Communication methods and communication devices

Abstract

This application provides a communication method and communication device for use in scenarios where AI and wireless networks are combined, particularly in scenarios where an AI model used for positioning is combined with a wireless network. The method includes: a network element acquires a measurement result of a first parameter based on channel measurement results, and determines a specific AI model corresponding to the current measurement result of the first parameter based on the correspondence between the first parameter and the AI ​​model. The AI ​​model corresponding to the measurement result of the first parameter is compared with the AI ​​model currently used for positioning, and a decision is made whether to switch or update the AI ​​model used for positioning. Changes in the channel environment affect positioning accuracy. The measurement result of the first parameter is acquired based on channel measurement results, and the measurement result of the first parameter can reflect the current channel environment. In this way, by deciding whether to switch or update the AI ​​model based on the measurement result of the first parameter, it is possible to implement timely switching or updating of the AI ​​model based on changes in the channel environment, thereby improving the positioning accuracy of the AI ​​model.

Description

Technical Field

[0001] The present invention relates to communication technologies, and more specifically, to communication methods and communication devices.

Background Art

[0002] In modern society, the location information of people and things has become increasingly important. Positioning devices and technologies are used in fields such as navigation and aviation, plotting and disaster relief, military defense, vehicle navigation, logistics tracking, and traffic management. With the development of mobile communication, positioning technologies based on wireless cellular networks have been widely applied. In these positioning technologies, in order to estimate the position of a terminal, characteristic parameters between the terminal and a network device are detected, and relative position or angle information between the terminal and the network device is obtained. Artificial Intelligence (AI) models can be used in positioning technologies. Therefore, how to ensure the positioning accuracy of the AI model is an issue worthy of consideration.

Summary of the Invention

Problems to be Solved by the Invention

[0003] This application provides a communication method and a communication device that determine whether to switch or update an AI model used for positioning based on channel measurement results, and improve the positioning accuracy of the AI model.

Means for Solving the Problems

[0004] According to a first aspect, a communication method is provided. This method can be executed by a network element. Hereinafter, the first network element is used as an example for explanation.

[0005] The method may include: a first network element receiving a reference signal; and the first network element performing a channel measurement based on the reference signal and obtaining a channel measurement result, wherein the channel measurement result or auxiliary positioning information based on the channel measurement result is an input to an artificial intelligence (AI) model used for positioning, and the measurement result and correspondence of a first parameter are used to determine whether to switch or update the AI ​​model, the measurement result of the first parameter is obtained based on the channel measurement result, and the correspondence includes the correspondence between the AI ​​model used for positioning and the value of the first parameter.

[0006] Optionally, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal.

[0007] Optionally, the first network element is a network device or positioning device, or a component of a network device or positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0008] According to the above technical solution, the network element can determine whether it is necessary to switch or update the AI ​​model based on the input characteristics of the AI ​​model. The AI ​​model may be an AI model used for positioning. Specifically, if the input to the AI ​​model is the channel measurement result, the measurement result of the first parameter may be obtained based on the channel measurement result, and there is a correspondence between the AI ​​model and the value of the first parameter. In this way, the network element can determine a specific AI model corresponding to the measurement result of the first parameter based on the measurement result of the first parameter and the correspondence, and by comparing this specific AI model with the current AI model, it can determine whether it is necessary to switch or update the AI ​​model. It is thought that changes in the channel environment may affect positioning accuracy. For example, when a single AI model is used, changes in the channel environment may cause a deterioration in positioning performance. Since the measurement result of the first parameter is obtained based on the channel measurement result, the measurement result of the first parameter can reflect the current channel environment. In this way, by determining whether to switch or update the AI ​​model based on the measurement result of the first parameter, it is possible to implement switching or updating in a timely manner based on changes in the channel environment, thereby improving the positioning accuracy of the AI ​​model.

[0009] In relation to the first aspect, in some implementations of the first aspect, the method further includes: a first network element receiving first instruction information and / or correspondence, the first instruction information indicating that a first parameter is used to determine whether to switch or update the AI ​​model used for positioning.

[0010] In relation to the first aspect, in some implementations of the first aspect, the method further includes: a first network element determining whether to switch or update the AI ​​model based on the measurement results and correspondence of the first parameter.

[0011] According to the above technical solution, the first network element can determine whether to switch or update the AI ​​model based on the measurement results and correspondence of the first parameter. In this way, the first network element can monitor the state of the channel environment by using the channel measurement results and make a timely decision on whether to switch or update the AI ​​model based on the state of the channel environment.

[0012] In relation to the first aspect, in some implementations of the first aspect, the first network element determining whether to switch or update an AI model based on a measurement result and correspondence of a first parameter includes: the first network element determining whether to switch or update the first type of AI model used for positioning from a first AI model to a second AI model based on a measurement result and correspondence of a first parameter, wherein the channel measurement result is the input to the first type of AI model used for positioning, and the correspondence includes the correspondence between the first type of AI model used for positioning and the value of the first parameter; the method further includes: the first network element transmitting request information which is used to request a second AI model; the first network element receiving a response to the request information which includes information about the second AI model; and the first network element performing positioning by using the second AI model.

[0013] For example, request information being used to request a second AI model includes: request information being used to request an update to the AI ​​model. The updated AI model is the second AI model.

[0014] For example, information about the second AI model includes information that can identify the second AI model or can be used to obtain the second AI model. For example, information about the second AI model includes an identifier for the second AI model.

[0015] According to the above technical solution, when a decision is made to switch or update the AI ​​model, the first network element can request the switched or updated AI model from another network element (e.g., the network element that stores and / or trains the AI ​​model). In this way, positioning can be performed using the switched or updated AI model.

[0016] In relation to the first aspect, in some implementations of the first aspect, the first network element determining whether to switch or update an AI model based on a measurement result and correspondence of a first parameter includes: the first network element determining, based on a measurement result and correspondence of a first parameter, to switch or update a second type AI model used for positioning from a third AI model to a fourth AI model, wherein the channel measurement result is an input to the first type AI model used for positioning, the auxiliary positioning information is an output to the first type AI model, the first type AI model is switched or updated from a first AI model corresponding to a third AI model to a second AI model corresponding to a fourth AI model, the output of the first type AI model used for positioning is an input to the second type AI model used for positioning, and the correspondence includes a correspondence between the second type AI model used for positioning and the value of the first parameter; and the first network element transmitting second instruction information, which instructs to switch or update the second type AI model used for positioning from a third AI model to a fourth AI model.

[0017] According to the above technical solution, if the first network element determines that it is necessary to switch or update an AI model configured within another network element, it can send instruction information to that other network element so that the other network element can switch or update the AI ​​model in a timely manner.

[0018] In relation to the first aspect, in some implementations of the first aspect, the method further includes: a first network element performing positioning by using a fifth AI model, the fifth AI model being an AI model that matches the fourth AI model.

[0019] According to the above technical solution, when an AI model configured within another network element is switched or updated, for example, when an AI model is switched to or updated to a fourth AI model, the first network element can perform positioning by using an AI model that matches the fourth AI model. In this way, positioning can be performed jointly by using matching AI models.

[0020] In relation to the first embodiment, in some implementations of the first embodiment, the method further includes: a first network element transmitting the measurement result of a first parameter.

[0021] According to the above technical solution, the first network element can transmit the measurement results of the first parameter to other network elements. In this way, the other network element can determine whether it needs to switch or update the AI ​​model based on the measurement results of the first parameter and their corresponding relationships.

[0022] In relation to the first aspect, in some implementations of the first aspect, the transmission of a measurement result of a first network element includes: the transmission of a measurement result of a first parameter when the measurement result of the first parameter satisfies a specified condition.

[0023] For example, the specified conditions may be set in advance or configured.

[0024] In connection with the first aspect, in some implementations of the first aspect, the method further includes: a first network element receiving third indication information, where the third indication information instructs to switch or update a first type of AI model used for positioning from a first AI model to a second AI model; and the first network element performing positioning by using the second AI model.

[0025] In connection with the first aspect, in some implementations of the first aspect, the third indication information includes information about the second AI model.

[0026] For example, the third indication information instructing to switch or update a first type of AI model used for positioning from a first AI model to a second AI model includes instructing to switch or update a first type of AI model used for positioning from a first AI model to a second AI model by including information about the second AI model.

[0027] For example, the information about the second AI model may include information that can identify the second AI model or can be used to obtain the second AI model. For example, the information about the second AI model includes an identifier of the second AI model and / or all or some of the neural network parameters of the second AI model. The neural network parameters may include at least one of the number of layers of the neural network, the width of the neural network, the weights of the neurons, or some of the parameters in the activation function of the neurons.

[0028] In connection with the first aspect, in some implementations of the first aspect, the method further includes: the first network element receiving fourth indication information that instructs to switch or update a second type of AI model used for positioning from a third AI model to a fourth AI model.

