSCH or a PUCCH. For example, the first task is a dynamically scheduled CSI reporting task. The second downlink signal can be a PDCCH for triggering the terminal device to perform the CSI reporting task, or can be a reference signal or an SSB for measuring CSI. The second uplink signal is a PUSCH or a PUCCH carrying a CSI report.

[0226] In the embodiments of the present application, the time point further includes a time unit, which can be one of a second (s), a millisecond (ms), a microsecond (us), a time slot (slot), a symbol (symbol), or at least one continuous symbol. The embodiments of the present application do not limit the specific manner of the time unit.

[0227] The embodiments of the present application provide a communication method, and FIG. 6 is a schematic diagram of the communication method. The communication method is described from the perspective of interaction between a network device and a terminal device, and the communication method includes but is not limited to the following steps:

[0228] S601. The network device sends first information to the terminal device, and the first information is used to instruct the terminal device to perform a first task, wherein the first model or function is in a first state, and at least one of the following items corresponding to the first task is different between the first model or function being in the first state and the first model or function being in a second state: a latency requirement, a usage time period of the first model or function, a storage unit occupation time period of the first model or function, a computing unit occupation time period of the first model or function, and the first model or function corresponding to the first task. Correspondingly, the terminal device receives the first information from the network device.

[0229] Optionally, the "instruction" in the first information used to instruct the terminal device to perform the first task can be understood as "triggering" or "configuration", for example, the first information is used to trigger the terminal device to perform the first task, and for another example, the first information is used to configure the terminal device to perform the first task.

[0230] In the embodiments of the present application, the first task is a task performed by running the first model or function, or in other words, the first task needs to use the first model or function for model inference. When the first model or function is in the first state (activated state), the model inference can be directly performed, and when the first model or function is in the second state (deactivated state), the model or function needs to be activated first, so that the model or function is switched to the first state (activated state), and then the model inference is performed. Therefore, the first information used to instruct the terminal device to perform the first task can be understood as implicitly instructing the terminal device to perform the activation and inference of the first model or function corresponding to the first task. That is, the network device can instruct the activation of the first model or function through the first information, and the terminal device determines whether to perform the actual model activation operation according to the state of the first model or function.

[0231] When the first task belongs to a periodic task, the first information can be carried in RRC signaling; when the first task belongs to a semi-persistent task, the first information can be carried in MAC-CE or PDCCH; when the first task belongs to a dynamically scheduled task, i.e., a non-periodic task, the first information can be carried in DCI, which is carried in PDCCH. Wherein, the first task belongs to a periodic task, indicating that the terminal device needs to periodically perform the first task. The first task belongs to a semi-persistent task, indicating that when the terminal device is triggered to perform the first task, the first task needs to be periodically performed. The first task belongs to a dynamically scheduled task, indicating that when the terminal device is triggered to perform the first task, the first task is performed once.

[0232] For example, when the first task belongs to a periodic CSI reporting task, the first information can be CSI reporting configuration information, which is carried in RRC signaling, and is used to configure the period and reporting amount of the terminal device for periodically reporting CSI. For another example, when the first task belongs to a semi-persistent CSI reporting task, the first information can be CSI reporting trigger information, which is carried in MAC-CE, and is also used to configure the period and reporting amount of the terminal device for reporting CSI. For another example, when the first task belongs to a dynamically scheduled CSI reporting task, the first information can also be CSI reporting trigger information, which is carried in DCI, which is carried in PDCCH, and is used to configure the resource and reporting amount of the terminal device for reporting CSI.

[0233] In addition, the first model or function corresponding to the first task can be one of the following: the model or function used by the terminal device to perform the first task is the first model or function, the terminal device uses the first model or function to perform the first task, the first task is performed by the first model or function, the first task is a task supported by the first function, the first task is one of one or more tasks supported by the first function, the first task is a task performed by the first model or function, and the first task is one of one or more tasks performed by the first model or function. For example, the first task is a CSI reporting task, and the first model or function corresponding to the first task is model or function 1, and the model used by the terminal device to perform model inference to obtain CSI is model or function 1. Similarly, the second model or function corresponds to the second task, and the third model or function corresponds to the third task, which has a similar understanding and will not be repeated.

[0234] Optionally, the first state is an active state, and the second state is a deactivated state. The active state of the model or function and the deactivated state of the model or function can be referred to the above description, and will not be repeated.

[0235] Optionally, the first state is an active and idle state, and the second state is a deactivated state. The first state being an active and idle state indicates that the model or function is in an active state and is not used. In this way, the state of the model or function further includes a third state, which is an active and occupied state, indicating that the model or function is in an active state and is occupied / used, or in other words, the third state is the state of the model or function when it is used for model inference.

[0236] Optionally, the first model or function being in the first state can be replaced by: the first task being in the first state, and the first task being in the first state indicating that the first model or function used by the terminal device to execute the first task is in the first state. Correspondingly, the first model or function being in the second state can be replaced by: the first task being in the second state, and the first task being in the second state indicating that the first model or function used by the terminal device to execute the first task is in the second state.

[0237] In an optional implementation, the latency requirement is a time interval requirement between a second downlink signal and a second uplink signal, and both the second downlink signal and the second uplink signal are related to the first task. The second downlink signal being a downlink channel for triggering the terminal device to execute the second task, or being a downlink signal used by the terminal device to execute the second task, and the second uplink signal being an uplink channel carrying report information corresponding to the second task, indicate that the second downlink signal and the second uplink signal are both related to the first task. For example, the first task is a CSI reporting task, the second downlink signal is RRC signaling or PDCCH or MAC-CE for triggering CSI reporting, or the second downlink signal is a reference signal or SSB used for measuring CSI in the CSI reporting process, and the second uplink signal is PUSCH or PUCCH carrying the CSI report.

[0238] For example, the first task is a CSI reporting task, the second downlink signal is PDCCH for triggering CSI reporting, and the second uplink signal is PUSCH carrying the CSI report. In this case, the latency requirement is a time interval requirement between the PDCCH and the PUSCH. For example, the latency requirement is that the start of the first uplink symbol carrying the corresponding CSI report (including the effect of timing advance) is not earlier than symbol Z ref , or the start of the first uplink symbol carrying the nth CSI report (including the effect of timing advance) is not earlier than symbol Z' ref (n), n being a positive integer. Wherein Z ref is an uplink symbol, and the interval between the start time of the CP of Z ref(n) is defined as an uplink symbol, the interval between the start time of the CP and the end time of the last symbol of the last resource in the resource used for measurement is greater than or equal to a specified time length.

[0239] For example, the first task is a CSI reporting task, the second downlink signal is a CSI RS used for measuring CSI, and the second uplink signal is a PUSCH carrying a CSI report. In this case, the latency requirement is the time interval requirement between the CSI RS and the PUSCH, or the time interval requirement between the reference resource of the CSI and the PUSCH.

[0240] In another optional implementation, the latency requirement is the time interval requirement between the second downlink signal and the end time of the use time period of the first model or function.

[0241] The use time period of the first model or function refers to the time period occupied by the terminal device in using the first model or function for model inference. For example, the first task is a CSI reporting task, and the second downlink signal is a CSI RS used for measuring CSI. In this case, the latency requirement is the time interval requirement between the CSI RS and the end time of the time period occupied by the terminal device in using the first model or function for CSI prediction.

[0242] In another optional implementation, the latency requirement is the time interval requirement between the second downlink signal and the end time of the storage unit occupation time period of the first model or function.

[0243] Optionally, the first model or function occupies the storage unit when it is in the first state, occupies the storage unit during the time period when the first model or function switches from the first state to the second state, occupies the storage unit during the time period when the first model or function switches from the second state to the first state, and does not occupy the storage unit when it is in the second state.

[0244] Therefore, the latency requirement between the second downlink signal and the end time of the storage unit occupation time period of the first model or function can be the time interval between the second downlink signal and the start time of the first model or function in the second state.

[0245] In another optional implementation, the latency requirement is the time interval requirement between the second downlink signal and the end time of the calculation unit occupation time period of the first model or function.

[0246] Optionally, the first model or function occupies the calculation unit when it is in the first state. Optionally, the first model or function occupies the calculation unit when it is used for model inference.

[0247] Optionally, the first model or function does not occupy the computing unit when in the second state, the first model or function does not occupy the computing unit during the time period from switching from the first state to the second state, and the first model or function does not occupy the computing unit during the time period from switching from the second state to the first state.

[0248] Optionally, at least one of the following is different when the first model or function in the first state corresponds to the first task, compared to when the first model or function in the second state corresponds to the first task: latency requirement, time period of use of the first model or function, time period of storage unit occupation of the first model or function, time period of computing unit occupation of the first model or function. Alternatively, when the first model or function in the first state corresponds to the first task, at least one of the following is different compared to when the first model or function in the second state corresponds to the first task: latency requirement, determination manner of time period of use of the first model or function, determination manner of time period of storage unit occupation of the first model or function, determination manner of time period of computing unit occupation of the first model or function. Alternatively, the determination manner of at least one of the following is different when the first model or function in the first state corresponds to the first task, compared to when the first model or function in the second state corresponds to the first task: latency requirement, time period of use of the first model or function, time period of storage unit occupation of the first model or function, time period of computing unit occupation of the first model or function.

[0249] Optionally, the storage unit can be replaced by "storage resource", and the computing unit can be replaced by "computing resource". For example, the time period of storage unit occupation of the first model or function can be replaced by "time period of storage resource occupation of the first model or function", and the time period of computing unit occupation of the first model or function can be replaced by "time period of computing resource occupation of the first model or function". Similarly, the time period of storage unit occupation of other models or functions can be replaced by "time period of storage resource occupation", and the time period of computing unit occupation of other models or functions can be replaced by "time period of computing resource occupation", which will not be described herein.

[0250] It can be seen that the first function or model used by the terminal device to perform the first task is in the first state and in the second state, and the following at least one of the first task is different: latency requirement, use time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function. Therefore, after the terminal device receives the first information, the state of the first model or function is determined, and then according to the state of the first model or function, the following at least one corresponding to the first task is determined: latency requirement, use time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function, and then the first task is executed according to the determined at least one. Wherein, the following at least one corresponding to the first task: latency requirement, use time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function can be regarded as the requirement corresponding to the first task, and this way is beneficial to match the requirement corresponding to the first task with the state of the first model or function, and then is beneficial to the terminal device to smoothly execute the first task, or in other words, is beneficial to the model inference corresponding to the first model or function to be smoothly performed.

[0251] S602. The terminal device performs the first task.

[0252] When the first model or function corresponding to the first task is in different states, the following at least one of the first task is not the same: latency requirement, use time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function. Then, the terminal device needs to execute the first task in combination with the state of the first model or function corresponding to the first task, so as to match the requirement corresponding to the first task with the state of the first model or function, and then to guarantee the smooth execution of the first task, or in other words, to guarantee the model inference corresponding to the first model or function to be smoothly performed.

[0253] In an optional implementation, the terminal device performs the first task, including: determining the state of the first model or function; according to the state of the first model or function, determining the following at least one corresponding to the first task: latency requirement, use time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function; and executing the first task according to the determined at least one. This way can make the determined at least one of the first task match the state of the first model or function, so as to guarantee the terminal device to smoothly execute the first task, or in other words, to guarantee the model inference of the first model or function to be smoothly performed.

[0254] In an optional implementation, the terminal device determines the state of the first model or function, including: determining a second time corresponding to the second model or function; and determining the state of the first model or function according to the second time and a starting time of the first task.

[0255] The second model or function corresponds to the second task. That is, the model or function used by the terminal device to perform the second task is the second model or function. The starting time of the second task is earlier than the starting time of the first task, or the triggering time / starting time of the second task is earlier than the triggering time or starting time of the first task.

[0256] The starting time of the second task corresponds to one of the following: the starting time of the first downlink signal, the ending time of the first downlink signal, a fourth offset time, the starting time of the usage time period of the second model or function, the starting time of the storage unit occupation time period of the second model or function, and the starting time of the calculation unit occupation time period of the second model or function. The time interval between the fourth offset time and the starting time or ending time of the first downlink signal corresponds to a seventh time offset. The first downlink signal is related to the second task. Specifically, the first downlink signal can be a channel, such as a PDCCH or DCI, for triggering the terminal device to perform the second task, or can be a signal, such as one of CSI RS, TRS, PTRS, PRS, and SSB, used by the terminal device to perform the second task.

[0257] The starting time of the second task corresponds to the starting time of the first downlink signal, which can be that the starting time of the second task is the starting time of the first downlink signal, or can be that the starting time of the second task is determined according to the starting time of the first downlink signal. The starting time of the second task is determined based on the starting time of the first downlink signal, which can be that the starting time of the second task is the starting time or ending time of the time unit in which the starting time of the first downlink signal is located, or can be that the starting time of the second task is the starting time of the first time unit after the time unit in which the starting time of the first downlink signal is located. Similarly, the starting time of the second task corresponds to other time points, which has a similar understanding and will not be repeated.

[0258] The second model or function corresponding to the second task is the same as the first model or function corresponding to the first task. The second model or function being the same as the first model or function can be understood as: the second model or function and the first model or function are used to implement the same function, such as both being used for CSI prediction or beam prediction, or the physical model corresponding to the second model or function and the first model or function is the same. The second model or function and the first model or function are the same in structure and all parameters, or the second model or function and the first model or function are the same in structure and part of the parameters are not the same, for example, the number of layers, the width, and the inter-layer connection relationship of the neural network, and the weight value and the bias of the neural network. For example, the first model or function and the second model or function are both used for CSI prediction, the structure of the first model or function and the second model or function is the same, and part of the weight values are not the same.

[0259] Optionally, the second time corresponding to the second model or function can be understood as: the second time is the duration time of the first state of the second model or function. That is, the first state of the second model or function can last until the second time, or after the second time, the second model or function can not be in the first state.

[0260] It can be seen that the terminal device can determine the state of the first model or function based on the duration time of the first state of the second model or function and the starting time of the first task. Or, the duration time of the first state of the second model or function can be used to determine the state of the first model or function.

[0261] In an optional implementation, the terminal device determines the second time corresponding to the second model or function, including: determining the second time according to the first time and the first time length. The second time is later than the first time.

[0262] Optionally, the first time can also be referred to as timing start, and the first time length can also be referred to as timer.

[0263] In an optional implementation, the first time corresponds to one of: the starting time of the first downlink signal, the ending time of the first downlink signal, the starting time of the first uplink signal, the ending time of the first uplink signal, the first offset time, and the second offset time. The time interval between the first offset time and the starting time or the ending time of the first downlink signal corresponds to the first time offset, and the time interval between the second offset time and the starting time or the ending time of the first uplink signal corresponds to the second time offset.

[0264] The first downlink signal and the first uplink signal are both related to the second task. The first uplink signal can be a channel carrying the reported information when the terminal device performs the first task, such as PUCCH or PUSCH.

[0265] The first time corresponds to the starting time of the first downlink signal. The first time can be the starting time of the first downlink signal, or the first time is determined based on the starting time of the first downlink signal, for example, the first time is determined based on the starting time of the first downlink signal and the first time offset, or the first time is the starting time or the ending time of the time unit in which the starting time of the first downlink signal is located, or the first time is the starting time of the first time unit after the time unit in which the starting time of the first downlink signal is located. Similarly, the first time corresponds to other times, and has the same meaning, which will not be repeated.