[0029] For example, the fourth instruction information indicating to switch or update the second type of AI model used for positioning from the third AI model to the fourth AI model may be replaced by the fourth instruction information indicating to switch or update the second type of AI model used for positioning.

[0030] Optionally, the fourth instruction information is used to trigger the first network element to perform positioning by using the fifth AI model, and the fifth AI model is an AI model that matches the fourth AI model; alternatively, the fourth instruction information is used to trigger the first network element to execute whether to switch or update the first type of AI model used for positioning.

[0031] Optionally, the method further includes: in response to the fourth instruction information, the first network element performs positioning by using the fifth AI model, and the fifth AI model is an AI model that matches the fourth AI model.

[0032] According to the above technical solution, the first network element can learn that the second type of AI model is switched or updated. In addition, optionally, when an AI model configured in another network element is switched or updated, for example, when the AI model is switched or updated to the fourth AI model, the first network element may determine whether it is necessary to switch or update the first type of AI model used for positioning based on the measurement result of the first parameter. Alternatively, when an AI model configured in another network element is switched or updated, for example, when the AI model is switched or updated to the fourth AI model, the first network element may perform positioning by using an AI model that matches the fourth AI model. In this way, by using a matching AI model, positioning can be jointly executed.

[0033] In relation to the first embodiment, in some implementations of the first embodiment, the first parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or moving velocity of the first network element, where k is a number greater than 0.

[0034] For example, k is a number greater than 0 and less than or equal to 1.

[0035] In relation to the first embodiment, in some implementations of the first embodiment, the first type of AI model used for positioning is an AI model used for positioning configured within a first network element.

[0036] In relation to the first embodiment, in some implementations of the first embodiment, the second type of AI model used for positioning is an AI model used for positioning that is configured within a second network element.

[0037] In relation to the first embodiment, in some implementations of the first embodiment, the measurement result of the first parameter is obtained by performing channel measurements on a reference signal received at least at two time points by the first network element.

[0038] According to the above technical solution, the measurement result of the first parameter can be obtained by performing channel measurements on a reference signal received at least at two time points by the first network element. In this way, power consumption caused by frequent measurements can be reduced, and errors caused by a single inaccurate measurement can be further reduced.

[0039] According to a second embodiment, a communication method is provided. This method can be performed by a network element. Hereinafter, a second network element will be used as an example for explanation.

[0040] The method may include: a second network element receiving a measurement result of a first parameter, the measurement result of the first parameter being based on a channel measurement result, and the measurement result of the first parameter or auxiliary positioning information based on the measurement result of the first parameter being input to an artificial intelligence (AI) model used for positioning; and the second network element determining whether to switch or update the AI ​​model based on the measurement result of the first parameter and a correspondence, the correspondence including the correspondence between the AI ​​model used for positioning and the value of the first parameter.

[0041] Optionally, the second network element is a network device, a core network device, a positioning device, or a server or cloud device within an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device within an OTT system, such as a chip or circuit.

[0042] Optionally, channel measurement results are obtained based on measurements of a reference signal. Optionally, the reference signal may include an uplink reference signal or a downlink reference signal.

[0043] According to the above technical solution, there is a correspondence between the AI ​​model and the value of the first parameter. In this way, after obtaining the measurement result of the first parameter, the network element can determine a specific AI model corresponding to the measurement result of the first parameter based on the measurement result of the first parameter and the correspondence, and compare this specific AI model with the current AI model to determine whether it is necessary to switch or update the AI ​​model. It is thought that changes in the channel environment may affect positioning accuracy. For example, when a single AI model is used, changes in the channel environment may cause a deterioration in positioning performance. Since the measurement result of the first parameter is obtained based on the channel measurement result, the measurement result of the first parameter can reflect the current channel environment. In this way, by determining whether to switch or update the AI ​​model based on the measurement result of the first parameter, it is possible to implement timely switching or updating of the AI ​​model based on changes in the channel environment, thereby improving the positioning accuracy of the AI ​​model.

[0044] In relation to the second aspect, in some implementations of the second aspect, the second network element determining whether to switch or update the AI ​​model based on the measurement results and correspondence of the first parameter includes: the second network element determining whether to switch or update the first type of AI model used for positioning from the first AI model to the second AI model, where the channel measurement results are the inputs to the first type of AI model used for positioning, and the correspondence includes the correspondence between the first type of AI model used for positioning and the value of the first parameter; and the method further includes: the second network element transmitting third instruction information, where the third instruction information instructs to switch or update the first type of AI model used for positioning from the first AI model to the second AI model.

[0045] In relation to the second aspect, in some implementations of the second aspect, the third instruction information includes information about the second AI model.

[0046] In relation to the second aspect, in some implementations of the second aspect, the determination by which a second network element decides whether to switch or update an AI model based on the measurement results and correspondence of a first parameter includes: the determination by which the second network element decides to switch or update the second type of AI model used for positioning from a third AI model to a fourth AI model based on the measurement results and correspondence of a first parameter, wherein the channel measurement results are inputs to the first type of AI model used for positioning, the auxiliary positioning information is an output of the first type of AI model, the first type of AI model is switched or updated from a first AI model corresponding to a third AI model to a second AI model corresponding to a fourth AI model, the output of the first type of AI model used for positioning is an input to the second type of AI model used for positioning, and the correspondence includes a correspondence between the second type of AI model used for positioning and the value of the first parameter; and the second network element performs positioning by using the fourth AI model.

[0047] In relation to the second aspect, in some implementations of the second aspect, the method further includes a second network element transmitting fourth instruction information, the fourth instruction information instructing to switch or update the second type of AI model used for positioning from the third AI model to the fourth AI model.

[0048] In relation to the second aspect, in some implementations of the second aspect, the method further includes a second network element transmitting first instruction information, the first instruction information indicating that a first parameter is used to determine whether to switch or update the AI ​​model used for positioning.

[0049] In relation to the second aspect, in some implementations of the second aspect, the first parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or moving velocity of the first network element, where k is a number greater than 0.

[0050] For example, k is a number greater than 0 and less than or equal to 1.

[0051] In relation to the second aspect, in some implementations of the second aspect, the first type of AI model used for positioning is an AI model used for positioning configured within a first network element.

[0052] In relation to the second aspect, in some implementations of the second aspect, the second type of AI model used for positioning is an AI model used for positioning configured within a second network element.

[0053] In relation to the second aspect, in some implementations of the second aspect, the measurement result of the first parameter is obtained by performing the measurement at least once at a time point by at least one first network element.

[0054] For the advantages and possible designs of the second embodiment, please refer to the relevant description of the first embodiment. Further details will not be provided here.

[0055] According to a third embodiment, a communication method is provided. This method can be performed by a network element. Below, a second network element is used as an example for explanation.

[0056] The method includes: a second network element transmitting first instruction information and / or a correspondence, the correspondence including a correspondence between an artificial intelligence (AI) model used for positioning and a value of a first parameter, the measurement result of the first parameter being based on a channel measurement result, the measurement result of the first parameter or auxiliary positioning information based on the measurement result of the first parameter being an input to the AI ​​model, and the first instruction information indicating that the first parameter is used to determine whether to switch or update the AI ​​model used for positioning.

[0057] Optionally, the second network element is a network device, a core network device, a positioning device, or a server or cloud device within an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device within an OTT system, such as a chip or circuit.

[0058] Optionally, channel measurement results are obtained based on measurements of a reference signal. Optionally, the reference signal may include an uplink reference signal or a downlink reference signal.

[0059] According to the above technical solution, the second network element can provide, for example, a correspondence between the AI ​​model used for positioning and the value of the first parameter for another network element. In this way, the other network element can decide whether to switch or update the AI ​​model based on the correspondence. By establishing a correspondence between the AI ​​model used for positioning and the value of the first parameter, the AI ​​model can be switched or updated in a timely manner based on changes in the channel environment, thereby improving the positioning accuracy of the AI ​​model.

[0060] In relation to the third aspect, in some implementations of the third aspect, the method further includes: a second network element receiving request information, which is used to request a second AI model; and a second network element transmitting a response to the request information, which includes information about the second AI model.

[0061] For example, request information being used to request a second AI model includes: request information being used to request an update to the AI ​​model. The updated AI model is the second AI model.

[0062] For example, information about the second AI model includes information that can identify the second AI model or can be used to obtain the second AI model. For example, information about the second AI model includes an identifier for the second AI model.

[0063] In relation to the third aspect, in some implementations of the third aspect, the method further includes: a second network element receiving second instruction information, the second instruction information instructing to switch or update the second type AI model used for positioning from a third AI model to a fourth AI model, the channel measurement result being the input to the first type AI model used for positioning, auxiliary positioning information being the output to the first type AI model, the first type AI model being switched or updated from the first AI model corresponding to the third AI model to the second AI model corresponding to the fourth AI model, the output of the first type AI model used for positioning being the input to the second type AI model used for positioning, and the correspondence including the correspondence between the second type AI model used for positioning and the value of the first parameter; and the second network element performing positioning by using the fourth AI model based on the second instruction information.

[0064] In relation to the third aspect, in some implementations of the third aspect, the second type of AI model used for positioning is an AI model used for positioning that is configured within a second network element.

[0065] In relation to the third aspect, in some implementations of the third aspect, the first parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or moving velocity of the first network element, where k is a number greater than 0. For example, k is a number greater than 0 and less than or equal to 1.

[0066] For the advantages and possible designs of the third embodiment, please refer to the relevant description of the first embodiment. Further details will not be provided here.

[0067] According to a fourth aspect, a communication method is provided. This method can be performed by a network element. Below, a second network element is used as an example for explanation.

[0068] The method may include: a second network element receiving a measurement result of a second parameter from a first network element, wherein the measurement result of the second parameter includes a measurement result obtained based on the output of an artificial intelligence (AI) model configured within the first network element; and a second network element determining whether to switch or update an AI model used for positioning, configured within the second network element, based on the measurement result of the second parameter and a correspondence, wherein the correspondence includes a correspondence between the AI ​​model used for positioning and the value of the second parameter.

[0069] Optionally, the second network element is a network device, a core network device, a positioning device, or a server or cloud device within an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device within an OTT system, such as a chip or circuit.

[0070] Optionally, channel measurement results are obtained based on measurements of a reference signal. Optionally, the reference signal may include an uplink reference signal or a downlink reference signal.

[0071] For example, the AI ​​model configured within the first network element is an AI model used for positioning.

[0072] For example, the measurement result of the second parameter includes a measurement result obtained based on the output of an AI model configured within the first network element, and includes any one of the following: that the measurement result of the second parameter is the output of an AI model configured within the first network element; that the measurement result of the second parameter is a measurement result obtained through computation based on the output of an AI model configured within the first network element; that the measurement result of the second parameter is the output of an AI model configured within the first network element and a different measurement result; or that the measurement result of the second parameter is a different measurement result from the measurement result obtained through computation based on the output of an AI model configured within the first network element. The above different measurement result may be obtained through computation by the second network element or may be provided for the second network element by another network element. This is not limited to this.

[0073] For example, an AI model configured within the first network element is used for auxiliary positioning, while an AI model configured within the second network element is used for direct positioning.

[0074] In relation to the fourth aspect, in some implementations of the fourth aspect, the first parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or moving velocity of the first network element, where k is a number greater than 0.

[0075] For example, k is a number greater than 0 and less than or equal to 1.

[0076] In relation to the fourth aspect, in some implementations of the fourth aspect, the determination by which a second network element switches or updates an AI model used for positioning, configured within the second network element, based on the measurement results and correspondence of a second parameter, includes: the second network element determining, based on the measurement results and correspondence of a second parameter, to switch or update the AI ​​model used for positioning, configured within the second network element, from a third AI model to a fourth AI model; and the second network element transmitting a fifth instruction information, the fifth instruction information instructing to switch or update the AI ​​model used for positioning, configured within the second network element, from a third AI model to a fourth AI model.

[0077] According to a fifth aspect, a communication method is provided. This method may be performed by a network element. Hereinafter, a first network element will be used as an example for explanation.

[0078] The method may include: a first network element receiving a reference signal; and the first network element performing a channel measurement based on the reference signal and obtaining a channel measurement result, the channel measurement result being the input to a first artificial intelligence (AI) model used for positioning; and the output accuracy, preset thresholds, and correspondences of the first AI model being used to determine whether to switch or update the first AI model, the correspondences including the complexity of the AI ​​model used for positioning and the correspondence between the AI ​​model used for positioning.

[0079] Optionally, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal.

[0080] Optionally, the first network element is a network device or positioning device, or a component of a network device or positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0081] For example, a pre-configured threshold may indicate the expected output accuracy. This pre-configured threshold may be transmitted to the first network element by another network element, pre-configured, or pre-defined. This is not limited to these pre-configured thresholds.

[0082] According to the above technical solution, the network element can determine whether it is necessary to switch or update the AI ​​model based on the output of the AI ​​model. The AI ​​model may be an AI model used for positioning. Specifically, there is a correspondence between the AI ​​model used for positioning and the complexity of the AI ​​model used for positioning. In this way, the network element can determine whether a preset threshold is met based on the output accuracy of the current AI model. If the preset threshold is not met, the network element may further determine a specific AI model to switch or update based on the correspondence. A higher complexity of the AI ​​model indicates a higher accuracy of output that can be achieved by the AI ​​model. Therefore, since a correspondence is established between the AI ​​model used for positioning and the complexity of the AI ​​model used for positioning, the AI ​​model can be adaptively switched or updated based on changes in the channel state.

[0083] In relation to the fifth aspect, in some implementations of the fifth aspect, the first AI model is an AI model configured within a first network element and used for auxiliary positioning, or the first AI model is an AI model configured within a first network element and used for LOS identification.

[0084] According to the above technical solution, positioning can be performed by jointly using a first AI model and another AI model. The output accuracy of the first AI model affects the final positioning accuracy, and the output accuracy of the first AI model is thought to be related to complexity. For example, a higher output accuracy of the first AI model indicates a higher complexity of the first AI model, and a more complex AI model is more adaptive to complex channels. Therefore, whether or not the first AI model needs to be switched or updated can be determined by monitoring the output accuracy of the first AI model, or the first AI model can be switched or updated adaptively in a timely manner based on changes in the channel state to ensure the final positioning accuracy performance.

[0085] In relation to the fifth aspect, in some implementations of the fifth aspect, the method further includes: a first network element receiving sixth instruction information and / or correspondence, the sixth instruction information indicating that the output accuracy of the first AI model is used to determine whether to switch or update the AI ​​model used for positioning.

[0086] In relation to the fifth aspect, in some implementations of the fifth aspect, the method further includes: the first network element determining whether to switch or update the first AI model based on the output accuracy of the first AI model, a preset threshold, and a correspondence.

[0087] In relation to the fifth aspect, in some implementations of the fifth aspect, the first network element determining whether to switch or update the first AI model based on the output accuracy of the first AI model, a preset threshold, and a correspondence relationship includes: the first network element determining whether to switch or update the first AI model to a second AI model based on the output accuracy of the first AI model, a preset threshold, and a correspondence relationship, the method further includes the first network element transmitting request information, the request information being used to request the second AI model; the first network element receiving a response to the request information, the response being related to the second AI model; and the first network element performing positioning by using the second AI model.

[0088] For example, request information being used to request a second AI model includes: request information being used to request an update to the AI ​​model. The updated AI model is the second AI model.

[0089] For example, information about the second AI model includes information that can identify the second AI model or can be used to obtain the second AI model. For example, information about the second AI model includes an identifier for the second AI model.

[0090] In relation to the fifth aspect, in some implementations of the fifth aspect, the method further includes: the first network element transmitting the output of the first AI model.

[0091] In relation to the fifth aspect, in some implementations of the fifth aspect, the transmission of the output of the first AI model by the first network element includes: the transmission of the output of the first AI model when the output of the first AI model satisfies a specified condition.

[0092] For example, one of the specified conditions is that the output of the first AI model is below a predetermined threshold.

[0093] For example, the specified conditions may be set in advance or configured.

[0094] In relation to the fifth aspect, in some implementations of the fifth aspect, the output of the first AI model includes at least one of the line-of-sight probability, line-of-sight hardness determination result, angle, and time.

[0095] For example, the angle can be the angle of arrival (AOA).

[0096] For example, time may be the time difference of arrival (TDoA), or it may be the time of arrival (TOA).

[0097] According to the sixth aspect, a communication method is provided. This method can be performed by a network element. Below, a second network element is used as an example for explanation.

[0098] The method includes: a second network element receiving the output of a first artificial intelligence (AI) model used for positioning, wherein the input to the first AI model is a channel measurement result; and the second network element determining whether to switch or update the AI ​​model based on the output accuracy of the first AI model, a preset threshold, and a correspondence, wherein the correspondence includes a correspondence between the complexity of the AI ​​model used for positioning and the AI ​​model used for positioning.

[0099] Optionally, the second network element is a network device, a core network device, a positioning device, or a server or cloud device within an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device within an OTT system, such as a chip or circuit.

[0100] Optionally, channel measurement results are obtained based on measurements of a reference signal. Optionally, the reference signal may include an uplink reference signal or a downlink reference signal.

[0101] In relation to the sixth aspect, in some implementations of the sixth aspect, the first AI model is an AI model used for positioning and / or an AI model used for auxiliary positioning, configured within the first network element.

[0102] In relation to the sixth aspect, in some implementations of the sixth aspect, the second network element determining whether to switch or update an AI model based on the output accuracy of the first AI model, a preset threshold, and a correspondence relationship includes the second network element determining whether to switch or update the first AI model to the second AI model based on the output accuracy of the first AI model, a preset threshold, and a correspondence relationship, and the method further includes the second network element transmitting third instruction information, the third instruction information instructing to switch or update the first AI model to the second AI model.