[0266] For example, the first task belongs to a periodic CSI reporting task, and FIG. 7 is a schematic diagram of periodic CSI reporting. As shown in FIG. 7, the first time is the starting time of the CSI RS for measuring the CSI. For another example, the first task belongs to a dynamically scheduled CSI reporting task, and FIG. 8 is a schematic diagram of aperiodic CSI reporting. As shown in FIG. 8, the network device triggers the terminal device to perform aperiodic CSI reporting through DCI at t1, and the first time is the starting time or the ending time of the DCI. For another example, the first task belongs to a periodic CSI reporting task, and FIG. 9 is another schematic diagram of periodic CSI reporting. As shown in FIG. 9, the first time is the starting time of the PUSCH carrying the CSI report.

[0267] In addition, the first time offset and / or the second time offset can be reported by the terminal device, can be preset, or can be configured or indicated by the network device. The value of the first time offset and the value of the second time offset can be positive or negative.

[0268] When the value of the first time offset is positive, the time interval between the first offset time and the starting time or the ending time of the first downlink signal corresponds to the first time offset, which means that the first offset time is the time when the starting time or the ending time of the first downlink signal is delayed by the first time offset. For example, FIG. 10 is another schematic diagram of periodic CSI reporting. As shown in FIG. 10, the first time is the first offset time t3, t3 is the time when the starting time t2 of the CSI RS for measuring the CSI is delayed by the first time offset, that is, the first time offset is positive. When the value of the first time offset is negative, the time interval between the first offset time and the starting time or the ending time of the first downlink signal corresponds to the first time offset, which means that the first offset time is the time corresponding to the starting time or the ending time of the first downlink signal after the starting time or the ending time of the first downlink signal is advanced by the absolute value of the first time offset.

[0269] Similarly, when the value of the second offset is positive, the time interval between the second offset time and the starting time or the ending time of the first uplink signal corresponds to the second time offset, indicating that the second offset time is the time when the starting time or the ending time of the first uplink signal is delayed by the second time offset; when the value of the second offset is negative, the time interval between the second offset time and the starting time or the ending time of the first uplink signal corresponds to the second time offset, indicating that the second offset time is the time corresponding to the time when the starting time or the ending time of the first uplink signal is advanced by the absolute value of the second time offset. For example, FIG. 11a is another CSI reporting schematic diagram, as shown in FIG. 11a, the first time is the second offset time t3, t3 is the time when the starting time of the PUSCH carrying the CSI report is delayed by the second time offset, that is, the second time offset is positive. For another example, FIG. 11b is another CSI reporting schematic diagram, as shown in FIG. 11b, the first time is the second offset time t2, t2 is the time when the starting time of the PUSCH carrying the CSI report is advanced by the second time offset, that is, the second time offset is negative.

[0270] Optionally, one or more of the first time offset between the starting time of the first downlink signal and the first offset time, the first time offset between the ending time of the first downlink signal and the first offset time, the second time offset between the starting time of the first uplink signal and the second offset time, and the second time offset between the ending time of the first uplink signal and the first offset time can be equal or not equal, and the embodiments of the present application do not limit this. For example, the first time offset between the starting time of the first downlink signal and the first offset time is equal to or not equal to the first time offset between the ending time of the first downlink signal and the first offset time. For another example, the first time offset between the starting time of the first downlink signal and the first offset time is equal to or not equal to the second time offset between the ending time of the first uplink signal and the second offset time.

[0271] In another optional embodiment, the first time corresponds to one of the following: the starting time of the use time period of the second model or function, the starting time of the storage unit occupation time period of the second model or function, and the starting time of the calculation unit occupation time period of the second model or function.

[0272] For example, the second model or function is the model or function used for the periodic CSI reporting in FIG. 7, the usage time period of the model or function is the time period of the model inference shown in FIG. 7, the first time point in FIG. 7 can not be t2, but can be the starting time point t4 of the time period occupied by the model inference in FIG. 7. For another example, the second model or function is the model or function used for the periodic CSI reporting in FIG. 7, the starting time point of the storage unit occupancy time period of the model or function is the starting time point t5 of the model activation in FIG. 7, the first time point in FIG. 7 can not be t2, but can be the starting time point t5 of the model activation in FIG. 7. For another example, the second model or function is the model or function used for the periodic CSI reporting in FIG. 7, the computation unit occupancy time period of the model or function is also the time period of the model inference shown in FIG. 7, the first time point in FIG. 7 can not be t2, but can be the starting time point t4 of the time period occupied by the model inference in FIG. 7.

[0273] In yet another optional implementation, the first time point corresponds to one of the following: the ending time point of the usage time period of the second model or function, the ending time point of the storage unit occupancy time period of the second model or function, and the ending time point of the computation unit occupancy time period of the second model or function. In this case, the first state is the activated and idle state, and before the first time point, the second model or function is in the third state. The third state is the activated and occupied state, i.e., the third state is the state in which the model or function is in model inference.

[0274] For example, FIG. 11c is another diagram of CSI reporting. As shown in FIG. 11c, the first time point is the time point t3, i.e., the first time point is the ending time point of the usage time period of the second model or function corresponding to the second task, i.e., the first time point is the ending time point of the model inference of the second model or function.

[0275] Optionally, when the first time point corresponds to one of the following: the ending time point of the usage time period of the second model or function, the ending time point of the storage unit occupancy time period of the second model or function, and the ending time point of the computation unit occupancy time period of the second model or function, the first state of the second model or function is expected to continue from the first time point to a second time point, and the second time point is later than the first time point. The first state of the second model or function is expected to continue from the first time point to the second time point means that, without considering the first task, according to the second task, it can be expected that the second model or function corresponding to the second task is in the first state in the time period from the first time point to the second time point. For example, in FIG. 11c, without considering the influence of the first task, the first state of the second model or function corresponding to the second task is expected to continue from the first time point t3 to the second time point t5, i.e., in the time period between t3 and t5, the second model or function is expected to be in the activated and idle state.

[0276] Optionally, when the second task belongs to a periodic task or a semi-persistent task, the first time instant corresponds to one of: a starting time instant or an ending time instant of a reference signal corresponding to the second task; a starting time instant or an ending time instant of a time unit in which the reference signal corresponding to the second task is located; a starting time instant of a first time unit after the time unit in which the reference signal corresponding to the second task is located; a starting time instant or an ending time instant of a reference resource corresponding to the second task; a starting time instant or an ending time instant of a time unit in which the reference resource corresponding to the second task is located; a starting time instant of a first time unit after the time unit in which the reference resource corresponding to the second task is located; a starting time instant or an ending time instant of a physical uplink shared channel (PUSCH) or a physical uplink control channel (PUCCH) carrying the reporting information of the second task; a starting time instant or an ending time instant of a time unit in which the PUSCH or the PUCCH carrying the reporting information of the second task is located; a starting time instant of a first time unit after the time unit in which the PUSCH or the PUCCH carrying the reporting information of the second task is located; a time instant in which the starting time instant or the ending time instant of the reference signal corresponding to the second task is advanced or delayed by a time offset; a time instant in which the starting time instant or the ending time instant of the time unit in which the reference signal corresponding to the second task is located is advanced or delayed by the time offset; a time instant in which the starting time instant of the first time unit after the time unit in which the reference signal corresponding to the second task is located is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the reference resource corresponding to the second task is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the time unit in which the reference resource corresponding to the second task is located is advanced or delayed by the time offset; a time instant in which the starting time instant of the first time unit after the time unit in which the reference resource corresponding to the second task is located is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the reporting information of the second task is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the time unit in which the PUSCH or the PUCCH carrying the reporting information of the second task is located is advanced or delayed by the time offset; a time instant in which the starting time instant of the first time unit after the time unit in which the PUSCH or the PUCCH carrying the reporting information of the second task is located is advanced or delayed by the time offset. The reference signal corresponding to the second task can be the first downlink signal, and the PUSCH or the PUCCH carrying the reporting information of the second task can be the first uplink signal.

[0277] Optionally, when the second task belongs to a dynamically scheduled task, the first time instant corresponds to one of: a starting time instant or an ending time instant of a reference signal corresponding to the second task; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the reference signal corresponding to the second task is located; a starting time instant of a first time unit after the time unit in which the starting time instant or the ending time instant of the reference signal corresponding to the second task is located; a starting time instant or an ending time instant of a PDCCH triggering the second task; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the PDCCH triggering the second task is located; a starting time instant of a first time unit after the time unit in which the starting time instant or the ending time instant of the PDCCH triggering the second task is located; a starting time instant or an ending time instant of a PUSCH or a PUCCH carrying reporting information of the second task; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the reporting information of the second task is located; a starting time instant of a first time unit after the time unit in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the reporting information of the second task is located; a time instant at which the starting time instant or the ending time instant of the reference signal corresponding to the second task is advanced or delayed by a time offset; a time instant at which the starting time instant or the ending time instant of the time unit in which the starting time instant or the ending time instant of the reference signal corresponding to the second task is located is advanced or delayed by the time offset; a time instant at which the starting time instant of the first time unit after the time unit in which the starting time instant or the ending time instant of the reference signal corresponding to the second task is located is advanced or delayed by the time offset; a time instant after the starting time instant or the ending time instant of the PDCCH triggering the second task is advanced or delayed by the time offset; a time instant after the starting time instant or the ending time instant of the time unit in which the starting time instant or the ending time instant of the PDCCH triggering the second task is located is advanced or delayed by the time offset; a time instant after the starting time instant of the first time unit after the time unit in which the starting time instant or the ending time instant of the PDCCH triggering the second task is located is advanced or delayed by the time offset; a time instant at which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the reporting information of the second task is advanced or delayed by the time offset; a time instant at which the starting time instant or the ending time instant of the time unit in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the reporting information of the second task is located is advanced or delayed by the time offset; a time instant at which the starting time instant of the first time unit after the time unit in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the reporting information of the second task is located is advanced or delayed by the time offset. The reference signal corresponding to the second task and the PDCCH triggering the second task can be the first downlink signal, and the PUSCH or the PUCCH carrying the reporting information of the second task can be the first uplink signal.

[0278] In an optional embodiment, the first time instant is determined based on a sixth time instant and a sixth time length. The sixth time instant and the sixth time length have similar embodiments as the first time instant and the first time length, and are not repeated here.

[0279] In addition, the first time length can be an absolute time length without distinguishing sub-carrier spacing, such as the unit of the first time length being millisecond (ms). Alternatively, the first time length can also be a time length distinguishing sub-carrier spacing, such as the unit of the first time length being slot or symbol.

[0280] In an optional embodiment, the first time length can be determined according to an index or identifier corresponding to the first time length. For example, the network device configures the first time length for the terminal device through an index or identifier corresponding to the first time length, and the terminal device can determine the first time length according to the index or identifier corresponding to the first time length.

[0281] For example, the network device determines that there are 8 optional values of the first time length, and each index corresponding to 0 to 7 corresponds to a specified time length. The correspondence between the index and the time length is determined by a protocol preset or configuration.

[0282] For another example, the network device determines that the value of the first time length is 16 integers from 0 to 15, wherein "0" represents that the second model or function is always in the first state, "16" represents that the second model or function is used up and then the state of the second model or function is switched, and the remaining values represent the values of the first time length. Then, the terminal device can determine the first time length according to the value indicated by the network device, such as the network device indicating the value of 6, and the terminal device determining the first time length as 6 time units.

[0283] In another optional embodiment, when the second task belongs to a periodic task or a semi-persistent task, the first time length can be determined according to the reporting period of the second task, such as the first time length being 1 / M times of the reporting period of the second task, M being a positive integer.

[0284] In yet another optional embodiment, the first time length can be determined according to the time characteristics of the second task. For example, when the second task belongs to a periodic task or a semi-persistent task, the first time length needs to be configured, such as the first time length being 1 / M times of the reporting period of the second task or equal to the reporting period. For another example, when the second task belongs to a dynamically scheduled task, the first time length is 0, i.e., for a non-periodic task, the first time length does not need to be configured.

[0285] Optionally, the first time and the first time length can be configured by the network device to the terminal device. For example, the network device configures the first time and the first time length to the terminal device through the first information, i.e., the first information includes the first time and the first time length. Optionally, the first time and the first time length can be preset or pre-agreed by the network device and the terminal device.

[0286] For each task, the first time, the first time length and / or the second time can be independently configured or set. For example, for a periodic or semi-persistent task, the network device configures a set of first time and first time length, and the second time is determined according to the first time and the first time length; for a dynamically scheduled task, the network device configures another set of first time and first time length, and the second time is determined according to the first time and the first time length. For another example, for a periodic or semi-persistent task, the network device configures a set of first time and first time length, and the second time is determined according to the first time and the first time length; for a dynamically scheduled task, the network device configures or the protocol presets the second time as the end time of the task. For another example, for a periodic or semi-persistent task, the network device configures or the protocol presets the first time as the start time of multiple periods, and the second time as the end time of the multiple periods; for a dynamically scheduled task, the network device configures a set of first time and first time length, and the second time is determined according to the first time and the first time length. For another example, for a periodic or semi-persistent task, the network device configures or the protocol presets the first time as the start time of multiple periods, and the second time as the end time of the multiple periods; for a dynamically scheduled task, the network device configures or the protocol presets the second time as the end time of the task.

[0287] After the terminal device determines the first time and the first time length, the terminal device can determine the second time according to the first time and the first time length, i.e., the second time is the time when the first time is delayed by the first time length, and the time interval between the first time and the second time is the first time length. For example, the second time in FIGS. 7-9 is t3. For another example, the second time in FIGS. 10 and 11a is t4.

[0288] In an optional embodiment, the embodiment in which the terminal device determines the second time according to the first time and the first time length is applicable to the scenario in which the second task belongs to a periodic task or a semi-persistent task, such as the second task belongs to a periodic CSI reporting task or a semi-persistent CSI reporting task. Alternatively, the terminal device can determine the type of the second task, and when the type of the second task belongs to a periodic task or a semi-persistent task, the terminal device determines the second time according to the first time and the first time length.

[0289] Optionally, the terminal device determines the second time according to the first time and the first time length. The implementation manner of determining the second time according to the first time and the first time length can also be applied to a scenario in which the second task belongs to a dynamically scheduled task, for example, the second task belongs to a dynamically scheduled CSI reporting task.

[0290] In another optional implementation manner, when the second task belongs to a periodic task or a semi-persistent task, the first time corresponds to a start time of a first period in N periods, and the first time length is a time length occupied by the N periods, and the N periods are a plurality of periods to which the second task belongs. Thus, the second time corresponds to an end time of an Nth period in the N periods, and N is a positive integer greater than 1.

[0291] Alternatively, when the second task belongs to a periodic task or a semi-persistent task, the first time can be a start time of a first period in N periods to which the second task belongs, and the second time can be an end time of an Nth period in the N periods. Alternatively, when the second task belongs to a periodic task or a semi-persistent task, the first state of the second model or function lasts from a start time of a first period in N periods to which the second task belongs to an end time of an Nth period in the N periods. Alternatively, when the second task belongs to a periodic task or a semi-persistent task, the second model or function is in the first state during a time period of the N periods to which the second task belongs.