[0103] In relation to the sixth aspect, in some implementations of the sixth aspect, the third instruction information includes information about the second AI model.

[0104] For example, information about the second AI model includes information that can identify the second AI model or can be used to obtain the second AI model. For example, information about the second AI model includes an identifier for the second AI model.

[0105] In relation to the sixth aspect, in some implementations of the sixth aspect, the method further includes: a second network element transmitting sixth instruction information, the sixth instruction information indicating that the output accuracy of the first AI model is used to determine whether to switch or update the AI ​​model used for positioning.

[0106] In relation to the sixth aspect, in some implementations of the sixth aspect, the output of the first AI model includes at least one of the line-of-sight probability, line-of-sight hardness determination result, angle, and time.

[0107] For example, the angle can be AOA.

[0108] For example, time can be TDoA or TOA.

[0109] According to the seventh aspect, a communication device is provided. The device is configured to perform the actions provided in any one of the first to sixth aspects. Specifically, the device may include units and / or modules, such as processing units and / or communication units, configured to perform the actions according to any one of the aforementioned implementations of any one of the first to sixth aspects. The communication device may be a first network element or a second network element.

[0110] In one implementation, this device is a communication device. When this device is a communication device, the communication unit may be a transceiver or an input / output interface, and the processing unit may be at least one processor. Optionally, the transceiver may be a transceiver circuit. Optionally, the input / output interface may be an input / output circuit.

[0111] In another implementation, this device is a circuit used in a chip, chip system, or communication device. When this device is a circuit used in a chip, chip system, or terminal device, the communication unit is an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit in the chip, chip system, or circuit, and the processing unit is at least one processor, processing circuit, or logic circuit.

[0112] According to the eighth aspect, a communication device is provided. The device includes a memory configured to store a program and at least one processor configured to execute a computer program or instructions stored in the memory in a manner provided in any implementation of any one of the first to sixth aspects. The communication device may be a first network element or a second network element.

[0113] In one implementation, this device is a communication device.

[0114] In another implementation, this device is a circuit used in a chip, chip system, or communication device.

[0115] According to the ninth aspect, the present application provides a processor configured to perform the method provided in the above aspects.

[0116] Unless otherwise specified, or insofar as it does not contradict the actual function or internal logic of the operation in the relevant description, operations such as transmit and receive operations performed by the processor's outputs and inputs, or by radio frequency circuits and antennas, may be understood as such. This is not limited to the present application.

[0117] According to a tenth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores program code to be executed by a device, and the program code is used to execute a method provided in any implementation of any one of the first to sixth aspects.

[0118] According to the eleventh aspect, a computer program product is provided. When the computer program product is executed on a computer, the computer becomes capable of performing the methods provided in any implementation of any one of the first to sixth aspects.

[0119] According to the twelfth aspect, a chip is provided. The chip includes a processor and a communication interface. The processor reads instructions stored in memory via the communication interface and performs a method provided in any one implementation of the first to sixth aspects.

[0120] Optionally, in one implementation, the chip further includes memory. The memory stores computer programs or instructions. The processor is configured to execute computer programs or instructions stored in memory. Once the computer programs or instructions are executed, the processor is configured to perform the actions provided in any one of the implementations of the first to sixth embodiments.

[0121] According to the 13th aspect, a communication system is provided. The communication system includes the first network element and / or the second network element. [Brief explanation of the drawing]

[0122] [Figure 1] This is a diagram of a wireless communication system 100 applicable to the embodiments of this application.

[0123] [Figure 2] This is a diagram of a wireless communication system 200 applicable to the embodiments of this application.

[0124] [Figure 3] This is a diagram of the structure of a neuron.

[0125] [Figure 4] This is a diagram showing the layer relationships of a neural network.

[0126] [Figure 5] This is a diagram showing the TDoA positioning.

[0127] [Figure 6] This is a diagram of the time-domain channel response.

[0128] [Figure 7] This is a diagram of LOS and NLOS.

[0129] [Figure 8] This is a diagram of a communication method 800 according to an embodiment of the present application.

[0130] [Figure 9] This is a schematic flowchart of the communication method 900 according to the embodiment of this application.

[0131] [Figure 10] This is a schematic flowchart of communication method 1000 according to another embodiment of the present application.

[0132] [Figure 11] This is a schematic flowchart of a communication method 1100 according to another embodiment of the present application.

[0133] [Figure 12] This is a schematic flowchart of a communication method 1200 according to another embodiment of the present application.

[0134] [Figure 13] This is a diagram of a communication method 1300 according to an embodiment of the present application.

[0135] [Figure 14] This is a schematic flowchart of communication method 1400 according to another embodiment of the present application.

[0136] [Figure 15] This is a diagram of a communication device 1500 according to an embodiment of the present application.

[0137] [Figure 16] This is a diagram of another communication device 1600 according to an embodiment of the present application.

[0138] [Figure 17] This is a diagram of a chip system 1700 according to an embodiment of the present application. [Modes for carrying out the invention]

[0139] The technical solutions of the embodiments of this application will be described below with reference to the drawings.

[0140] The technical solutions provided in this application can be applied to various communication systems, such as future communication systems like 5th generation (5G) or new radio (NR) systems, long-term evolution (LTE) systems, LTE frequency division duplex (FDD) systems, LTE time division duplex (TDD) systems, wireless local area network (WLAN) systems, satellite communication systems, 6th generation mobile communication systems, or systems that combine multiple systems. The technical solutions provided in this application can further be applied to device-to-device (D2D) communication, vehicle-to-everything (V2X) communication, machine-to-machine (M2M) communication, machine-type communication (MTC), Internet of Things (IoT) communication systems, or other communication systems.

[0141] Network elements within a communication system may transmit signals to or receive signals from other network elements. These signals may include information, signaling, data, etc. Network elements may be replaced by entities, network entities, devices, communication devices, communication modules, nodes, communication nodes, etc. In this disclosure, network elements are used as illustrative examples. For example, a communication system may include at least one terminal device and at least one network device. The network device may transmit downlink signals to the terminal device, and / or the terminal device may transmit uplink signals to the network device. The terminal device in this disclosure may be replaced by a first network element, the network device may be replaced by a second network element, and it may be understood that the terminal device and the network device perform the corresponding communication methods in this disclosure.

[0142] The terminal devices in the embodiments of this application include various devices having wireless communication capabilities, and the terminal devices may be configured to connect to people, objects, machines, etc. Terminal devices can be widely used in a variety of scenarios, such as cellular communication, D2D, V2X, peer-to-peer (P2P), M2M, MTC, IoT, virtual reality (VR), augmented reality (AR), industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, unmanned aerial vehicles, robots, remote sensing, passive sensing, positioning, navigation and tracking, and autonomous delivery. The terminal device may be a terminal in any one of the aforementioned scenarios, for example, an MTC terminal or an IoT terminal.Terminal devices include user equipment (UE), terminals, fixed devices, mobile station devices or mobile devices, subscriber units, handheld devices, in-vehicle devices, wearable devices, cellular phones, smartphones, session initiation protocol (SIP) phones, wireless data cards, personal digital assistants (PDAs), computers, tablet computers, notebook computers, wireless modems, handheld devices (handsets), laptop computers, computers with wireless transceiver functions, smartbooks, vehicles, satellites, and global positioning systems. System, GPS) devices, target tracking devices, flight devices (e.g., unmanned aerial vehicles, helicopters, multi-helicopter, four-helicopter, airplanes), ships, remote control devices, smart home devices, or industrial equipment; devices built into the above devices (e.g., communication modules, modems, or chips in these devices); or other processing devices connected to a wireless modem. For ease of explanation, the following examples will use terminal devices where the terminal device is a terminal or UE.

[0143] It should be understood that in some scenarios, an UE can alternatively function as a base station. For example, an UE can function as a scheduling entity providing sidelink signals between UEs in V2X, D2D, and P2P scenarios.

[0144] In embodiments of this application, the device configured to implement the functions of a terminal device may be a terminal device, or a device that can assist a terminal device in implementing its functions, such as a chip system or a chip, which may be mounted within the terminal device. In embodiments of this application, the chip system may include a chip, or it may include a chip and other separate components.

[0145] The network device in the embodiments of this application may be a device configured to communicate with a terminal device. The network device may also be called an access network device or a radio access network device. For example, the network device may be a base station. The network device in the embodiments of this application may be a radio access network (RAN) node (or device) that connects terminal devices to a radio network. In a broad sense, a base station may encompass, or be replaced by, the following various names: for example, NodeB (NodeB), evolved NodeB (eNB), next-generation NodeB (gNB), relay station, access point, transmission reception point (TRP), transmission point (TP), primary station, secondary station, motor slide retainer (MSR) node, home base station, network controller, access node, radio node, access point (AP), transmitting node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), positioning node, etc. A base station may also be a macro base station, micro base station, relay node, donor node, analog, or a combination thereof. A base station may alternatively be a communication module, a modem, or a chip located within the aforementioned device or apparatus. Alternatively, a base station may be a mobile switching station, a device performing base station functions in D2D, V2X, or M2M communications, a network-side device in a 6G network, or a device performing base station functions in a future communication system. A base station may support networks with the same or different access technologies.The specific technologies used in network devices and the specific device forms of network devices are not limited to those in the embodiments of this application.

[0146] Base stations may be fixed or mobile. For example, a helicopter or unmanned aerial vehicle may be configured as a mobile base station, and one or more cells may move based on the location of the mobile base station. In another example, a helicopter or unmanned aerial vehicle may be configured as a device for communicating with another base station.

[0147] In embodiments of this application, the device configured to implement the functions of a network device may be a network device, or a device capable of assisting the network device in implementing its functions, such as a chip system or a chip, which may be mounted on the network device. In embodiments of this application, the chip system may include a chip or include a chip and other separate components.

[0148] Network devices and terminal devices may be deployed on land, where deployment may include indoors or outdoors, or may be handheld or vehicle-mounted; they may be deployed on water; or they may be deployed in the air on aircraft, balloons, and satellites. The scenarios in which network devices and terminal devices are located are not limited to the embodiments of this application. In addition, each of terminal devices and network devices may be an entity including a hardware device, software functions running on dedicated hardware, software functions running on general-purpose hardware, for example virtualization functions instantiated on a platform (e.g., a cloud platform), or dedicated or general-purpose hardware devices and software functions. The specific forms of terminal devices and network devices are not limited in this application.

[0149] First, a brief description of the communication system applicable to the embodiments of this application will be given.

[0150] Figure 1 is a diagram of a wireless communication system 100 applicable to embodiments of the present application. As shown in Figure 1, the wireless communication system includes a wireless access network 100. The wireless access network 100 may be a next-generation (e.g., 6G or higher version) wireless access network or a conventional (e.g., 5G, 4G, 3G, or 2G) wireless access network. One or more terminal devices (120a to 110j, collectively referred to as 120) may be interconnected or connected to one or more network devices (120a and 110b, collectively referred to as 110) within the wireless access network 100. Figure 1 is merely a diagram. The wireless communication system may further include other devices, such as a core network device, a wireless relay device, and / or a wireless backhaul device, which are not shown in Figure 1.

[0151] In practical applications, a wireless communication system may include multiple network devices (also called access network devices) and multiple terminal devices, but is not limited to these. One network device may provide services to one or more terminal devices, and one terminal device may access one or more network devices. The number of terminal devices and network devices included in the wireless communication system is not limited to the embodiments of this application.

[0152] Figure 2 is a diagram of a wireless communication system 200 applicable to one embodiment of the present application. As shown in Figure 2, the wireless communication system 200 may include at least one network device, for example, the network device 210 shown in Figure 2. The wireless communication system 200 may further include at least one terminal device, for example, terminal devices 220 and 230 shown in Figure 2. The wireless communication system 200 may further include a positioning device, for example, the positioning device 240 shown in Figure 2. For example, the positioning device 240 is a Location Management Function (LMF).

[0153] Positioning devices and network devices can communicate with each other using interface messages. For example, network device 210 is a gNB and positioning device 240 is an LMF. The gNB and LMF can exchange information using NR positioning protocol A (NRPPa) messages. In another example, network device 210 is an eNB and positioning device 240 is an LMF. The eNB and LMF can exchange information using LTE positioning protocol (LPP) messages.

[0154] Terminal devices and positioning devices may communicate directly with each other, or they may communicate via another device. This other device may be, for example, a network device and / or a core network element. For example, as shown in Figure 2, terminal devices 220 and 230 may communicate with positioning device 240 via network device 210.

[0155] Optionally, the positioning device and the network device may be different modules of the same device, or they may be different, separate devices.

[0156] Optionally, the communication system further includes at least one artificial intelligence (AI) node.

[0157] Optionally, an AI node may be deployed in one or more of the following: a network device, a terminal device, a core network, or a positioning device, or an AI node may be deployed independently, for example, in a location other than one of the aforementioned devices. An AI node may communicate with another device in a communication system, which may be, for example, one or more of the following: a network device, a terminal device, a core network element, or a positioning device.

[0158] Optionally, an AI node is configured to perform AI-related operations. For example, AI-related operations may include one or more of the following: model failure testing, model performance testing, model training testing, or data set testing.

[0159] For example, a network device transfers data related to an AI model and reported by a terminal device to an AI node, and the AI ​​node performs AI-related operations. In another example, a network device or terminal device may transfer data related to an AI model to an AI node, and the AI ​​node performs AI-related operations. In yet another example, the AI ​​node may send one or more outputs of the AI-related operations, such as a trained neural network model, model evaluation, or test results, to the network device and / or terminal device. For example, the AI ​​node may send the outputs of the AI-related operations directly to the network device and terminal device. In yet another example, the AI ​​node may send the outputs of the AI-related operations to a terminal device via a network device. In yet another example, the AI ​​node may send the outputs of the AI-related operations to a network device via a terminal device.

[0160] It should be understood that the number of AI nodes is not limited in this application. For example, when there are multiple AI nodes, the multiple AI nodes may be divided based on their function. For example, different AI nodes may perform different functions.

[0161] It can be further understood that an AI node may be an independent device, may be integrated into the same device to implement different functions, may be a network element within a hardware device, may be a software function running on dedicated hardware, or may be a virtualized function instantiated on a platform (e.g., a cloud platform). Specific forms of AI nodes are not limited in this application.

[0162] It should be further understood that Figures 1 and 2 are simplified illustrative diagrams for ease of understanding. The wireless communication system may further include other network devices, other terminal devices, or AI nodes, etc., which are not shown in Figures 1 and 2.

[0163] To facilitate understanding of the embodiments of this application, the terminology used in these embodiments will be briefly explained below.

[0164] 1. Artificial Intelligence: Artificial intelligence enables machines to learn and accumulate experience, solving problems such as natural language comprehension, image recognition, and chess play that humans can solve through experience. Artificial intelligence can be understood as intelligence represented by machines created by humans. In general, artificial intelligence is the technology that presents human intelligence through the use of computer programs. The goal of artificial intelligence includes understanding intelligence by constructing computer programs for symbolic reasoning or inference.

[0165] 2. Machine Learning: Machine learning is an implementation of artificial intelligence. It is a method that provides machines with the ability to learn, enabling them to implement functions that cannot be implemented directly through programming. In practice, machine learning is a method for training models by using data and then using those models for prediction. There are many types of machine learning, such as neural networks (NN), decision trees, and support vector machines. Machine learning theory mainly involves designing and analyzing several algorithms that enable computers to learn automatically. Machine learning algorithms are algorithms that automatically analyze data to obtain rules and use those rules to predict unknown data.

[0166] 3. Neural Networks: Neural networks are a specific embodiment of machine learning methods. Neural networks are mathematical models that mimic the behavioral features of animal neural networks and process information. The idea of ​​neural networks originates from the neuronal structure of brain tissue. Each neuron performs a weighted sum operation on the input values ​​of the neuron and outputs the result obtained through the weighted sum operation via an activation function.

[0167] Figure 3 is a diagram of the structure of a neuron. As shown in Figure 3, the inputs of the neuron are x=[x_0,x_1,...,x_n], the weights corresponding to the inputs are w=[w,w_1,...,w_n], and the bias of the weighted sum is b. b can be an integer, a decimal, or any of the various possible values ​​such as a complex number. The form of the activation function can be diverse. For example, if the activation function of the neuron is y=f(z)=max(0,z), then the output of the neuron is:

number

number

[0168] A neural network typically has a multilayer structure, and each layer may contain one or more logic decision units. These logic decision units are sometimes called neurons. Increasing the depth and / or width of a neural network can improve its expressive power, providing more powerful information extraction and abstraction modeling capabilities for complex systems. The depth of a neural network may be understood as the number of layers it contains, and the number of neurons in each layer may be called the layer width.

[0169] Figure 4 shows the layer relationships of a neural network.

[0170] In a possible implementation, a neural network includes an input layer and an output layer. The input layer of the neural network performs neuronal processing on the received input and transfers the results to the output layer. The output layer retrieves the output result of the neural network.

[0171] In another possible implementation, the neural network includes an input layer, hidden layers, and an output layer, as shown in Figure 4. The input layer of the neural network performs neuronal processing on the received input and forwards the results to an intermediate hidden layer. The hidden layer forwards the computation results to the output layer or an adjacent hidden layer. Finally, the output layer obtains the output result of the neural network. A single neural network may include one or more sequentially connected hidden layers; this is not limited to this.