[0292] This is because a periodic task or a semi-persistent task lasts for a plurality of periods, and the second model or function is in the first state during a time period of the plurality of periods, which can avoid frequent switching between the first state and the second state, save switching time and switching overhead, and in addition, the terminal device does not need to determine the state each time, and the operation can be simplified.

[0293] In addition, when the second task belongs to a periodic task or a semi-persistent task and the second task is a task in a first period, the terminal device needs to determine the state of the second model or function corresponding to the second task when the second task is triggered. Optionally, the terminal device determines the state of the second model or function according to a duration of the first state of the model or function corresponding to a task before the second task. For details, refer to the implementation manner of determining the state of the first model or function according to the second time. Optionally, during a period other than the first period in a plurality of periods corresponding to the second task, the terminal device does not need to determine the state of the model or function used when performing another task related to the second task. During the period, the model or function used by the terminal device is in the first state.

[0294] In another optional implementation, when the second task belongs to a periodic task or a semi-persistent task, if the second task is not the last one of the N periods, the first time corresponds to the end time of the period in which the second task is located, and the second time corresponds to the start time of the next period of the period in which the second task is located; if the second task is the last one of the N periods, the first time corresponds to the end time of the period in which the second task is located, and the second time corresponds to the end time of the period in which the second task is located. The N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1. The end time of the period in which the second task is located can also be understood as the end time of model reasoning of the second model or function corresponding to the second task; and the start time of the next period of the period in which the second task is located can also be understood as the start time of model reasoning of the first model or function corresponding to the first task.

[0295] Alternatively, when the second task belongs to a periodic task or a semi-persistent task, the first state of the second model or function is periodically from the end time of each period to the start time of the next period. That is, in the N periods to which the second task belongs, the second model or function is periodically in the first state, which is the active and idle state.

[0296] For example, FIG. 11d is another CSI reporting schematic diagram. If the second task is the second periodic CSI reporting task in FIG. 11d, that is, the CSI reporting task corresponding to the second CSI RS in FIG. 11d, the first time is t3, the second time is t4, t3 is the end time of the period in which the second task is located, and t4 is the start time of the next period of the period in which the second task is located.

[0297] Optionally, when the second task belongs to a periodic task or a semi-persistent task, the third state of the second model or function starts from the second time, or the second time is the time when the terminal device switches the first state of the second model or function to the third state. The third state is a state in which the model or function is in the active state and is used, or the time period in which the second model or function is in the third state is a time period in which the terminal device performs model reasoning using the second model or function, or when the second model or function is in the third state, the terminal device is performing model reasoning using the second model or function.

[0298] Optionally, the first time is the time when the terminal device switches the third state of the second model or function to the first state, or the second model or function is in the third state in a period of time before the first time.

[0299] Optionally, the second model or function is in the second state at the end of the Nth period of the N periods, or in other words, the end of the Nth period of the N periods is the time at which the terminal device switches the third state of the second model or function to the second state, or in other words, the second state of the second model or function starts from the end of the Nth period of the N periods.

[0300] In yet another optional implementation, when the second task belongs to a periodic task or a semi-persistent task, the first time corresponds to the start of the first period of the N periods, and the second time corresponds to the end of the Nth period of the N periods. The N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1. Within the N periods to which the second task belongs, the second model or function periodically switches between the first state and the third state. In particular, at the end of each period of the N periods except the Nth period, the second model or function switches from the third state to the first state, and at the start of each period of the N periods except the first period, the second model or function switches from the first state to the third state. In particular, at the start of the first period of the N periods, the second model or function switches from the first state to the third state, or the second model or function switches from the second state to the third state. In particular, at the end of the Nth period of the N periods, the terminal device switches the third state of the second model or function to the second state, or in other words, the second model or function is in the second state at the end of the Nth period of the N periods, or in other words, the second state of the second model or function starts from the end of the Nth period of the N periods. The third state is a state in which the model or function is in an active state and is used, or in other words, the time period in which the second model or function is in the third state is a time period in which the terminal device performs model inference using the second model or function, or in other words, when the second model or function is in the third state, the terminal device is performing model inference using the second model or function.

[0301] For example, FIG. 11e is another schematic diagram of CSI reporting. FIG. 11e includes three periodic CSI reporting tasks, in which the second task is a CSI reporting task in the first period, and the first task is a CSI reporting task in the second period. As shown in FIG. 11e, the first time is the start time (t2) of the first period of the three periods, and the second time is the end time (t7) of the third period of the three periods. In addition, as shown in FIG. 11e, the periodicity of the second model or function corresponding to the second task switches between the first state and the third state. Specifically, at the end time t3 of the first period, the second model or function switches from the third state to the first state; at the start time t4 of the second period, the second model or function switches from the first state to the third state; at the end time t5 of the second period, the second model or function switches from the third state to the first state; and at the start time t6 of the third period, the second model or function switches from the first state to the third state. In addition, at the start time t2 of the first period, the second model or function switches from the second state to the third state; and at the end time t7 of the third period, the second model or function switches from the third state to the second state.

[0302] Optionally, when the second task belongs to a periodic task or a semi-persistent task, the second task is any one of the plurality of periods.

[0303] In an optional implementation, when the time interval between the first time and the second time is the first time length, the second task belongs to a task that does not need to report information.

[0304] Alternatively, the terminal device determines the second time based on the first time and the first time length, when the terminal device determines that the second task does not need to report information. When the second task belongs to a task that does not need to report information, or when the second task does not need to report information, it means that the result obtained by the terminal device executing the second task does not need to be directly reported through the uplink channel or the uplink signal. The result obtained by the terminal device executing the second task can be applied to other tasks, and the result obtained by the terminal device executing the other tasks can be directly reported through the uplink channel or the uplink signal. Alternatively, when the second task belongs to a task that does not need to report information, or when the second task does not need to report information, it means that the network device does not configure the terminal device with an uplink channel or an uplink signal for result feedback.

[0305] In another optional implementation, the second time point can not be determined according to the first time point and the first time length, and the second time point can be preset or configured or specified by the network device to the terminal device. In this case, the second time point can correspond to one of the following: a starting time point or an ending time point of a channel carrying the second task report information, a starting time point or an ending time point of a time unit in which the starting time point or the ending time point of the channel carrying the second task report information is located, or a starting time point of a first time unit after the time unit in which the starting time point or the ending time point of the channel carrying the second task report information is located. The channel carrying the second task report information can be, for example, a PUSCH or a PUCCH.

[0306] In this case, the second time point can correspond to a starting time point or an ending time point of a channel carrying the second task report information, and can be: the second time point is the starting time point or the ending time point of the channel carrying the second task report information, or can be: the second time point is determined according to the starting time point or the ending time point of the channel carrying the second task report information. Similarly, the second time point corresponds to other time points, and has similar meanings, which will not be described here.

[0307] In an optional implementation, the second time point corresponds to one of the following: a starting time point or an ending time point of a channel carrying the second task report information, a starting time point or an ending time point of a time unit in which the starting time point or the ending time point of the channel carrying the second task report information is located, or a starting time point of a first time unit after the time unit in which the starting time point or the ending time point of the channel carrying the second task report information is located. The implementation is applicable to a case where the second task belongs to a dynamically scheduled task, i.e., the second task belongs to a non-periodic task. This is because there is a false alarm and a missed detection probability for the DCI of the non-periodic task. If the network device indicates the first time length to the terminal device through the DCI, when a false alarm or a missed detection occurs, the network device and the terminal device cannot align the state of the first model or function. For example, the network device indicates a first time length of 3 slots to the terminal device through the DCI. When the first task is triggered within the first time length, the network device considers that the first model or function corresponding to the first task is in the first state, while the terminal device is in the second state in the case of missed detection, which leads to inconsistent understanding of the state of the first model or function between the network device and the terminal device.

[0308] Optionally, the terminal device determines the second time point to be the end time point of the first uplink signal when the second time point determined by the terminal device is earlier than the start time point or the end time point of the first uplink signal. For example, the second task is a CSI reporting task, and the terminal device determines the second time point to be earlier than the start time point of the PUSCH carrying the CSI report according to the first time point and the first time length. The terminal device determines the second time point to be the end time point of the PUSCH, that is, the first state of the second model or function corresponding to the second task lasts to the end time point of the PUSCH carrying the CSI report. This is because if the network device configures the second time point too early, the terminal device side may not have used the second model or function completely. At this time, if the second time point configured by the network device is used, the second model or function may be switched to the second state too early, thereby causing the second task to be unable to be executed normally.

[0309] In an optional implementation, the first state of the second model or function lasts from the first time point to the second time point, and the second model or function is the same as the first model or function. Then, when the terminal device is triggered to execute the first task, the state of the first model or function can be determined based on the state of the second model or function, that is, the state of the first model or function can be determined based on the second time point.

[0310] Specifically, the terminal device determines the state of the first model or function according to the second time point and the start time point of the first task, including: determining the state of the first model or function according to the first time point, the second time point, and the start time point of the first task. The terminal device determines the state of the first model or function according to the first time point, the second time point, and the start time point of the first task, including: when the start time point of the first task is located between the first time point and the second time point, determining that the first model or function is in the first state; and when the start time point of the first task is later than the second time point, determining that the first model or function is in the second state.

[0311] The start time point of the first task corresponds to one of the following: the start time point of the second downlink signal, the end time point of the second downlink signal, the third offset time point, the start time point of the use time period of the first model or function, the start time point of the storage unit occupation time period of the first model or function, and the start time point of the calculation unit occupation time period of the first model or function. The time interval between the third offset time point and the start time point or the end time point of the second downlink signal corresponds to a third time offset, which can be a positive number or a negative number. The third time offset can be configured or indicated by the network device to the terminal device, can be predefined, or can be reported by the terminal device.

[0312] The starting moment of the first task corresponds to the starting moment of the second downlink signal, which can be understood as that the starting moment of the first task is the starting moment of the second downlink signal, or can be understood as that the starting moment of the first task is determined based on the starting moment of the second downlink signal. Similarly, the second downlink signal corresponds to other moments, and has similar meanings, which will not be repeated.

[0313] It can be seen that the second model or function is in the first state between the first moment and the second moment, if the starting moment of the first task is between the first moment and the second moment, the first model or function is in the first state; if the starting moment of the first task is later than the second moment, the first model or function is in the second state.

[0314] Exemplarily, FIG. 12 is another schematic diagram of CSI reporting. As shown in FIG. 12, the periodic CSI reporting task triggered at t1 is the second task, and the aperiodic CSI reporting task triggered at t4 is the first task. The first moment and the second moment corresponding to the second task are t3 and t5 respectively, and the first state of the second model or function corresponding to the second task lasts from t3 to t5. The starting moment of the first task is the starting moment of the second downlink signal for scheduling the terminal device to perform the first task, that is, the moment t4, t4 is between t3 and t5, and the terminal device determines that the first model or function corresponding to the first task is in the active state.

[0315] Exemplarily, FIG. 13 is another schematic diagram of CSI reporting. As shown in FIG. 13, the periodic CSI reporting task triggered at t1 is the second task, and the aperiodic CSI reporting task triggered at t5 is the first task. The first moment and the second moment corresponding to the second task are t3 and t4 respectively, and the first state of the second model or function corresponding to the second task lasts from t3 to t4. The starting moment of the first task is the starting moment of the second uplink signal for scheduling the terminal device to perform the first task, that is, the moment t5, t5 is later than t4, and the terminal device determines that the first model or function corresponding to the first task is in the deactivated state.

[0316] It can be seen that the first model or function corresponding to the first task is in the first state, which includes that the starting moment of the first task is between the first moment and the second moment. The starting moment of the first task is between the first moment and the second moment, which can be that the starting moment of the first task is the first moment, or the starting moment of the first task is later than the first moment and earlier than the second moment.

[0317] In an optional embodiment, the start time and the end time of the first task are both between the first time and the second time. The end time of the first task corresponds to one of the following: the start time of the second uplink signal, the end time of the second uplink signal, the end time of the usage time period of the first model or function, the end time of the storage unit occupation time period of the first model or function, and the end time of the calculation unit occupation time period of the first model or function. When the end time of the first task corresponds to the start time of the second uplink signal, the end time of the first task can be the start time of the second uplink signal, or the end time of the first task can be determined based on the start time of the second uplink signal. Similarly, the end time of the first task corresponding to other time points has similar meanings, which will not be repeated here.

[0318] In the manner that the start time and the end time of the first task are both between the first time and the second time, the start time of the first task is the first time, and the end time of the first task is earlier than the second time or the second time. Alternatively, the start time of the first task is later than the first time, and the end time of the first task is earlier than the second time or the second time. For example, in FIG. 12, the first time is t3, the second time is t5, the start time of the first task is t4, t4 is between t3 and t5, the end time of the first task is t5, and the first function or model corresponding to the first task is in the active state.

[0319] In another optional embodiment, the start time of the first task is between the first time and the second time, and the end time of the first task is after the first time and the second time. In this manner, the start time of the first task is the first time, or the start time of the first task is later than the first time and earlier than the second time, and the end time of the first task is later than the second time. For example, FIG. 14a is another CSI reporting schematic diagram. In FIG. 14a, the first time is t3, the second time is t5, the start time of the first task is t4, t4 is between t3 and t5, the end time of the first task is t6, t6 is later than t5, and the first function or model corresponding to the first task is in the active state.

[0320] Similarly, the first model or function corresponding to the first task is in the second state, including: the first function or model is not in the first state; or the start time of the first task is after the second time, i.e., the start time of the first task is later than the second time. For example, in FIG. 13, the start time t5 of the first task is after the second time t4, and the first function or model corresponding to the first task is in the inactive state.

[0321] Optionally, when the second time is not determined based on the first time and the first time length, the first state of the second model or function continues to the second time, and the terminal device determines the state of the first model or function according to the second time and the start time of the first task, including: when the start time of the first task is earlier than the second time, determining that the first model or function is in the first state; and when the start time of the first task is later than the second time, determining that the first model or function is in the second state. That is, the first model or function is in the first state, including: the start time of the first task is earlier than the second time; and the first model or function is in the second state, including: the start time of the first task is later than the second time.

[0322] In an optional implementation, the first state is the active and idle state, and when the first model or function corresponding to the first task is in the first state, the first state of the first model or function continues to a seventh time, or the first state of the first model or function ends at the seventh time, or the seventh time is the time at which the terminal device switches the first state of the first model or function to a third state. The seventh time corresponds to the first task. Optionally, the seventh time corresponds to the start time of the first task. The seventh time corresponding to the start time of the first task can be that the seventh time is the start time of the first task, or the seventh time can be determined based on the start time of the first task, and the start time of the first task is the start time of model inference of the first model or function.

[0323] For example, in FIG. 11c, the first time is the end time of the use time period of the second model or function corresponding to the second task (t3 time), the second time is t5 time, and the seventh time is the start time of the first task (t6 time). The terminal device determines that the first model or function is in the active and idle state based on the start time of the first task, the first time, and the second time, and then the first state of the first model or function continues to t6 time.