[0172] A loss function may be defined during the neural network training process. The loss function is used to measure the difference between the model's predicted values ​​and the actual values. In the neural network training process, the loss function describes the gap or difference between the neural network's output values ​​and the ideal target values. The neural network training process is the process of tuning the neural network's parameters so that the value of the loss function falls below a threshold or satisfies the target requirement. The neural network parameters may include at least one of the following: the number of layers in the neural network, the width of the neural network, the weights of the neurons, or at least one of the parameters in the neuron activation function.

[0173] 4. AI Model: An AI model is an algorithm or computer program capable of implementing AI functionality. An AI model represents a mapping relationship between the inputs and outputs of a model, or an AI model is a functional model that maps inputs of a particular dimension to outputs of a particular dimension. The parameters of a functional model can be obtained through machine learning training. For example, f(x) = ax² + b is a quadratic functional model and may be considered an AI model, where a and b are parameters of the AI ​​model, and a and b can be obtained through machine learning training. For example, the AI ​​models referred to in the following embodiments of this application are not limited to neural networks, linear regression models, decision tree models, support vector machines (SVMs), Bayesian networks, Q-learning models, or other machine learning (ML) models.

[0174] AI model design primarily includes a data collection phase (e.g., collection of training data and / or inference data), a model training phase, and a model inference phase. It may also include an inference result application phase. In the aforementioned data collection phase, a data source is used to provide training datasets and inference data. In the model training phase, the training data provided by the data source is analyzed or trained to obtain an AI model. Obtaining an AI model through learning using model training nodes is equivalent to obtaining a mapping relationship between the inputs and outputs of the AI ​​model through learning based on the training data. In the model inference phase, the AI ​​model trained in the model training phase is used to perform inference based on the inference data provided by the data source, and the inference results are obtained. This phase may be understood as follows: inference data is input to the AI ​​model, and output is obtained through the AI ​​model. The output is the inference result. The inference result may indicate configuration parameters used (executed) by actor objects, and / or operations performed by actor objects. The inference result is released in the inference result application phase. For example, the inference result may be uniformly planned by actor entities. For example, an actor entity may send inference results to one or more actor objects (e.g., core network devices, access network devices, or terminal devices) for execution. In another example, an actor entity may further feed back the performance of an AI model to a data source to facilitate subsequent AI model update training.

[0175] It can be understood that AI models may be implemented using hardware circuits, software, or a combination of software and hardware. This is not limited to these. Non-exclusive examples of software include program code, programs, subprograms, instructions, instruction sets, code, code segments, software modules, application programs, software application programs, etc.

[0176] 5. Dataset: A dataset is the data used for model training, model validation, or model testing in machine learning. The quantity and quality of the data affect the effectiveness of machine learning. Training data may include the inputs of an AI model, or it may include the inputs of an AI model and the target output. The target output is the target value output by the AI ​​model and may also be called the output truth value, comparison truth value, label, or label sample.

[0177] 6. Hyperparameters: Hyperparameters are one or more parameters of a neural network, such as the number of layers, the number of neurons, the activation function, and the loss function.

[0178] 7. Model Training: Model training is the process of training model parameters by selecting an appropriate loss function and using an optimization algorithm to make the value of the loss function below a threshold or to make the value of the loss function satisfy the target requirements.

[0179] 8. Model Application: Model application involves using a trained model to solve real-world problems.

[0180] 9. Fingerprint: Multipath propagation of signals is environment-dependent. The multipath structure of a channel at each location is unique. Reflection and refraction of the radiotelegraph transmitted by the terminal generate a specific mode of multipath signal that is closely related to the surrounding environment. This multipath feature is sometimes called the "fingerprint" information of the location. The method of performing positioning based on this "fingerprint" information is sometimes called fingerprint positioning.

[0181] 10. Time Difference of Arrival (TDoA): TDoA is a method of positioning that uses time difference.

[0182] Figure 5 is a diagram of TDoA positioning. As shown in Figure 5, assume that the distance between network device #1 and the terminal device is d1, the transmission time of the signal between network device #1 and the terminal device is t1, the distance between network device #2 and the terminal device is d2, the transmission time of the signal between network device #2 and the terminal device is t2, the distance between network device #3 and the terminal device is d3, and the transmission time of the signal between network device #3 and the terminal device is t3.

[0183] In TDoA, for example, multiple network devices may transmit reference signals, such as positioning reference signals (PRSs), to a terminal device. The terminal device determines its position by measuring the time difference in the arrival of the reference signals. This time difference in the arrival of the reference signals is sometimes also called the reference signal time difference (RSTD). Let's use Figure 5 as an example. Suppose that P1 is the reference signal transmitted to the terminal device by network device #1, P2 is the reference signal transmitted to the terminal device by network device #2, and P3 is the reference signal transmitted to the terminal device by network device #3. The terminal device measures the time difference in arrival between P2 and P1, i.e., t2-t1. Based on t2-t1, the terminal device may estimate the distance difference d2-d1 between network device #2 and network device #1 and obtain a curve. Each point on the curve coincides with the distance difference d2-d1 between network device #2 and network device #1. Similarly, the terminal device measures the time difference in arrival between P3 and P1, i.e., t3-t1. Based on t3-t1, the terminal device can estimate the distance difference d3-d1 between network device #3 and network device #1 and obtain another curve. Each point on the curve corresponds to the distance difference d3-d1 between network device #3 and network device #1. The position of the terminal device can be determined by using the intersection of the two curves above. Using a mathematical model, this can be expressed as Equation 1 below:

number

[0184] (ai,bi) represents the position coordinates of network device #i. (a,b) represents the position coordinates of the terminal device whose position should be determined. For example, a represents the position coordinates on the X axis of the terminal device whose position should be determined, and b represents the position coordinates on the Y axis of the terminal device whose position should be determined. c represents the speed of light.

[0185] Because certain synchronization errors exist between different network devices, there is uncertainty in the corresponding measurements, which corresponds to the intervals shown by the dashed lines in Figure 5.

[0186] The above describes a method for performing positioning based on a reference signal transmitted from a network device to a terminal device, with reference to Figure 5. This method is sometimes called downlink TDoA (DL-TDoA) or observed time difference of arrival (OTDoA). Similarly, positioning may alternatively be performed based on a reference signal transmitted from a terminal device to a network device, such as a sounding reference signal (SRS). This positioning method is sometimes called uplink TDoA (UL-TDoA).

[0187] In addition to time difference measurement, it may be understood that positioning may be performed by measuring angles. The angle may be the angle of arrival (AOA) or the angle of departure (AoD). The angle of arrival represents the angle between the direction in which the receiving end receives the signal and the reference direction. The angle of departure represents the angle between the direction in which the transmitting end transmits the signal and the reference direction. The reference direction may be a direction determined based on the position and / or shape of the antenna.

[0188] In actual communication scenarios, noise and interference can introduce specific measurement errors in time or angle measurements, resulting in corresponding errors in positioning results.

[0189] Figure 6 shows the time-domain channel response. In Figure 6, the horizontal coordinates are sampling points, and the vertical coordinates are the power at different time-domain sampling points of the time-domain channel response obtained by measuring a reference signal. As shown in Figure 6, the signal power at the time-domain sampling points of the starting segment is weak and may correspond to a noise signal, while the signal power at the time-domain sampling points of the intermediate segments is strong and may correspond to the multipath response of the actual signal. In real-world scenarios, it is necessary to determine the starting position of the actual signal in an environment with interference and noise in order to obtain accurate signal propagation times.

[0190] In addition, in positioning measurements, signals from non-line of sight (NLOS) can typically introduce significant estimation errors; therefore, the measurement signal must be from the line of sight (LOS).

[0191] Figure 7 shows the LOS and NLOS. As shown in Figure 7, the LOS between the network device and the terminal device (dashed line in Figure 7) is blocked by an obstacle, and the reference signal transmitted between the network device and the terminal device is actually the reflected NLOS (solid line in Figure 7). From the figure, it can be seen that the distance of the NLOS (i.e., d2 + d3) is greater than the distance of the LOS (i.e., d1). If the NLOS is treated as the LOS during position estimation, a large measurement error can occur. Therefore, the classification of LOS and NLOS is also important for positioning accuracy.

[0192] In this application, “instruction” may include direct, indirect, explicit, or implicit instruction. When an instruction is described as instructing A, it may be understood that the instruction transmits A, directly instructs A, or indirectly instructs A. Indirect instruction may mean that the instruction directly instructs B and the correspondence between B and A in order to instruct A by using the instruction. The correspondence between B and A may be predefined or pre-stored in the protocol, or obtained by using the configuration between network elements.