[0324] Optionally, after the seventh time, the first state of the first model or function starts again at an eighth time, or the eighth time is the time at which the terminal device switches the third state of the first model or function to the first state. The eighth time corresponds to the first task. Optionally, the eighth time corresponds to the end time of the first task. The eighth time corresponding to the end time of the first task can be that the eighth time is the end time of the first task, or the eighth time can be determined based on the end time of the first task. For example, in FIG. 11c, the eighth time is t7 time, and the first state of the first model or function starts again at t7 time, that is, t7 time is the time at which the terminal device switches the third state of the first model or function to the first state.

[0325] Optionally, the first state of the first model or function starts at the ninth time, or in other words, the ninth time is the time when the first model or function switches from the third state to the first state. Optionally, the ninth time corresponds to the end time of the second task or the ninth time is the first time. In this case, the ninth time corresponds to the end time of the second task, which can be that the ninth time is the end time of the second task, or the ninth time is determined based on the end time of the second task, and the end time of the second task is the end time of the model inference of the second model or function. For example, in FIG. 11c, the ninth time is t3, which is the time when the first model or function switches from the third state to the first state, that is, the first state of the first model or function starts at t3, and t3 is the first time and the end time of the second task.

[0326] Optionally, the third state of the first model or function lasts until the eighth time, or in other words, the third state of the first model or function ends at the eighth time, or in other words, the eighth time is the time when the terminal device switches the third state of the first model or function to the second state. In this case, the eighth time corresponds to the first task or the eighth time is the second time. Optionally, the eighth time corresponds to the end time of the first task. In this case, the eighth time corresponds to the end time of the first task, which can be that the eighth time is the end time of the first task, or the eighth time is determined based on the end time of the first task, and the end time of the first task refers to the end time of the model inference.

[0327] For example, FIG. 14b is another CSI reporting diagram. In FIG. 14b, the third state of the first model or function lasts from t5 to t6, that is, the third state of the first model or function ends at t6, and t6 is the second time and the end time of the first task.

[0328] It can be understood that when the first model or function is in the first state, at least one of the following is different from when the first model or function is in the second state: latency requirement, usage time period of the first model or function, storage unit occupation time period of the first model or function, and computing unit occupation time period of the first model or function. Then, the terminal device can determine at least one of the following based on the state of the first model or function: latency requirement, usage time period of the first model or function, storage unit occupation time period of the first model or function, and computing unit occupation time period of the first model or function, to execute the first task according to the determined at least one.

[0329] The terminal device determines at least one of the following according to the state of the first model or function: a latency requirement, a usage time period of the first model or function, a storage unit occupation time period of the first model or function, a computing unit occupation time period of the first model or function, and executes the first task according to the determined at least one.

[0330] Embodiment 1: The terminal device determines a latency requirement according to the state of the first model or function, and executes the first task according to the latency requirement.

[0331] In an optional embodiment, the latency requirement is a time interval requirement between a second downlink signal and a second uplink signal, and the terminal device determines the time interval requirement between the second downlink signal and the second uplink signal according to the state of the first model or function. When the first model or function is in the first state, the time interval requirement between the second downlink signal and the second uplink signal includes a time requirement for model inference, that is, a time for the terminal device to perform model inference using the first model or function; when the first model or function is in the second state, the latency requirement includes a time requirement for model activation and model inference, that is, a time for the terminal device to activate the first model or function and a time for the terminal device to perform inference using the first model or function. Therefore, when the first model or function is in the first state, the latency requirement corresponding to the first task is different from that when the first model or function is in the second state. Moreover, when the first model or function is in the first state, the latency requirement corresponding to the first task is smaller than that when the first model or function is in the second state.

[0332] Further, the terminal device executes the first task according to the determined latency requirement, including: when the first model or function is in the first state, performing model inference using the first model or function based on the latency requirement to obtain an inference result; and when the first model or function is in the second state, activating the first model or function based on the latency requirement, and performing model inference using the activated first model or function to obtain an inference result.

[0333] It can be seen that, when the first model or function is in the first state, compared with when the first model or function is in the second state, the terminal device does not need to activate the first model or function when executing the first task based on the latency requirement, and can directly perform model inference using the first model or function, thereby saving resources.

[0334] For example, in FIG. 12, the terminal device determines that the first model or function is in the activated state, and determines that the latency requirement is the time interval requirement between the starting time t6 of the CSI RS when the terminal device performs the first task and the ending time t5 of the PUSCH carrying the CSI report, which includes the time for the terminal device to perform model inference using the first model or function, and does not include the time for the terminal device to activate the first model or function. Then, when the terminal device performs the first task, the terminal device performs model inference using the first model or function to obtain predicted CSI based on the time interval requirement between t6 and t5, and the terminal device does not need to activate the first model or function, which can save resource overhead.

[0335] For example, in FIG. 13, the terminal device determines that the first model or function is in the deactivated state, and determines that the latency requirement is the time interval requirement between the starting time t7 of the CSI RS when the terminal device performs the first task and the ending time t6 of the PUSCH carrying the CSI report, which includes the model activation time for the terminal device to activate the first model or function and the time for the terminal device to perform model inference using the first model or function. Then, when the terminal device is triggered to perform the first task, the terminal device activates the first model or function based on the time interval requirement between t7 and t6, and then performs model inference using the activated first model or function to obtain predicted CSI.

[0336] In an optional implementation, the terminal device performs the first task according to the determined latency requirement, which can further include: when the first model or function is in the first state, it is determined that the time interval between the configured second downlink signal and the second uplink signal can meet the latency requirement, so that the first task can be performed, or in other words, the first task is not ignored; when the first model or function is in the second state, it is determined that the latency requirement is greater, and the time interval between the configured second downlink signal and the second uplink signal does not meet the latency requirement, so that the first task cannot be performed, or in other words, the first task is ignored.

[0337] In an optional implementation, when the first model or function is in the first state, after the terminal device receives the signal or SSB for performing the first task, the terminal device performs model inference by using the first model or function. In this manner, the time interval between the signal or SSB for performing the first task and the PUSCH or PUCCH for carrying the reported information is greater than a predefined first time interval by Δt3, where Δt3 is a value reported by the terminal device or a value predefined by a protocol. In other words: when the network device schedules time for the terminal device to perform model inference when the first model or function is in the first state, the network device schedules time resources for the terminal device by Δt3 more than the predefined time. Optionally, the predefined first time interval corresponds to the time interval requirement between the signal or SSB for performing the first task and the PUSCH or PUCCH for carrying the reported information in a non-AI mode, or corresponds to the first latency requirement for performing the task in the non-AI mode. For example, the predefined first time interval is T' proc,CSI , T' proc,CSI may be understood as T' proc,CSI corresponds to the first latency requirement for performing CSI reporting in the non-AI mode, and Δt3 is the time increment of the time length required for running the first AI model or function to perform model inference to perform CSI reporting relative to the first latency requirement for performing CSI reporting in the non-AI mode. In another possible implementation, the predefined first time interval corresponds to the first reference latency requirement for performing the task in the AI mode, for example, the predefined first time interval is the time interval requirement between the CSI-RS corresponding to running the AI model or function to perform CSI reporting and the PUSCH or PUCCH for carrying the reported information, which is predefined by a protocol, and the first reference latency requirement can be the minimum latency requirement or the average latency requirement, and Δt3 is the time increment of the time length required for running the first AI model or function to perform model inference to perform the task relative to the first reference latency requirement. If the time length required for running the first AI model or function to perform model inference to perform the task is less than the predefined first time interval, Δt3 can be 0. In this manner, it can be ensured that, after the terminal device receives the signal or SSB for performing the first task, there is sufficient time to perform inference of the first model or function, and thus it can be ensured that the terminal device can successfully perform the first task by using the first model or function, or in other words, it can be ensured that the model inference of the first model or function is successfully performed.

[0338] In an optional implementation, the terminal device activates the first model or function again after receiving the signal or SSB for performing the first task when the first model or function is in the second state. In this mode, the time interval between the signal or SSB for performing the first task and the PUSCH or PUCCH for carrying the reported information is greater than a predefined first time interval by Δt1, where Δt1 is a value reported by the terminal device or a value predefined by a protocol. In other words, the network device schedules more time resources than Δt1 for the terminal device to perform model activation and inference when the first model or function is in the second state than the predefined first time interval. Optionally, the predefined first time interval corresponds to the time interval between the signal or SSB for performing the first task and the PUSCH or PUCCH for carrying the reported information in the non-AI mode, or the first latency requirement for performing the task in the non-AI mode. For example, the predefined first time interval is T' proc,CSI , T' proc,CSI As described above, it can be understood that T' proc,CSI corresponds to the first latency requirement for performing CSI reporting in the non-AI mode, and Δt1 is the time increment of the time length required for running the first AI model or function to perform CSI reporting relative to the first latency requirement for performing CSI reporting in the non-AI mode. In another possible implementation, the predefined first time interval corresponds to the first reference latency requirement for performing the task in the AI mode, for example, the predefined first time interval is the time interval requirement between the CSI-RS corresponding to running the AI model or function to perform CSI reporting and the PUSCH or PUCCH for carrying the reported information defined by a protocol, and the first reference latency requirement can be the minimum latency requirement or the average latency requirement, and Δt1 is the time increment of the time length required for running the first AI model or function to perform the task relative to the first reference latency requirement. The time length required for running the first AI model or function to perform the task includes the time for model activation and model inference. If the time length required for running the first AI model or function to perform the task is less than the predefined first time interval, Δt1 can be 0.

[0339] Optionally, if the first task is a dynamically scheduled task, in this mode, the time interval between the PDCCH for triggering the first task and the PUSCH or PUCCH for carrying the reported information is greater than a predefined second time interval by Δt1. Optionally, the predefined second time interval corresponds to the time interval requirement between the PDCCH for triggering the first task and the PUSCH or PUCCH for carrying the reported information in the non-AI mode, or the second latency requirement for performing the task in the non-AI mode. For example, the predefined second time interval is T proc,CSI , T proc,CSIThe predefined time corresponds to a second reference time delay requirement for performing a task in an AI mode.

[0340] The method can ensure that, after the terminal device receives the signal or SSB for performing the first task, there is sufficient time to activate the first model or function, thereby ensuring that the terminal device can successfully perform the first task using the first model or function, or in other words, ensuring that the model inference of the first model or function is successfully performed.

[0341] For example, in FIG. 13, the time interval between the CSI RS for measuring CSI and the PUSCH for carrying the CSI report is increased from t9-t6 to t8-t6 when the terminal device performs the first task, and the interval between t9 and t8 is Δt1. Thus, after the terminal device receives the CSI RS, the terminal device can complete the activation of the first model or function within the t8-t10 time period, and then can perform model inference using the activated first model or function to obtain predicted CSI and complete CSI reporting.

[0342] In another optional implementation, when the first model or function is in the second state, the terminal device starts to activate the first model or function at a time point Δt1 before receiving the signal or SSB for performing the first task. In other words, the terminal device activates the first model or function before receiving the signal or SSB for performing the first task, and the time interval between the time point of activating the first model or function and the time of receiving the signal or SSB for performing the first task is Δt1. In this method, the time interval between the signal or SSB for performing the first task and the PUSCH or PUCCH for carrying the reported information satisfies the predefined time interval, but does not satisfy the time delay requirement when the first model or function is in the second state, because the time delay requirement when the first model or function is in the second state is greater. Therefore, the terminal device can perform model activation Δt1 in advance.

[0343] Optionally, the predefined first time interval corresponds to a time interval requirement between the signal or SSB for performing the first task and the PUSCH or PUCCH for carrying the reported information in a non-AI mode, or in other words, corresponds to a first time delay requirement for performing a task in a non-AI mode. For example, the predefined first time interval is T' proc,CSI , T' proc,CSI As described above, it can be understood that T' proc,CSIThe first time delay requirement corresponds to CSI reporting performed in a non-AI mode. Δt1 is a time increment of a time length required for running the first AI model or function to perform CSI reporting relative to the first time delay requirement of CSI reporting performed in the non-AI mode. In another possible implementation, the predefined time corresponds to a first reference time delay requirement of performing a task in an AI mode. For example, the predefined first time interval is a first reference time delay requirement of running an AI model or function to perform CSI reporting defined by a protocol. The first reference time delay requirement can be a minimum time delay requirement or an average time delay requirement. Δt1 is a time increment of a time length required for running the first AI model or function to perform the task relative to the first reference time delay requirement. The time length required for running the first AI model or function to perform the task includes a time of model activation and a time of model inference. If the time length required for running the first AI model or function to perform the task is less than the predefined first time interval, Δt1 can be 0.

[0344] In addition, the terminal device cannot completely activate the first model or function before receiving the signal or SSB for performing the first task. After the terminal device receives the signal or SSB for performing the first task, the terminal device activates the first model or function and switches the state of the first model or function to the first state based on the activation.

[0345] Optionally, if the first task is a dynamically scheduled task, in this mode, a time interval between a PDCCH for triggering the first task and a PUSCH or PUCCH for carrying the reported information is greater than a predefined second time interval by Δt1. Optionally, the predefined second time interval corresponds to a time interval requirement between the PDCCH for triggering the first task and the PUSCH or PUCCH for carrying the reported information in a non-AI mode, or corresponds to a second time delay requirement of performing the task in the non-AI mode. For example, the predefined time is T proc,CSI , T proc,CSI Reference can be made to the foregoing description. Alternatively, the predefined second time interval corresponds to a second reference time delay requirement of performing the task in an AI mode. This mode can also ensure that the terminal device can complete the activation of the first model or function using sufficient resources, and then perform model inference using the activated first model or function, to ensure smooth execution of the first task.

[0346] For example, FIG. 15a is another CSI reporting schematic diagram. In FIG. 15a, the starting time t5 of the first task is later than the second time t4, the first model or function corresponding to the first task is in the second state, and the terminal device starts model activation at a time point of t5-t1 before receiving the CSI RS. After the terminal device receives the CSI RS, the terminal device continues to activate the first model or function, and performs model inference using the activated first model or function from t9 after the first model or function is activated. Compared with activating the first model or function after the terminal device receives the CSI RS, this mode can ensure the smooth execution of the first task.

[0347] In another optional implementation, when the first model or function is in the second state, the terminal device starts to activate the first model or function at a time point of t4-t2 before the configured PUSCH or PUCCH for carrying the reported information. In other words, the time interval between the time point at which the terminal device activates the first model or function and the configured PUSCH or PUCCH for carrying the reported information is t2. t2 is a value reported by the terminal device or a value predefined by a protocol. In this mode, the terminal device can perform model activation t2 before the configured reporting time point. It can be understood that t2 is the time length required for running the first AI model or function to execute the task, or in other words, t2 includes the time of model activation and model inference. The time point at which the terminal device performs model activation is after receiving the signal for triggering the first task. The time point at which the terminal device performs model activation can be before or after receiving the signal for executing the first task or the SSB. Based on this mode, the terminal device can complete the activation and inference of the first model or function using sufficient resources before the configured PUSCH or PUCCH for carrying the reported information, and thus the smooth execution of the first task can be ensured.

[0348] Optionally, if the first task is a dynamically scheduled task, in this mode, the second time interval between the PDCCH for triggering the first task and the PUSCH or PUCCH for carrying the reported information needs to be greater than t2, or the time interval between the PDCCH for triggering the first task and the PUSCH or PUCCH for carrying the reported information is greater than the predefined second time interval by t4, t4 is a value reported by the terminal device or a value predefined by a protocol. Optionally, t4 corresponds to the difference between t2 and the predefined second time interval. If t2 is less than the predefined second time interval, t4 can be 0. Optionally, the predefined second time interval corresponds to the time interval requirement between the PDCCH for triggering the first task and the PUSCH or PUCCH for carrying the reported information in a non-AI mode, or in other words, corresponds to the second delay requirement for executing the task in the non-AI mode. For example, the predefined time is T proc,CSI , T proc,CSIThe above description can be referred to. Alternatively, the predefined second time interval corresponds to a second reference latency requirement for performing the task in the AI mode.

[0349] For example, FIG. 15b is another schematic diagram of CSI reporting. As shown in FIG. 15b, the starting time t5 of the first task is later than the second time t4, and the first model or function is in the second state. In FIG. 15b, the terminal device activates the first model or function at a time point (i.e., at the time point t8) that is Δt2 before the PUSCH for reporting information on the configured bearer, so as to ensure that the terminal device can complete the activation and inference of the first model or function using sufficient resources before the PUSCH for reporting information on the configured bearer, and thus the smooth execution of the first task can be ensured.

[0350] Similarly, when the latency requirement is a time interval requirement between the second downlink signal and the end time of the usage time period of the first model or function, or the latency requirement is a time interval requirement between the second downlink signal and the end time of the storage unit occupation time period of the first model or function, or the latency requirement is a time interval requirement between the second downlink signal and the end time of the calculation unit occupation time period of the first model or function, if the first model or function is in the first state, the latency requirement includes the model inference time; if the first model or function is in the second state, the latency requirement includes the model activation time and the model inference time. In this mode, the terminal device performs the first task according to the determined latency requirement, which can refer to the implementation mode of the latency requirement being a time interval requirement between the second downlink signal and the second uplink signal, which will not be described again.

[0351] Implementation Mode 2: The terminal device determines the usage time period of the first model or function according to the state of the first model or function, and performs the first task according to the usage time period of the first model or function.

[0352] When the first model or function is in the first state, the usage time period of the first model or function starts from the starting time or the ending time of the second downlink signal corresponding to the first task. When the first model or function is in the first state, after the terminal device receives the second downlink signal corresponding to the first task, the first model or function can be directly used for model inference, and thus the usage time period of the first model or function can start from the starting time or the ending time of the second downlink signal corresponding to the first task. For example, in FIG. 12, the first model or function corresponding to the first task is in the activated state, and the terminal device does not need to activate the first model or function when performing the first task, and thus the usage time period of the first model or function starts from the starting time or the ending time of the CSI RS corresponding to the first task.

[0353] The starting moment of the use time period of the first model or function is later than the starting moment or the ending moment of the second downlink signal corresponding to the first task when the first model or function is in the second state. In other words, the time interval between the starting moment of the use time period of the first model or function and the starting moment or the ending moment of the second downlink signal corresponding to the first task corresponds to the fourth time offset, and the fourth time offset is positive. After the terminal device receives the second downlink signal corresponding to the first task, the first model or function needs to be activated, and then the activated first model or function is used for model inference. Therefore, the starting moment of the use time period of the first model or function is later than the starting moment or the ending moment of the second downlink signal corresponding to the first task.

[0354] For example, in FIG. 13, the first model or function corresponding to the first task is in the deactivated state. The terminal device needs to activate the first model or function, and then uses the activated first model or function for model inference. Therefore, the starting moment of the use time period of the first model or function is later than the starting moment or the ending moment of the CSI RS corresponding to the first task, for example, the starting moment of the use time period of the first model or function is t10.

[0355] The terminal device determines the use time period of the first model or function according to the state of the first model or function. When the terminal device executes the first task, the first model or function can be used for inference in the determined use time period of the first model or function to obtain an inference result.

[0356] In an embodiment 3, the terminal device determines a time period occupied by a calculation unit of the first model or function according to the state of the first model or function, and executes the first task according to the time period occupied by the calculation unit of the first model or function.

[0357] In a possible implementation, the time period occupied by the calculation unit of the first model or function can be the use time period of the first model or function. When the first model or function is in the second state or in the time period of the transition from the second state to the first state, the calculation unit occupied by the first model or function is 0, or in other words, the first model or function does not occupy the calculation unit. When the first model or function is in the first state, the time period occupied by the calculation unit of the first model or function starts from the starting or ending moment of the second downlink signal corresponding to the first task. When the first model or function is in the second state, the starting moment of the time period occupied by the calculation unit of the first model or function is later than the starting or ending moment of the second downlink signal corresponding to the first task. Alternatively, when the first model or function is in the second state, the time interval between the starting moment of the time period occupied by the calculation unit of the first model or function and the starting or ending moment of the second downlink signal corresponding to the first task corresponds to the fourth time offset.

[0358] The terminal device determines the time period occupied by the computing unit of the first model or function according to the state of the first model or function. For details, refer to the implementation of the terminal device determining the use time period of the first model or function according to the state of the first model or function, which will not be repeated here.

[0359] After the terminal device determines the time period occupied by the computing unit of the first model or function according to the state of the first model or function, when the terminal device executes the first task, the terminal device can use the first model or function to perform reasoning in the time period occupied by the computing unit of the first model or function, and obtain a reasoning result. In addition, the terminal device can determine a time period outside the time period occupied by the computing unit of the first model or function, and the first model or function does not occupy the computing unit, so that the terminal device can use these computing units to execute other tasks, and improve the utilization efficiency of the computing unit.

[0360] Implementation 4: The terminal device determines the time period occupied by the storage unit of the first model or function according to the state of the first model or function, and executes the first task according to the time period occupied by the storage unit of the first model or function.

[0361] The time period occupied by the storage unit of the first model or function is calculated from the time when the first model or function is in the first state to the time when the first model or function is not in the first state; or the time period occupied by the storage unit of the first model or function is calculated from the time when the second state is converted to the first state to the time when the first model or function is not in the first state.

[0362] Therefore, when the first model or function corresponding to the first task is in the first state, the time period occupied by the storage unit of the first model or function corresponding to the first task starts from the starting time or the ending time of the second downlink signal corresponding to the first task. The second downlink signal can be a downlink channel used to trigger the first task, or can be a downlink signal used to execute the first task. For example, in FIG. 12, the first model or function corresponding to the first task is in the first state, and the time period occupied by the storage unit of the first model or function can start from the starting time t4 of the DCI used to trigger the CSI report, or can start from the starting time t6 of the CSI RS used to measure the CSI.

[0363] When the first model or function corresponding to the first task is in the second state, the starting moment of the time period occupied by the storage unit of the first model or function corresponding to the first task is earlier than the starting moment or the ending moment of the second downlink signal corresponding to the first task, and the second downlink signal is a downlink signal used to execute the first task. For example, in FIG. 15a, the first model or function corresponding to the first task is in the second state, and the terminal device activates the first model or function before receiving the CSI RS, so the starting moment of the time period occupied by the storage unit of the first model or function is earlier than the starting moment or the ending moment of the CSI RS.

[0364] Optionally, when the first model or function corresponding to the first task is in the second state, the starting moment of the time period occupied by the storage unit of the first model or function corresponding to the first task is later than the starting moment or the ending moment of the second downlink signal corresponding to the first task, and the second downlink signal is a downlink signal used to trigger the first task. For example, in FIG. 13, the first model or function corresponding to the first task is in the second state, and the terminal device activates the first model or function after receiving the CSI RS, so the starting moment of the time period occupied by the storage unit of the first model or function is later than the starting moment or the ending moment of the PDCCH triggering the first task.

[0365] Optionally, when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start or end time of the second uplink signal corresponding to the first task, and the start time of the storage occupation time period of the first model or function corresponding to the first task is determined by the start or end time of the second uplink signal corresponding to the first task, and the second uplink signal is an uplink channel carrying the reported information, such as PUSCH or PUCCH. For example, the time interval between the start time of the storage occupation time period of the first model or function corresponding to the first task and the start or end time of the second uplink signal corresponding to the first task corresponds to a fifth time offset, and the value of the fifth time offset is negative. That is, the start time of the storage occupation time period of the first model or function corresponding to the first task is the time point corresponding to the absolute value of the fifth time offset after the start time or the end time of the second uplink signal. The terminal device determines the start time of the storage unit occupation time period of the first model or function according to the state of the first model or function, and when the terminal device executes the first task, it can determine whether the first model or function needs to be activated based on the start time of the storage unit occupation time period of the first model or function. For example, if the storage unit occupation time period of the first model or function corresponding to the first task starts from the start time or the end time of the second downlink signal corresponding to the first task, when the terminal device executes the first task, the first model or function does not need to be activated, and the first model or function can be directly used for inference to obtain the inference result. For another example, if the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second downlink signal corresponding to the first task, or the start time of the storage unit occupation time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task, or the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start or end time of the second uplink signal corresponding to the first task, and the start time of the storage occupation time period of the first model or function corresponding to the first task is determined by the start or end time of the second uplink signal corresponding to the first task, when the terminal device executes the first task, the first model or function is activated, and the activated first model or function is used for model inference to obtain the inference result.

[0366] When the first model or function corresponding to the first task is in the second state, the terminal device successfully completes the first task before the starting time or the ending time of the second uplink signal scheduled by the network device, the first model or function is activated from the time earlier than the starting time or the ending time of the second uplink signal, so that the starting time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the starting time or the ending time of the second uplink signal corresponding to the first task. Further, the starting time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the starting time or the ending time of the second uplink signal corresponding to the first task, and the time interval between the starting time of the storage unit occupation time period of the first model or function corresponding to the first task and the starting time or the ending time of the second uplink signal corresponding to the first task needs to be greater than or equal to the sum of the model activation time and the time for which the first task uses the model for inference.

[0367] Optionally, when the first model or function corresponding to the first task is in the second state, the starting time of the use time period of the first model or function is earlier than the time when the first model or function is in the first state, and the starting time of the calculation unit occupation time period of the first model or function is earlier than the time when the first model or function is in the first state. In other words, the starting time of the use time period of the first model or function is earlier than the starting time or the ending time of the second uplink signal corresponding to the first task, and the starting time of the calculation unit occupation time period of the first model or function is earlier than the starting time or the ending time of the second uplink signal corresponding to the first task.

[0368] For example, in FIG. 15b, the first model or function corresponding to the first task is in the second state, the starting time t8 of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the starting time of the PUSCH corresponding to the first task, and the starting time of the storage occupation time period of the first model or function corresponding to the first task is determined by the starting or ending time of the PUSCH corresponding to the first task, such as the time t8 located before the starting time t10 of the PUSCH by Δt2.

[0369] It can be understood that, when the first model or function is in the first state, the storage unit occupation time period of the first model or function corresponding to the first task is longer than when the first model or function is in the second state. When the first model or function is in the first state, the first model or function corresponding to the first task does not need to be activated, so the storage unit occupation time period of the first model or function corresponding to the first task does not include the time for switching from the first state to the second state; when the first model or function is in the second state, the first model or function corresponding to the first task needs to be activated first, so the storage unit occupation time period of the first model or function corresponding to the first task also needs to include the time for switching from the first state to the second state.

[0370] The terminal device determines the storage unit occupation time period of the first model or function according to the state of the first model or function. When the terminal device executes the first task, the occupation demand of the first model or function for the storage unit can be determined based on the storage unit occupation time period of the first model or function, so as to ensure that the first task can be executed smoothly. In addition, the terminal device can determine a time period outside the storage unit occupation time period of the first model or function, and the first model or function does not occupy the computing unit, so that the terminal device can execute other tasks by using the storage unit.

[0371] In another optional implementation, the use time period of the first model or function can be the storage unit occupation time period of the first model or function. The implementation in which the terminal device determines the use time period of the first model or function according to the state of the first model or function can refer to the implementation of the terminal device determining the storage occupation time period of the first model or function according to the state of the first model or function, which will not be described herein again. The terminal device determines the use time period of the first model or function according to the state of the first model or function. When the terminal device executes the first task, the terminal device can use the first model or function to perform reasoning to obtain a reasoning result in the determined use time period of the first model or function.

[0372] In another optional implementation, the computing unit occupation time period of the first model or function can be the storage unit occupation time period of the first model or function. The implementation in which the terminal device determines the computing unit occupation time period of the first model or function according to the state of the first model or function can refer to the implementation of the terminal device determining the storage occupation time period of the first model or function according to the state of the first model or function, which will not be described herein again. The terminal device determines the computing unit occupation time period of the first model or function according to the state of the first model or function. When the terminal device executes the first task, the occupation demand of the first model or function for the computing unit can be determined based on the computing unit occupation time period of the first model or function, so as to ensure that the first task can be executed smoothly.

[0373] Embodiment 5: The terminal device determines at least two of the following according to the state of the first model or function: the latency requirement, the use time period of the first model or function, the storage unit occupation time period of the first model or function, and the computing unit occupation time period of the first model or function, and executes the first task according to the determined at least two.

[0374] In the embodiments, the terminal device determines the state of the first model or function according to any one of the following: a time delay requirement, a use time period of the first model or function, a storage unit occupation time period of the first model or function, and a computing unit occupation time period of the first model or function. The embodiments of the terminal device performing the first task according to any one of the following: a time delay requirement, a use time period of the first model or function, a storage unit occupation time period of the first model or function, and a computing unit occupation time period of the first model or function are similar to the embodiments 1 to 4, and are not described herein again.

[0375] It can be seen that the terminal device can determine the state of the first model or function corresponding to the first task, and determine at least one of the following according to the state of the first model or function: a time delay requirement, a use time period of the first model or function, a storage unit occupation time period of the first model or function, and a computing unit occupation time period of the first model or function. Therefore, the terminal device can perform the first task according to the determined at least one. When the first model or function is in the first state, compared with the terminal device being triggered to perform the first task directly, the network device does not need to schedule resources for activating the first model or function for the terminal device, and resource waste can be reduced. When the first model or function is in the second state, compared with the terminal device being triggered to perform the first task directly, the terminal device can activate the first model or function with sufficient resources, and perform model inference with the activated first model or function, that is, the terminal device can successfully perform the first task.

[0376] Optionally, after the terminal device performs model inference with the activated first model or function to obtain an inference result, the terminal device can report the inference result or report information obtained according to the inference result to the network device, for example, in the form of reporting.

[0377] In an optional embodiment, the terminal device can further determine a third time corresponding to the first task. The terminal device can determine the third time corresponding to the first task based on a fourth time corresponding to the first task and a second time length. The embodiments of the fourth time and the second time length are similar to the embodiments of the first time and the first time length, and are not described herein again. Optionally, the third time is configured or indicated by the network device, or is preset. For example, when the first task belongs to a dynamically scheduled task, the third time is an end time of the terminal device performing the first task. The embodiment of the third time is similar to the embodiment of the second time, and is not described herein again.

[0378] The third moment is used for the terminal device to determine the state of the third model or function corresponding to the third task, and is further used for the terminal device to determine at least one of the following corresponding to the third task: a latency requirement, a use time period of the third model or function, a storage unit occupation time period of the third model or function, and a computing unit occupation time period of the third model or function. That is, the third moment acts on at least one of the following corresponding to the third task: a latency requirement, a use time period of the third model or function, a storage unit occupation time period of the third model or function, and a computing unit occupation time period of the third model or function.