[0193] In this application, information indicated by indication information is referred to as indicated information. In a specific implementation process, indicated information can be indicated in several ways, including, but not limited to, ways that directly indicate the indicated information. For example, indicated information and an index of indicated information directly indicate the indicated information. Alternatively, indicated information may be indicated indirectly by indicating other information, in which case there is a relationship between the other information and the indicated information. Alternatively, only a part of the indicated information may be indicated, with the other part of the indicated information being known or pre-agreed. For example, a particular indication may be indicated by using a pre-agreed (e.g., specified in a protocol) sequence of multiple pieces of information, which can reduce the overhead of the indication to some extent. In addition, indicated information may be transmitted as a whole, or it may be divided into multiple sub-pieces for separate transmission, and the transmission cycle and / or timing of these sub-pieces may be the same or different.

[0194] It should be further understood that "at least one" in this application refers to one or more items. "Multiple" means two or more items. "And / or" describes a relationship of association to describe related objects, indicating that three relationships may exist. For example, A and / or B could refer to the following three cases: that only A exists, that both A and B exist, and that only B exists. The letter " / " generally indicates an "or" relationship between related objects. In addition, it should be understood that while terms such as the first and second may be used to describe objects in this application, these objects are not limited by these terms. These terms are simply used to distinguish objects from one another.

[0195] The communication method provided in the embodiments of this application will be described in detail below with reference to the attached drawings. The embodiments provided in this application may be applied to the communication system shown in Figure 1 or Figure 2, but are not limited thereto.

[0196] In the following embodiments, the first network element may be an inference network element of an AI model, or a memory network element of an AI model library and / or a training network element of an AI model. The second network element may be a training network element of an AI model and / or a memory network element of an AI model library, or an inference network element of an AI model. An AI model used for positioning may be deployed in the first network element; in other words, the first network element may perform positioning by using an AI model. An AI model used for positioning may also be placed in the second network element; in other words, the second network element may perform positioning by using an AI model. An AI model library may contain one or more AI models.

[0197] For example, the first and second network elements may be logically deployed separately. In different implementations, the first and second network elements may be physically located in the same network element or in different network elements. This is not limited. For example, the first network element may be a terminal device, and the second network element may be a server (also called a host) or cloud device in an over-the-top (OTT) system, and the terminal device may communicate with the server or cloud device in the OTT system via the internet. In another example, the first network element may be a module (e.g., a physical layer module) within a device (e.g., a terminal device), and the second network element may be another module (e.g., an application layer module, e.g., an application module connected to an OTT server) within the device (e.g., a terminal device). In embodiments of this application, it may be understood that the modules may be implemented by hardware or by software. This is not limited.

[0198] For example, the first network element may be a terminal device or a component of a terminal device (e.g., a chip or circuit), the second network element may be a network device or a component of a network device (e.g., a chip or circuit), the second network element may be a positioning device (e.g., an LMF) or a component of a positioning device (e.g., a chip or circuit), the second network element may be a core network element or a component of a core network element (e.g., a chip or circuit), or the second network element may be a server or cloud device in an OTT system or a component of a server or cloud device in an OTT system (e.g., a chip or circuit).

[0199] Figure 8 is a diagram of a communication method 800 according to an embodiment of the present application. The method 800 shown in Figure 8 may include the following steps:

[0200] 810: The first network element receives the reference signal.

[0201] For example, the reference signal may be a reference signal used for positioning, such as a PRS or a preamble.

[0202] In one example, the reference signal may be a reference signal between a terminal device and a network device, such as an uplink positioning SRS or a downlink PRS. For example, the first network element is a terminal device, and the first network element receives a downlink PRS from the network device. In another example, the first network element is a network device, and the first network element receives an uplink positioning SRS from the terminal device.

[0203] In another example, the reference signal may be a reference signal between terminal devices, such as a sidelink-positioning reference signal (SL-PRS). For example, the first network element is a terminal device, and the first network element receives an SL-PRS from another terminal device.

[0204] 820: A first network element performs channel measurement based on a reference signal and obtains channel measurement results, where the channel measurement results, or auxiliary positioning information based on the channel measurement results, are inputs to the AI ​​model used for positioning, and the measurement results and correspondences of the first parameter are used to determine whether to switch or update the AI ​​model, where the measurement results of the first parameter are obtained based on the channel measurement results, and the correspondences include the correspondence between the AI ​​model used for positioning and the value of the first parameter.

[0205] For example, an AI model library contains W AI models, where W is an integer greater than or equal to 1. Suppose it is determined that an AI model needs to be switched or updated based on the measurement results and correspondence of the first parameter, for example, switching or updating from AI model #1 to AI model #2. If AI model #2 is one of the W AI models, then AI model #1 is switched to AI model #2. Alternatively, if AI model #2 does not belong to any of the W AI models, then AI model #1 is updated to AI model #2. In other words, an AI model may be updated when there is no switchable AI model in the AI ​​model library. For example, an AI model may be updated in one of the following ways: retraining the original AI model using a dataset based on the original AI model, where the dataset may optionally contain a small number of samples; retraining some functional layers of the original AI model using a dataset based on the original AI model, where the dataset may optionally contain a small number of samples; retraining the AI ​​model using a dataset, where the dataset may optionally contain a large number of samples; or retraining some functional layers of the AI ​​model using a dataset, where the dataset may optionally contain a large number of samples. This is not limited to these methods.

[0206] For the sake of clarity, in the following explanation, "adjusting the AI ​​model" will be used to mean "switching or updating the AI ​​model." In other words, "adjusting the AI ​​model" as described below may be replaced with "switching the AI ​​model" or "updating the AI ​​model."

[0207] Based on embodiments of this application, a network element can determine whether or not an AI model needs to be adjusted based on the input features of the AI ​​model. The AI ​​model may be an AI model used for positioning. Specifically, if the input to the AI ​​model is a channel measurement result, the measurement result of the first parameter, i.e., the input features of the AI ​​model, may be obtained based on the channel measurement result, and a correspondence exists between the AI ​​model and the value of the first parameter. In this way, the network element can determine a specific AI model corresponding to the measurement result of the first parameter based on the correspondence and the measurement result of the first parameter obtained based on the channel measurement result, and can determine whether or not an AI model needs to be adjusted by comparing the specific AI model with the current AI model. It is thought that changes in the channel environment may affect positioning accuracy. For example, when a single AI model is used, changes in the channel environment may cause a deterioration in positioning performance. Since the measurement result of the first parameter is obtained based on the channel measurement result, the measurement result of the first parameter can reflect the current channel environment. In this way, by determining whether or not to adjust the AI ​​model based on the measurement result of the first parameter, adjustment of the AI ​​model based on changes in the channel environment can be implemented in a timely manner, thereby improving the positioning accuracy of the AI ​​model.

[0208] The embodiments of this application do not limit the method for performing the initial selection of the AI ​​model used for positioning. For example, the AI ​​model configured within the first network element may be trained by the first network element, trained by another network element (e.g., a second network element) and transmitted to the second network element, or selected for the first network element from the current AI model library by another network element (e.g., a second network element).

[0209] In one example, channel measurement results can be used as input to an AI model used for positioning. In the case of an AI model used for positioning, the input to the AI ​​model may be channel measurement results obtained through measurements based on a reference signal, and the output to the AI ​​model may be position-related information of the first network element. For simplicity, the AI ​​model used for positioning will be abbreviated as "AI model" below.

[0210] Where possible, the AI ​​model is used to directly locate the position of the first network element. In this case, the input to the AI ​​model is the channel measurement result obtained through measurements based on a reference signal, and the output is the position information of the first network element.

[0211] In another possible case, the AI ​​model is used to perform auxiliary positioning (or indirect positioning) for the position of a first network element. In this case, the input to the AI ​​model is the channel measurement result obtained through measurements based on a reference signal, and the output is auxiliary positioning information, which is used to determine the position of the first network element. For example, the auxiliary positioning information may include, but is not limited to, at least one of the following: LOS probability, AOA, AoD, TOA, or arrival time difference of the reference signal.

[0212] In another example, auxiliary positioning information based on channel measurement results may be used as input to an AI model. For an AI model used for positioning, the input to the AI ​​model may be channel measurement results-auxiliary positioning information, and the output of the AI ​​model may be position-related information of the first network element. The auxiliary positioning information based on channel measurement results may be obtained by conventional methods or may be the output of an AI model; however, this is not limited.

[0213] Where possible, the AI ​​model is used to directly determine the position of the first network element. In this case, the input to the AI ​​model is the channel measurement result—auxiliary positioning information—and the output is the position information of the first network element.

[0214] In another possible scenario, the AI ​​model is used to perform auxiliary positioning (or indirect positioning) for the position of the first network element. In this case, the input to the AI ​​model is channel measurement results—auxiliary positioning information, and the output is another piece of auxiliary positioning information, which is used to determine the position of the first network element.

[0215] It should be noted that in the embodiments of this application, examples are primarily used for illustrative purposes in which channel measurement results are obtained through channel measurement based on a reference signal. This is not limited to this. For example, the channel measurement results may further include pedestrian dead-reckoning (PDR) measurement results from a terminal, or the channel measurement results may further include environmental monitoring identification results from a camera, for example, environmental monitoring identification results from a surveillance camera in an indoor factory.