[0379] The starting moment of the third task is later than the starting moment of the first task. The third model or function corresponding to the third task is the same as the first model or function corresponding to the first task. The third model or function is the same as the first model or function, which means that the third model or function and the first model or function are used to implement the same function, or the physical model corresponding to the third model or function and the first model or function is the same. The third model or function is the same as the first model or function in structure and all parameters, or the third model or function is the same as the first model or function in structure and part of the parameters, such as the number of layers, width, and inter-layer connection relationship of a neural network, and the weight value and bias of the neural network.

[0380] Optionally, the starting moment of the third task corresponds to one of the following: the starting moment of the third downlink signal, the ending moment of the third downlink signal, the fifth offset moment, the starting moment of the use time period of the third model or function, the starting moment of the storage unit occupation time period of the third model or function, and the starting moment of the computing unit occupation time period of the third model or function. The time interval between the fifth offset moment and the starting moment or ending moment of the third downlink signal corresponds to the eighth time offset. The third downlink signal is related to the third task. Specifically, the third downlink signal can be a channel, such as a PDCCH or DCI, for triggering the terminal device to perform the third task, or can be a signal, such as one of a CSI RS, a TRS, a PTRS, a PRS, and an SSB, used by the terminal device to perform the third task.

[0381] The starting moment of the third task corresponds to the starting moment of the third downlink signal, which can be that the starting moment of the third task is the starting moment of the third downlink signal, or can be that the starting moment of the third task is determined according to the starting moment of the third downlink signal. The starting moment of the third task is determined based on the starting moment of the third downlink signal, which can be that the starting moment of the third task is the starting moment or ending moment of the time unit in which the starting moment of the third downlink signal is located, or can be that the starting moment of the third task is the starting moment of the first time unit after the time unit in which the starting moment of the third downlink signal is located. Similarly, the starting moment of the third task corresponds to other moments, which has a similar understanding and will not be repeated.

[0382] In an optional implementation, when the first model or function corresponding to the first task is in the first state, the terminal device further determines a duration of the first state of the first model or function. Specifically, the terminal device determines that the first state of the first model or function lasts until the second time corresponding to the second task, or determines that the first state of the first model or function lasts until the third time corresponding to the first task.

[0383] In an optional implementation, the terminal device determines the duration of the first state of the first model or function based on the priority of the first task and the priority of the second task. For example, when the priority of the first task is higher than the priority of the second task, the terminal device determines that the first state of the first model or function lasts until the third time corresponding to the first task. For another example, when the priority of the first task is lower than the priority of the second task, the terminal device determines that the first state of the first model or function lasts until the second time corresponding to the second task.

[0384] Optionally, the terminal device determines the duration of the first state of the first model or function according to the start time of the task. For example, when the start time of the first task is later than the start time of the second task, the terminal device determines that the first state of the first model or function lasts until the third time corresponding to the first task. For another example, when the start time of the first task is earlier than the start time of the second task, the terminal device determines that the first state of the first model or function lasts until the second time corresponding to the second task.

[0385] Optionally, the terminal device determines the duration of the first state of the first model or function according to the identity of the task. For example, when the identity of the first task is greater than the identity of the second task, the terminal device determines that the first state of the first model or function lasts until the third time. For another example, when the identity of the second task is greater than the identity of the first task, the terminal device determines that the first state of the first model or function lasts until the second time corresponding to the second task. The identity of the first task and the identity of the second task can be an identity used to identify the two tasks.

[0386] Optionally, when the first task is a reference task specified by the network device, the terminal device determines that the first state of the first model or function lasts until the third time corresponding to the first task. Optionally, when the second task is a reference task specified by the network device, the terminal device determines that the first state of the first model or function lasts until the second time corresponding to the second task.

[0387] Optionally, the terminal device determines the duration time of the first state of the first model or function according to the time characteristic of the task, the time characteristic of the task including periodicity, semi-persistent, and aperiodicity. For example, the terminal device determines that the first state of the first model or function lasts to the third time point when the first task is an aperiodic task. For another example, the terminal device determines that the first state of the first model or function lasts to the second time point when the first task is a periodic task.

[0388] In the various embodiments of the terminal device determining the duration time of the first state of the first model or function, the third time point is later than the second time point, or the third time point is earlier than the second time point.

[0389] When the first state of the first model or function lasts to the second time point, the second time point is the end time point of the first state of the first model or function; in other words, the first state of the first model or function ends at the second time point; in other words, the terminal device starts to convert the first state of the first model or function to the second state from the second time point; in other words, the terminal device can determine that the first state of the first model or function is the first state before the second time point according to the second time point. In an optional embodiment, after the second time point, the first model or function is in the second state, in other words, the second state of the first model or function starts from the second time point, or in other words, the second time point is the start time point of the second state of the first model or function. In another optional embodiment, the second time point is delayed by a sixth time offset, and the first model or function is in the second state, that is, the time period between the second time point and the time point delayed by the sixth time offset is the time period for the terminal device to switch the first model or function from the first state to the second state, and the value of the sixth time offset is positive, that is, the time period corresponding to the second state of the first model or function is included in the time period corresponding to the first state outside the time.

[0390] The first state of the first model or function lasts until the third time point, which is one of the following: the end time point of the first state of the first model or function; or the first state of the first model or function ends at the third time point; or the terminal device starts to convert the first state of the first model or function to the second state at the third time point; or the terminal device can determine the first state of the first model or function according to the third time point, and the first model or function is in the first state before the third time point. In an optional implementation, the first model or function is in the second state after the third time point, or the second state of the first model or function starts at the third time point, or the third time point is the start time point of the second state of the first model or function. In another optional implementation, the third time point is a time point after the third time point is delayed by the sixth time offset, and the first model or function is in the second state, that is, the time period between the third time point and the time point after the third time point is delayed by the sixth time offset is the time period during which the terminal device switches the first model or function from the first state to the second state, that is, the time period corresponding to the second state of the first model or function is outside the time period corresponding to the first state.

[0391] The sixth time offset is reported by the terminal device or is preset. In a possible implementation, the sixth time offset corresponds to the time delay during which the first model or function is switched from the first state to the second state.

[0392] For example, FIG. 16 is another CSI reporting schematic diagram. As shown in FIG. 16, the third time point corresponding to the first task is t5, the second time point corresponding to the second task is t6, and the terminal device determines that the activated state of the first model or function lasts until t5. Then the terminal device starts to deactivate the first model or function at t5, that is, t5 is the end time point of the first state of the first model or function, and the first model or function is in the deactivated state after t5, or the first model or function is in the deactivated state at a time point after t5 is delayed by a period of time.

[0393] It can be seen that when the duration of the first state of the first model or function corresponding to the first task overlaps with the duration of the first state of the second model or function corresponding to the second task, the terminal device can determine the duration time point of the first state of the first model or function according to one of the following: the priority of the task, the start time point of the task, the identifier of the task, and the time characteristic of the task. Therefore, the terminal device can convert the first state of the first model or function to the second state starting from the duration time point of the first state of the first model or function. The terminal device converts the first state of the first model or function to the second state at the duration time point of the first model or function, or converts the first state of the first model or function to the second state within a period of time after the duration time point of the first model or function.

[0394] It can be understood that, whether the first state of the first model or function lasts to the second moment or lasts to the third moment, the second moment is used to determine the state of the first model or function. Or, whether the first state of the first model or function lasts to the second moment or lasts to the third moment, the terminal device determines the state of the first model or function according to the second moment and the starting moment of the first task.

[0395] Optionally, when there is an overlap between the duration period of the first state of the first model or function corresponding to the first task and the duration period of the first state of the second model or function corresponding to the second task, the network device can also determine the duration moment of the first state of the first model or function according to one of the following: the priority of the task, the starting moment of the task, the identifier of the task, and the time characteristic of the task. Thus, the network device and the terminal device can align the duration moment of the first model or function, and the terminal device can successfully use the first model or function in the duration period of the first state of the first model or function.

[0396] In an optional implementation, when the first function or model corresponding to the first task lasts from the starting moment of the first task to the third moment, the third model or function corresponding to the third task is in the first state, including that the starting moment of the third task is located between the starting moment of the first task and the third moment. Or, the terminal device can determine the state of the third model or function corresponding to the third task according to the starting moment of the first task, the third moment and the starting moment of the third task; and the terminal device determines that the third model or function is in the first state when the starting moment of the third task is located between the starting moment of the first task and the third moment.

[0397] Optionally, since the first model or function is in the first state, at least one of the following corresponding to the first task is different from that when the first model or function is in the second state: the latency requirement, the use time period of the first model or function, the storage unit occupation time period of the first model or function, and the calculation unit occupation time period of the first model or function. Therefore, the network device can also determine the state of the first model or function corresponding to the first task before instructing the terminal device to execute the first task, or before sending the first information, and determine at least one of the following according to the state of the first model or function: the latency requirement, the use time period of the first model or function, the storage unit occupation time period of the first model or function, and the calculation unit occupation time period of the first model or function. In turn, the network device can schedule related resources for the terminal device to execute the first task according to the determined at least one. For example, the network device schedules resources of a second downlink signal for the terminal device to execute the first task according to the determined latency requirement, or configures resources of a second uplink signal for the terminal device to carry the reported information.

[0398] The network device also determines the state of the first model or function according to the second time and the start time of the first task. For details, refer to the above description of the terminal device determining the state of the first model or function according to the second time and the start time of the first task, which will not be repeated here.

[0399] It can be seen that the network device can also determine the state of the first model or function corresponding to the first task, so that the network device and the terminal device can align the state of the first model or function, and then the network device and the terminal device can both determine the requirement corresponding to the first task according to the state of the first model or function, which can ensure that the terminal device can successfully execute the first task.

[0400] In the embodiments of the present application, when the first model or function corresponding to the first task is in the first state, at least one of the following items corresponding to the first task is different from that when the first model or function is in the second state: latency requirement, usage time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function. Therefore, when the terminal device receives the first information for indicating execution of the first task, the terminal device determines the state of the first model or function, and determines at least one of the following items corresponding to the first task according to the state of the first model or function: latency requirement, usage time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function, and then executes the first task according to the determined at least one item. This way can make the requirement corresponding to the first task match the state of the first model or function corresponding to the first task, so that the terminal device can successfully execute the first task, or in other words, the model inference of the first model or function is successfully executed.

[0401] The embodiments of the present application also take the second task as a periodically scheduled CSI measurement and reporting task, the first task as a dynamically scheduled CSI measurement and reporting task (i.e., a non-periodically scheduled CSI measurement and reporting task), the start time of the first task is later than the start time of the second task, the model or function used by the terminal device to execute the first task is the first model or function, and the model or function used by the terminal device to execute the second task is the second model or function as an example, and illustrate the communication method shown in FIG. 6. FIG. 17 is an interaction diagram of a network device and a terminal device. As shown in FIG. 17, the interaction between the network device and the terminal device includes but is not limited to the following steps:

[0402] S1701. The network device sends CSI reporting trigger information to the terminal device, and the CSI reporting trigger information is used to trigger the terminal device to execute the first task. Correspondingly, the terminal device receives the CSI reporting trigger information from the network device.

[0403] The first task belongs to aperiodic beam prediction CSI measurement and reporting. The CSI reporting trigger information further includes a resource and a reporting quantity for the terminal device to perform aperiodic beam prediction CSI measurement and reporting.

[0404] In the embodiments of the application, the first task is a task executed by running a first model or function, or in other words, the first task needs to use the first model or function for model inference. The first model or function can directly perform model inference when it is in the first state (activated state). When the first model or function is in the second state (deactivated state), it needs to be activated first, so that the model or function is switched to the first state (activated state), and then model inference is performed. Therefore, the CSI reporting trigger information is used to trigger the terminal device to perform the first task, which can be understood as an implicit instruction for the terminal device to perform activation and inference of the first model or function corresponding to the first task. That is, the network device can indicate the activation of the first model or function through the CSI reporting trigger information, and the terminal device determines whether to perform the actual model activation operation according to the state of the first model or function.

[0405] S1702. The terminal device determines the state of the first model or function corresponding to the first task.

[0406] In an optional embodiment, the terminal device determines the state of the first model or function, including: determining a second time corresponding to the second task, the second time being a duration of the first state of the second model or function; and determining the state of the first model or function according to the second time and a starting time of the first task.

[0407] The terminal device determines the second time corresponding to the second task, including: determining the second time according to the first time and the first time length, and the implementation of which can be referred to the specific description of the communication method in FIG. 6 described above, and will not be repeated. Optionally, the first time and the first time length are configured by the network device to the terminal device through the CSI reporting trigger information, or are pre-defined by a protocol.

[0408] Optionally, the first time instant corresponds to one of: a starting time instant or an ending time instant of a reference signal used for measuring the CSI; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the reference signal used for measuring the CSI is located; a starting time instant of a first time unit after a time unit in which the starting time instant or the ending time instant of the reference signal used for measuring the CSI is located; a starting time instant or an ending time instant of a PUSCH or a PUCCH carrying the report of the CSI; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the report of the CSI is located; a starting time instant of a first time unit after a time unit in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the report of the CSI is located; a starting time instant or an ending time instant of a corresponding reference resource of the CSI reporting; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the corresponding reference resource of the CSI reporting is located; a starting time instant of a first time unit after a time unit in which the starting time instant or the ending time instant of the corresponding reference resource of the CSI reporting is located; a starting time instant or an ending time instant of a PDCCH triggering the CSI reporting; a starting time instant or an ending time instant of a time unit in which the starting time instant or the ending time instant of the PDCCH triggering the CSI reporting is located; a starting time instant of a first time unit after a time unit in which the starting time instant or the ending time instant of the PDCCH triggering the CSI reporting is located; a time instant in which the starting time instant or the ending time instant of the reference signal used for measuring the CSI is advanced or delayed by a time offset; a time instant in which the starting time instant or the ending time instant of the time unit in which the reference signal used for measuring the CSI is located is advanced or delayed by the time offset; a time instant in which the starting time instant of the first time unit after the time unit in which the reference signal used for measuring the CSI is located is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the PUSCH or the PUCCH carrying the report of the CSI is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the time unit in which the PUSCH or the PUCCH carrying the report of the CSI is located is advanced or delayed by the time offset; a time instant in which the starting time instant of the first time unit after the time unit in which the PUSCH or the PUCCH carrying the report of the CSI is located is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the corresponding reference resource of the CSI reporting is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the time unit in which the corresponding reference resource of the CSI reporting is located is advanced or delayed by the time offset; a time instant in which the starting time instant of the first time unit after the time unit in which the corresponding reference resource of the CSI reporting is located is advanced or delayed by the time offset; a time instant in which the starting time instant or the ending time instant of the PDCCH triggering the CSI reporting is advanced or delayed by the time offset;A time unit in which the starting moment or ending moment of the PDCCH triggering the CSI reporting is located, a starting moment or ending moment of a time unit after the starting moment or ending moment of the PDCCH triggering the CSI reporting is advanced or delayed by a time offset, or a starting moment of a first time unit after a time unit in which the starting moment or ending moment of the PDCCH triggering the CSI reporting is located is advanced or delayed by a time offset. The reference signal for measuring the CSI is an RS or an SSB.

[0409] Optionally, the first moment corresponds to a starting moment of a first period in the N periods, and the second moment corresponds to an ending moment of an Nth period in the N periods. The N periods are a plurality of periods to which the second task belongs, the second task is a task corresponding to one period in the plurality of periods, and N is a positive integer greater than 1.

[0410] In addition, the implementation of determining the state of the first model or function according to the second moment and the starting moment of the first task by the terminal device can refer to the specific description in the communication method described in FIG. 6, and will not be repeated here.

[0411] S1703. The terminal device determines at least one of the following according to the state of the first model or function: a time delay requirement, a use time period of the first model or function, a storage unit occupation time period of the first model or function, and a calculation unit occupation time period of the first model or function.

[0412] The implementation of S1703 can refer to the above-mentioned embodiments 1 to 5, and will not be repeated here.

[0413] S1704. The network device sends the CSI RS to the terminal device. Correspondingly, the terminal device receives the CSI RS from the network device.

[0414] The CSI RS is used for the terminal device to perform CSI measurement for beam prediction.

[0415] In an optional implementation, in S1703, the terminal device determines the latency requirement, and the terminal device receiving the CSI RS from the network device includes: receiving the CSI RS from the network device based on the determined latency requirement. For example, the terminal device receives the CSI RS at t6 based on the latency requirement between the CSI RS corresponding to the first task in FIG. 12 and the PUSCH. Alternatively, the terminal device determines to use the CSI RS received at t6 to perform the CSI measurement based on the latency requirement between the CSI RS corresponding to the first task in FIG. 12 and the PUSCH. For another example, the terminal device receives the CSI RS at t7 based on the latency requirement between the CSI RS corresponding to the first task in FIG. 13 and the PUSCH. Alternatively, the terminal device determines to use the CSI RS received at t7 to perform the CSI measurement based on the latency requirement between the CSI RS corresponding to the first task in FIG. 13 and the PUSCH. For another example, the terminal device receives the CSI RS at t7 based on the latency requirement between the CSI RS corresponding to the first task in FIG. 14a and the PUSCH. Alternatively, the terminal device determines to use the CSI RS received at t7 to perform the CSI measurement based on the latency requirement between the CSI RS corresponding to the first task in FIG. 14a and the PUSCH.

[0416] It can be understood that the CSI RS is transmitted according to the configuration of the CSI RS, and therefore the configuration and transmission of the CSI RS are two independent processes from the CSI reporting configuration / triggering command. Therefore, the execution order of S1704 can be before or after the execution order of S1701. Alternatively, the CSI RS used by the terminal device to perform the CSI measurement can be the CSI RS transmitted before the execution of S1701 and / or the CSI RS transmitted after the execution of S1701.

[0417] S1705a. The terminal device determines the optimal beam and the RSRP of the optimal beam based on the CSI RS and the first model or function when the first model or function is in an activated state.

[0418] In the method, the terminal device determines the optimal beam and the RSRP of the optimal beam based on the CSI RS and the first model or function, including: obtaining a first beam and a RSRP of the first beam based on the CSI RS; and performing model inference on the first beam and the RSRP of the first beam using the first model or function within a time period corresponding to the first model or function to obtain the optimal beam and the RSRP of the optimal beam. The time period corresponding to the first model or function includes at least one of the following determined by the terminal device: a use time period of the first model or function, a storage unit occupation time period of the first model or function, and a calculation unit occupation time period of the first model or function.

[0419] For example, in S1703, the terminal device determines the use time period of the first model or function, and then in S1705a: the terminal device obtains the first beam and the RSRP of the first beam based on the CSI RS; and in the use time period of the first model or function, the terminal device performs model inference on the first beam and the RSRP of the first beam using the first model or function to obtain the optimal beam and the RSRP of the optimal beam.

[0420] S1705b. When the first model or function is in the deactivated state, the terminal device activates the first model or function, and determines the optimal beam and the RSRP of the optimal beam based on the CSI RS and the activated first model or function.

[0421] In the method, when the first model or function is in the deactivated state, the terminal device activates the first model or function, and determines the optimal beam and the RSRP of the optimal beam based on the CSI RS and the activated first model or function, including: activating the first model or function based on the time period corresponding to the first model or function, and determining the optimal beam and the RSRP of the optimal beam based on the CSI RS and the activated first model or function. The time period corresponding to the first model or function includes at least one of the following determined by the terminal device: the use time period of the first model or function, the storage unit occupation time period of the first model or function, and the calculation unit occupation time period of the first model or function.

[0422] For example, in S1703, the terminal device determines the use time period of the first model or function, and the starting time of the use time period of the first model or function is later than the starting time of the CSI RS, and then in S1705b: after receiving the CSI RS, the terminal device activates the first model or function, and performs model inference on the first beam and the RSRP of the first beam using the first model or function in the use time period of the first model or function to obtain the optimal beam and the RSRP of the optimal beam.

[0423] S1706. The terminal device sends a CSI report to the network device, and the CSI report includes the optimal beam and the RSRP of the optimal beam.

[0424] Optionally, before the network device sends the CSI reporting trigger information to the terminal device, the network device also determines the state of the first model or function corresponding to the first task, so as to align the state of the first model or function with the terminal device. The implementation of the network device to determine the state of the first model or function can refer to the implementation of the terminal device to determine the state of the first model or function, which will not be described herein.

[0425] Optionally, the network device further determines, according to the state of the first model or function, at least one of the following: a latency requirement, a usage time period of the first model or function, a storage unit occupation time period of the first model or function, a computing unit occupation time period of the first model or function, and then configures, for the terminal device, a related requirement corresponding to the first task according to the determined at least one. For example, the terminal device determines the latency requirement, and then schedules, for the terminal device, a resource for receiving the CSI RS according to the latency requirement, or schedules, for the terminal device, a resource for carrying the PUSCH for the CSI report according to the latency requirement. In this way, the network device also determines at least one requirement corresponding to the first task according to the state of the first model or function. The at least one requirement corresponding to the first task determined by the network device and the at least one requirement corresponding to the first task determined by the terminal device are both determined according to the state of the first model or function, and thus are aligned, so that the terminal device can successfully perform the first task, or the resources for the terminal device to perform the first task can be saved.

[0426] It can be seen that after the terminal device receives the CSI reporting trigger information for triggering the terminal device to perform the first task, the terminal device determines the state of the first model or function corresponding to the first task, and determines, according to the state of the first model or function, at least one of the following corresponding to the first task: a latency requirement, a usage time period of the first model or function, a storage unit occupation time period of the first model or function, and a computing unit occupation time period of the first model or function, and then performs beam prediction using the activated first model or function according to the determined at least one, to obtain an optimal beam and an RSRP of the optimal beam, and then reports the optimal beam and the RSRP of the optimal beam through the CSI report.

[0427] Embodiments of the present application also propose another communication method. This communication method is also described from the perspective of interaction between the network device and the terminal device. FIG. 18 is an interaction diagram of the communication method. As shown in FIG. 18, the communication method includes but is not limited to the following steps:

[0428] S1801. The network device sends first information to the terminal device, the first information being used to instruct the terminal device to perform a first task, a computing unit occupation time period of a first model or function corresponding to the first task starting at a fourth time, the fourth time being later than a starting time of a second downlink signal corresponding to the first task, or the fourth time being earlier than the starting time of the second downlink signal corresponding to the first task. Correspondingly, the terminal device receives the first information from the network device.

[0429] S1802. The terminal device performs the first task.

[0430] The first information can be referred to as described in S601 above, and will not be described again.

[0431] Optionally, the time period occupied by the computing unit of the first model or function corresponding to the first task can be replaced by: a time period of use of the first model or function corresponding to the first task.

[0432] In addition, the time period occupied by the computing unit of the first model or function corresponding to the first task starts at the fourth time point, and the fourth time point is later than the starting time point of the second downlink signal corresponding to the first task. The second downlink signal is a downlink signal or a downlink channel. Compared with the fourth time point starting from the starting time point of the second downlink signal, this mode can reduce the time period occupied by the computing unit of the model, so that the terminal device can have more idle computing units to perform tasks using other models. For example, as shown in FIG. 13, the time period occupied by the computing unit of the first model or function starts at t10, instead of t7 or t8, thereby saving the time period occupied by the computing unit of the first model or function. Therefore, during the time period from t5 to t10, there can be idle computing units, and the terminal device can use other models or functions to perform model inference.

[0433] For example, FIG. 19 is another schematic diagram of CSI reporting. As shown in FIG. 19, t1 is a time point at which the terminal device is configured to perform periodic CSI reporting, and t3 is a time point at which the terminal device is triggered to perform aperiodic CSI reporting. The periodic CSI reporting task triggered at t1 is the first task, and the aperiodic CSI reporting task triggered at t3 is the second task. The terminal device is allowed to use at most 2 CPUs, the time period occupied by the computing unit of the first model or function is t7 to t8, and 2 computing units CPUs are occupied, so that there is no idle CPU during the time period from t7 to t8. The aperiodic CSI reporting task triggered at t3 needs to occupy 1 CPU, and since the time period occupied by the computing unit of the first model or function corresponding to the first task is t7 to t8, there are 2 idle CPUs from t3 to t7. Therefore, the terminal device can use the idle CPUs to perform the second task, that is, the terminal device can use the second model or function corresponding to the second task to perform model inference during the time period from t5 to t6, and obtain the inference result.

[0434] For example, FIG. 20 is another schematic diagram of CSI reporting. As shown in FIG. 20, t1 is a time point at which the terminal device is configured to perform periodic CSI reporting, t3 is a time point at which the terminal device is triggered to perform aperiodic CSI reporting, the periodic CSI reporting triggered at t1 is a first task, and the aperiodic CSI reporting triggered at t3 is a second task. The terminal device is allowed to use at most 2 CPUs, the time period in which the calculation unit of the first model or function occupies is t2-t4, and 1 CPU is occupied, so there is 1 idle CPU in the time period t2-t4. The second task triggered at t3 needs to occupy 2 CPUs, but the CPU occupation time period of the second task is t5-t6, that is, the CPU for executing the second task starts to be occupied from t5, and there are 2 idle CPUs between t5 and t6, so the terminal device can first perform model activation after receiving the trigger request of the second task, and then can use the second model or function corresponding to the second task to perform model inference between t5 and t6 to obtain an inference result.

[0435] Optionally, the fourth time point is earlier than a starting time point of a second downlink signal corresponding to the first task, and the second downlink signal is a downlink signal. When the fourth time point is earlier than the starting time point of the second downlink signal corresponding to the first task, it can be ensured that the terminal device can use sufficient resources to perform model inference based on the activated first model or function to obtain an inference result. For example, in FIG. 15a, if the time period in which the calculation unit of the first model or function occupies starts from t8, that is, the calculation resources are reserved for the first task at t8, the terminal device can immediately use sufficient resources between t8 and t10 to successfully perform model inference based on the activated first model or function after the first model or function is switched from the activated state to the deactivated state, and complete the execution of the first task.

[0436] In addition, a time interval between the fourth time point and the starting time point of the second downlink signal is a second time length, and the second time length is reported by the terminal device or preset.

[0437] Optionally, the fourth time point is earlier than a starting time point or an ending time point of a second uplink signal corresponding to the first task, and a time interval between the fourth time point and the starting time point or the ending time point of the second uplink signal is a third time length. The third time length is reported by the terminal device or preset. For example, in FIG. 19, the starting time point t7 of the time period in which the calculation unit of the first model or function corresponding to the first task occupies is earlier than the starting time point t8 of the PUSCH carrying the CSI report or earlier than the ending time point t9 of the PUSCH carrying the CSI report.

[0438] Optionally, the first model or function computing unit occupation time period ends at a fifth moment, and the fifth moment is earlier than an ending moment of the second uplink signal corresponding to the first task. For example, in FIG. 19, the first model or function computing unit occupation time period ends at t8, and t8 is earlier than the ending moment t9 of the PUSCH carrying the CSI report.

[0439] Optionally, a time interval between the fifth moment and a starting or ending moment of the second downlink signal is a fourth time length. The fourth time length is reported by the terminal device or is preset. For example, in FIG. 19, there is a time interval between the ending moment t8 of the first model or function computing unit occupation time period and the CSI RS corresponding to the first task.

[0440] Optionally, a time interval between the fifth moment and a starting or ending moment of the second uplink signal is a fifth time length. The fifth time length is reported by the terminal device or is preset. For example, in FIG. 19, there is a time interval between the ending moment t8 of the first model or function computing unit occupation time period and the ending moment t9 of the PUSCH carrying the CSI report corresponding to the first task.

[0441] Optionally, the terminal device performs the first task, including: performing the first task based on the first model or function computing unit occupation time period corresponding to the first task. Specifically, the terminal device performs model inference using the first model or function in the first model or function computing unit occupation time period, and obtains an inference result. Optionally, the terminal device further reports the inference result or report information generated according to the inference result to the network device.

[0442] It can be seen that, in the embodiments of the present application, the first model or function computing unit occupation time period corresponding to the first task starts at a fourth moment, and the fourth moment is later than a starting moment of the second downlink signal corresponding to the first task, or the fourth moment is earlier than the starting moment of the second downlink signal corresponding to the first task. Then, after the terminal device receives the first information used for indicating to perform the first task, the terminal device determines the first model or function computing unit occupation time period, and performs the first task according to the determined first model or function computing unit occupation time period.

[0443] The time period occupied by the first model or the function of the first task starts at the fourth moment, and the fourth moment is later than the starting moment of the second downlink signal corresponding to the first task. The terminal device performs the first task according to the determined time period occupied by the first model or the function, so that more computing units are idle, thereby the terminal device can perform more tasks by using the idle computing units, or in other words, the terminal device can fully utilize the idle computing unit resources to perform multiple parallel tasks. When the fourth moment is earlier than the starting moment of the second downlink signal corresponding to the first task, the terminal device performs the first task according to the determined time period occupied by the first model or the function, so that when the first model or the function is in the first state, the terminal device can use sufficient resources to perform model inference based on the first model or the function, and then the terminal device can successfully perform the first task.

[0444] For the technical solutions described above, the corresponding device implementation solutions are further described below.

[0445] To implement the functions in the methods provided in the embodiments of the present application, the terminal device and the network device can include hardware structures and / or software modules to implement the above functions in the form of hardware structure, software module, or hardware structure plus software module. Whether a certain function in the above functions is implemented in the form of hardware structure, software module, or hardware structure plus software module depends on the specific application and design constraints of the technical solutions.

[0446] As shown in FIG. 21, the embodiment of the present application provides a communication device 2100. The communication device 2100 can be a component (for example, an integrated circuit, a chip, etc.) of a terminal device, and can also be a component (for example, an integrated circuit, a chip, etc.) of a network device. The communication device 2100 can also be other communication units for implementing the methods in the method embodiments of the present application. The communication device 2100 can include a communication unit 2101 and a processing unit 2102. Optionally, it can also include a storage unit 2103.

[0447] In a possible design, one or more units in FIG. 21 can be implemented by one or more processors, or by one or more processors and memories; or by one or more processors and transceivers; or by one or more processors, memories, and transceivers, and the embodiments of the present application do not make any limitations in this regard. The processor, memory, transceiver can be separately arranged, or integrated.

[0448] The communication apparatus 2100 has the functions of the terminal device or the network device described in the embodiments of the present application. For example, the communication apparatus 2100 includes a module or unit or means corresponding to the terminal device involved steps in each method embodiment described above, and the functions or units or means can be implemented by software, or by hardware, or by a combination of hardware and software. For details, further reference can be made to the corresponding description in the foregoing method embodiments.

[0449] In a possible design, the communication apparatus 2100 can include a processing unit 2102 and a communication unit 2101, and the apparatus is applied to a terminal device.

[0450] The communication unit 2101 is configured to receive first information, where the first information is used to instruct the terminal device to perform a first task.

[0451] The processing unit 2102 is configured to perform the first task.

[0452] The first model or function is in a first state, and at least one of the following is different from that when the first model or function is in a second state: a latency requirement, a usage time period of the first model or function, a storage unit occupation time period of the first model or function, and a computing unit occupation time period of the first model or function, where the first model or function corresponds to the first task.

[0453] In an optional embodiment, the first state of the second model or function lasts from a first time point to a second time point, and the second model or function corresponds to a second task, where a start time point of the second task is earlier than a start time point of the first task.

[0454] In an optional implementation, the first time point corresponds to one of: a start time point of the first downlink signal, an end time point of the first downlink signal, a start time point of the first uplink signal, an end time point of the first uplink signal, a first offset time point, and a second offset time point; the time interval between the first offset time point and the start time point or the end time point of the first downlink signal corresponds to a first time offset, and the time interval between the second offset time point and the start time point or the end time point of the first uplink signal corresponds to a second time offset; or the first time point corresponds to one of: a start time point of a usage time period of the second model or function corresponding to the second task, a start time point of a storage unit occupation time period of the second model or function, a start time point of a calculation unit occupation time period of the second model or function, an end time point of the usage time period of the second model or function, an end time point of the storage unit occupation time period of the second model or function, and an end time point of the calculation unit occupation time period of the second model or function; the first downlink signal and the first uplink signal are both related to the second task.

[0455] In an optional implementation, a time interval between the first time point and the second time point is a first time length, and the first time length is configured or indicated by the network device or is preset.

[0456] In an optional implementation, the second task belongs to a periodic task or a semi-persistent task.

[0457] In an optional implementation, when the second task belongs to a periodic task or a semi-persistent task, the first time point corresponds to a start time point of a first period in N periods, and the second time point corresponds to an end time point of an Nth period in the N periods; the N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1.

[0458] In an optional implementation, the second time point corresponds to one of: a start time point or an end time point of a channel carrying the reported information of the second task, a start time point or an end time point of a time unit in which the start time point or the end time point of the channel carrying the reported information of the second task is located, and a start time point of a first time unit after the time unit in which the start time point or the end time point of the channel carrying the reported information of the second task is located.

[0459] In an optional implementation, the second task belongs to a dynamically scheduled task.

[0460] In an optional implementation, the first model or function corresponding to the first task is in a first state, including that a start time point of the first task is located between the first time point and the second time point.

[0461] In an optional implementation, the first state of the first model or function lasts to a third time point, the third time point corresponding to the first task, and the third time point being later than the second time point.

[0462] In an optional implementation, the third time point corresponds to at least one of the following for a third task: a latency requirement, a usage time period of a third model or function, a storage unit occupation time period of the third model or function, and a computing unit occupation time period of the third model or function; the third model or function corresponds to the third task, and a start time point of the third task is later than a start time point of the first task.

[0463] In an optional implementation, the first state of the first model or function ends at the second time point.

[0464] In an optional implementation, the second state of the first model or function starts from the second time point, or a time period corresponding to the second state is included in a time period other than a time period corresponding to the first state.

[0465] In an optional implementation, the first model or function is in a second state, including that the first model or function is not in the first state, or a start time point of the first task is after the second time point.

[0466] In an optional implementation, the start time point of the first task corresponds to one of the following: a start time point of a second downlink signal, an end time point of the second downlink signal, a third offset time point, a start time point of a usage time period of the first model or function, a start time point of a storage unit occupation time period of the first model or function, and a start time point of a computing unit occupation time period of the first model or function; the third offset time point corresponds to a third time offset between a time interval between the start time point or the end time point of the second downlink signal; and the second downlink signal is related to the first task.

[0467] In an optional implementation, the second model or function is the same as the first model or function.

[0468] In an optional implementation, the time delay requirement is one of the following: a time interval requirement between the second downlink signal and the second uplink signal; a time interval requirement between the second downlink signal and an end time of a time period during which the first model or function is used; a time interval requirement between the second downlink signal and an end time of a time period during which a storage unit of the first model or function is occupied; a time interval requirement between the second downlink signal and an end time of a time period during which a calculation unit of the first model or function is occupied; and the second downlink signal and the second uplink signal are both related to the first task.

[0469] In an optional implementation, when the first model or function is in the first state, the first task corresponds to a smaller time delay requirement than when the first model or function is in the second state.

[0470] In an optional implementation, when the first model or function is in the first state, a time period during which the first model or function is used corresponding to the first task starts from a start time or an end time of a second downlink signal corresponding to the first task; and when the first model or function is in the second state, a start time of the time period during which the first model or function is used corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task.

[0471] In an optional implementation, when the first model or function is in the first state, a time period during which a calculation unit of the first model or function is occupied corresponding to the first task starts from a start time or an end time of a second downlink signal corresponding to the first task; and when the first model or function is in the second state, a start time of the time period during which the calculation unit of the first model or function is occupied corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task.

[0472] In an optional implementation, when the first model or function corresponding to the first task is in the first state, the storage unit occupation time period of the first model or function corresponding to the first task starts from the start or end time of the second downlink signal corresponding to the first task; when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or end time of the second downlink signal corresponding to the first task; or, when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is later than the start time or end time of the second downlink channel corresponding to the first task; or, when the first model or function corresponding to the first task is in the second state, the start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or end time of the second uplink signal corresponding to the first task, and the start time of the storage unit occupation time period of the first model or function corresponding to the first task is determined by the start time or end time of the second uplink signal corresponding to the first task.

[0473] In an optional implementation, the first state is an active state, and the second state is a deactivation state.

[0474] In an optional implementation, the first time is configured by a network device or is preset.

[0475] In an optional implementation, the first time length is determined according to at least one of the following: an index or an identifier corresponding to the first time length; 1 / M times of a reporting period of the second task, where M is a positive integer; and a time characteristic of the second task.

[0476] In an optional implementation, the first time offset and / or the second time offset are reported by the terminal device or are preset.

[0477] In an optional implementation, the third time offset is reported by the terminal device or is preset.

[0478] Embodiments of the present application and the above-described method embodiments are based on the same concept and bring the same technical effects. For specific principles, refer to the description of the above-described embodiments, which will not be repeated here.

[0479] The embodiment of the present application further provides a communication device 2200, and FIG. 22 is a structural schematic diagram of the communication device 2200. The communication device 2200 can be a terminal device, or can be a chip, a chip system, or a processor supporting the terminal device to implement the above method; or can be a network device, or can be a chip, a chip system, or a processor supporting the network device to implement the above method. The device can be used to implement the method described in the above method embodiment, and details can be referred to the description in the above method embodiment.

[0480] The communication device 2200 can include one or more processors 2201. The processor 2201 can be a general-purpose processor or a special-purpose processor. For example, it can be a baseband processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, or a central processing unit (CPU). The baseband processor can be used to process a communication protocol and communication data, and the central processing unit can be used to control the communication device (such as a base station, a baseband chip, a terminal, a terminal chip, a DU or a CU, etc.), execute a software program, and process data of the software program.

[0481] Optionally, the communication device 2200 can include one or more memories 2202, and the memories 2202 can have instructions 2204 stored thereon. The instructions can be run on the processor 2201, so that the communication device 2200 performs the method described in the above method embodiment. Optionally, the memory 2202 can also store data. The processor 2201 and the memory 2202 can be separately arranged, or can be...

Claims

1. A communication method characterized by comprising: The method comprises: receiving first information, the first information being used to instruct a terminal device to perform a first task; performing the first task; at least one of the following corresponding to the first task is different between when a first model or function is in a first state and when the first model or function is in a second state: latency requirement, usage time period of the first model or function, storage unit occupation time period of the first model or function, and calculation unit occupation time period of the first model or function; the first model or function corresponds to the first task.

2. The method of claim 1, wherein, the first model or function corresponding to the first task is in the first state, comprising: a starting moment of the first task is between a first moment and a second moment.

3. The method of claim 1 or 2, wherein: the first state of the second model or function lasts from the first moment to the second moment; the second model or function corresponds to a second task, and a starting moment of the second task is earlier than a starting moment of the first task.

4. The method of claim 3, wherein: the first moment corresponds to one of the following: a starting moment of a first downlink signal, an ending moment of the first downlink signal, a starting moment of a first uplink signal, an ending moment of the first uplink signal, a first offset moment, and a second offset moment; wherein the first offset moment corresponds to a first time offset between the starting moment or the ending moment of the first downlink signal, and the second offset moment corresponds to a second time offset between the starting moment or the ending moment of the first uplink signal; or, the first moment corresponds to one of the following: a starting moment of a usage time period of a second model or function corresponding to the second task, a starting moment of a storage unit occupation time period of the second model or function, a starting moment of a calculation unit occupation time period of the second model or function, an ending moment of the usage time period of the second model or function, an ending moment of the storage unit occupation time period of the second model or function, and an ending moment of the calculation unit occupation time period of the second model or function; the first downlink signal and the first uplink signal are both related to the second task.

5. The method of any one of claims 2 to 4, wherein: a time interval between the first moment and the second moment is a first time length, and the first time length is configured or instructed by a network device or is preset.

6. The method of claim 3 or 4, wherein: the second task belongs to a periodic task or a semi-persistent task.

7. The method of claim 3, wherein: when the second task belongs to a periodic task or a semi-persistent task, the first moment corresponds to a starting moment of a first period in N periods, and the second moment corresponds to an ending moment of an Nth period in the N periods; the N periods are a plurality of periods to which the second task belongs, and N is a positive integer greater than 1.

8. The method of claim 3 or 4, wherein, the second moment corresponds to one of the following: a starting moment or an ending moment of a channel carrying reported information of the second task. a start time or an end time of a time unit in which a start time or an end time of a channel carrying the reporting information of the second task is located; a start time of a first time unit after the time unit in which the start time or the end time of the channel carrying the reporting information of the second task is located. 9.The method of claim 8, wherein the second task belongs to a dynamically scheduled task. 10.The method of any one of claims 2 to 9, wherein: the first state of the first model or function lasts to a third time, the third time corresponding to the first task, the third time being later than the second time.

11. The method of claim 10, wherein, the third time acts on at least one of the following corresponding to a third task: a latency requirement, a usage time period of a third model or function, a storage unit occupation time period of the third model or function, a computing unit occupation time period of the third model or function; the third model or function corresponds to the third task, a start time of the third task being later than a start time of the first task.

12. The method according to any one of claims 2 to 9, characterized in that, the first state of the first model or function ends at the second time. 13.The method of any one of claims 2 to 9, or claim 12, wherein: the second state of the first model or function starts from the second time, or a time period corresponding to the second state of the first model or function is contained in a time outside a time period corresponding to the first state.

14. The method according to any one of claims 1 to 13, characterized in that, the first model or function is in the second state, including: the first model or function is not in the first state; or a start time of the first task is after the second time.

15. The method according to any one of claims 2 to 11, characterized in that, including: the start time of the first task corresponds to one of the following: a start time of a second downlink signal, an end time of the second downlink signal, a third offset time, a start time of a usage time period of the first model or function, a start time of a storage unit occupation time period of the first model or function, a start time of a computing unit occupation time period of the first model or function; wherein the third offset time corresponds to a third time offset between the start time or the end time of the second downlink signal; the second downlink signal is related to the first task. 16.The method of any one of claims 2 to 15, wherein: the second model or function is the same as the first model or function.

17. The method according to any one of claims 1 to 16, characterized in that, the latency requirement is one of the following: a time interval requirement between a second downlink signal and a second uplink signal; a time interval requirement between the second downlink signal and an end time of a usage time period of the first model or function; a time interval requirement between the second downlink signal and an end time of a storage unit occupation time period of the first model or function; a time interval requirement between the second downlink signal and an end time of a computing unit occupation time period of the first model or function; the second downlink signal and the second uplink signal are both related to the first task. 18.The method of any one of claims 1 to 17, wherein: The first task corresponds to a smaller latency requirement when the first model or function is in the first state than when the first model or function is in the second state.

19. The method of any one of claims 1-18, wherein, The usage time period of the first model or function corresponding to the first task starts from a start time or an end time of the second downlink signal corresponding to the first task when the first model or function is in the first state. The start time of the usage time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task when the first model or function is in the second state.

20. The method of any one of claims 1-19, wherein, The computation unit occupation time period of the first model or function corresponding to the first task starts from a start time or an end time of the second downlink signal corresponding to the first task when the first model or function is in the first state. The start time of the computation unit occupation time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task when the first model or function is in the second state.

21. The method of any one of claims 1-20, wherein, The storage unit occupation time period of the first model or function corresponding to the first task starts from a start time or an end time of the second downlink signal corresponding to the first task when the first model or function corresponding to the first task is in the first state. The start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second downlink signal corresponding to the first task when the first model or function corresponding to the first task is in the second state; or The start time of the storage unit occupation time period of the first model or function corresponding to the first task is later than the start time or the end time of the second downlink signal corresponding to the first task when the first model or function corresponding to the first task is in the second state; or The start time of the storage unit occupation time period of the first model or function corresponding to the first task is earlier than the start time or the end time of the second uplink signal corresponding to the first task when the first model or function corresponding to the first task is in the second state, and the start time of the storage unit occupation time period of the first model or function corresponding to the first task is determined by the start time or the end time of the second uplink signal corresponding to the first task.

22. The method of any one of claims 1-21, wherein, The first state is an active state, and the second state is a deactivation state.

23. The method of any one of claims 2 to 22, wherein, The first time is configured by a network device or is preset.

24. The method of claim 5, wherein, The first time length is determined according to at least one of the following: an index or an identifier corresponding to the first time length; and an index or an identifier corresponding to the first model or function. 1 / M times of a reporting period of the second task, where M is a positive integer; a time characteristic of the second task.

25. The method of claim 4, wherein, The first time offset and / or the second time offset are reported by the terminal device or are preset.

26. The method of claim 15, wherein, The third time offset is reported by the terminal device or is preset.

27. A communications device, characterized by The communication apparatus comprises modules for performing the method according to any one of claims 1 to 26.

28. A communications device, characterized by The communication apparatus comprises a processor configured to perform the method according to any one of claims 1 to 26.

29. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions which, when executed on a computer, cause the method according to any one of claims 1 to 26 to be performed.

30. A computer program product comprising instructions, wherein: The computer readable storage medium stores instructions which, when executed on a computer, cause the method according to any one of claims 1 to 26 to be performed.

Images