[0216] The correspondence includes the relationship between the AI ​​model and the value of the first parameter. In other words, different values ​​of the first parameter correspond to different AI models. Thus, the AI ​​model corresponding to the measurement result of the first parameter can be learned by using the measurement result of the first parameter (i.e., the value of the first parameter obtained through this measurement).

[0217] The first and second types of AI models will be mentioned multiple times below.

[0218] For example, the first type of AI model represents an AI model deployed within a first network element, or an AI model configured within a first network element. AI-related operations may be implemented by configuring AI functions (e.g., AI modules) within the first network element. Positioning, for example, is performed using an AI model. In this example, the correspondence may include the correspondence between the first type of AI model and the value of the first parameter. Whether the first type of AI model (i.e., the AI ​​model configured within the first network element) needs to be adjusted can be determined based on the measurement result of the first parameter (i.e., the value of the first parameter obtained through the current measurement) and the correspondence between the first type of AI model and the value of the first parameter.

[0219] For example, a second type of AI model represents an AI model deployed within a second network element, or an AI model configured within a second network element. AI-related operations may be implemented by configuring AI functions (e.g., AI modules) within the second network element. Positioning, for example, is performed using an AI model. In this example, the correspondence may include the correspondence between the second type of AI model and the value of the first parameter. Whether the second type of AI model (i.e., the AI ​​model configured within the second network element) needs to be adjusted can be determined based on the measurement result of the first parameter (i.e., the value of the first parameter obtained through the current measurement) and the correspondence between the second type of AI model and the value of the first parameter.

[0220] Optionally, the first type AI model and the second type AI model may be used jointly for positioning. For example, the first type AI model may be used for auxiliary positioning, and the second type AI model may be used for direct positioning. For example, the input to the first type AI model may be the channel measurement result, and the output of the first type AI model may be auxiliary positioning information. The input to the second type AI model may be auxiliary positioning information; in other words, the input to the second type AI model may be the output of the first type AI model, and the output of the second type AI model may be the position information of the first network element.

[0221] Optionally, the first type AI model and the second type AI model may be used separately. For example, the first type AI model may be used for direct positioning or auxiliary positioning. In another example, the second type AI model may be used for direct positioning or auxiliary positioning. When the first type AI model or the second type AI model is used for auxiliary positioning, the first network element or the second network element may acquire position information based on auxiliary positioning information in a way that is not based on the AI ​​model, for example, in a conventional way.

[0222] In the following embodiments, unless otherwise stated, the correspondences in the following embodiments may be applicable to the correspondence between a first type AI model and the value of the first parameter (abbreviated as the first type correspondence), and may also be applicable to the correspondence between a second type AI model and the value of the first parameter (abbreviated as the second type correspondence). It can be understood that the first type correspondence and the second type correspondence may be the same or different. For example, there may be an intersection between the first type correspondence and the second type correspondence, for example, the first type correspondence may be a subset of the second type correspondence, and the second type correspondence may be a subset of the first type correspondence, or the first type correspondence and the second type correspondence may be completely different.

[0223] In addition, in this application, there may be multiple AI models that correspond to a specific value of the first parameter. Therefore, the correspondence between an AI model and the value of the first parameter in this application may be replaced with a correspondence between an AI model group and the value of the first parameter. An AI model group corresponding to the same value of the first parameter may include one or more AI models. These AI models may have several different characteristics. For example, the characteristics may include one or more of the following: implementation complexity, performance, etc. After a specific value of the first parameter is determined, the specific AI model selected from the corresponding AI model group may be predefined according to a protocol, or it may be determined based on system requirements or settings. This is not limited to this specification. When an AI model to be switched or updated is indicated, an AI model group may be indicated, or a specific AI model within an AI model group may be indicated.

[0224] For example, a correspondence can exist in the form of a table, function, text, or string, and can be stored or transmitted. Table 1 is an example of presenting a correspondence in table format. [Table 1]

[0225] The values ​​N1, N2, N3, and N4 of the first parameter in Table 1 may be specific values ​​or a range of values. This is not limited to these values.

[0226] Table 1 is used as an example. The value of the first parameter corresponding to AI model #1 is N1. In other words, if the value of the first parameter N is N1, the AI ​​model is AI model #1. The value of the first parameter corresponding to AI model #2 is N2. In other words, if the value of the first parameter N is N2, the AI ​​model is AI model #2. For example, if the AI ​​model currently used for positioning is AI model #1, and the value of the first parameter actually obtained through measurement is N2, this means that the AI ​​model needs to be adjusted, in other words, AI model #1 needs to be adjusted to AI model #2.

[0227] The measurement result of the first parameter is obtained based on the channel measurement result, and specifically, the first parameter represents a parameter that can reflect the channel environment or the channel state. In other words, changes in the channel environment can be reflected in the measurement result of the first parameter. For example, the first parameter represents a parameter that can reflect the degree of NLOS of the channel.

[0228] Optionally, the first parameter includes one or more of the following: a channel time-domain quality indicator, a channel frequency-domain quality indicator, the number of paths in the channel impulse response (CIR) having energy greater than k times the energy of the first path, the average power of multiple sampling points, the LOS probability, the signal-to-interference plus noise ratio (SINR), the reference signal received power (RSRP), the received signal strength indicator (RSSI), the Rice coefficient, or the moving velocity of the first network element. It can be understood that these indicators are obtained by measuring the reference signal or by performing a corresponding process on the measurement results of the reference signal. The specific processing process is not limited herein and may be, for example, any known or future processing.

[0229] Example 1: The first parameter includes a channel time-domain quality indicator.

[0230] In this example, the correspondence may include the correspondence between an AI model and a channel time-domain quality indicator. In other words, different values ​​of the channel time-domain quality indicator correspond to different AI models. Thus, the corresponding AI model can be learned based on the channel time-domain quality indicator values ​​obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0231] Specifically, the first network element performs channel measurements based on a reference signal to obtain channel measurement results, and obtains a channel time-domain quality indicator through calculations based on the channel measurement results. For example, the first network element performs channel measurements based on a reference signal to obtain a CIR, and obtains a channel time-domain quality indicator through calculations based on the CIR. In the embodiments of this application, when the first parameter is calculated based on the CIR, the number of sampling points is not limited. In the example shown in Figure 6, the number of sampling points is 100. In addition, the CIR may be replaced with any one of the following: time-aligned CIR, cross-correlated sequences of multiple CIR sequences, normalized CIR, etc. Further details are not described again below.

[0232] A channel time-domain quality indicator may include, for example, one or more of the following: timing error group (TEG), inter-site synchronization error, or delay spread. TEG and delay spread may be measured by a first network element. Inter-site synchronization error may be provided by another network element, which may be, for example, a different network element from the first and second network elements. If the channel time-domain quality indicator is TEG, TEG can reflect changes in the channel environment, allowing the AI ​​model to be adaptively adjusted based on those changes. In addition, since TEG is a feedback value (in other words, the first network element measures TEG), the feedback value may be reused to determine whether the AI ​​model needs to be adjusted, thereby reducing the overhead caused by monitoring the first parameter. If the channel time-domain quality indicator is delay spread, delay spread can reflect changes in the channel environment, allowing the AI ​​model to be adaptively adjusted based on those changes. In addition, because delayed spreads can accurately reflect changes in the channel environment, using delayed spreads provides a high degree of accuracy in determining whether or not to adjust the AI ​​model.

[0233] For example, Table 2 shows an example of the correspondence between the AI ​​model and the site-to-site synchronization error. [Table 2]

[0234] Table 2 shows that different site-to-site synchronization errors correspond to different AI models. For example, when the site-to-site synchronization error is level 1, the corresponding AI model is AI model #1. Therefore, the corresponding AI model can be learned based on the site-to-site synchronization error obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0235] It should be understood that Table 2 is an example for illustrative purposes only and is not limited thereto. For example, Table 2 may have more inter-site synchronization error levels.

[0236] For example, Table 3 shows an example of the correspondence between AI models and lazy spreads. [Table 3]

[0237] Delay Spread: After propagating through a multipath radio channel, transmitted signals arrive at the receiving end at different times. The signal that arrives at the receiving end fastest is used as the reference. The time difference between the reference signal and another signal corresponds to the delay spread. Measurement is usually performed using the mean delay spread or the root mean square delay spread. The smaller the delay spread, the closer it is to a pure LOS (Low-Levels-of-Situation) path.

[0238] Table 3 shows that different delay spreads correspond to different AI models. For example, when the delay spread is 'a', the corresponding AI model is AI model #1. Therefore, the corresponding AI model can be learned based on the delay spread obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0239] It should be understood that Table 3 is an example for illustrative purposes only. This is not limited to this. For example, in Table 3, more delay spreads may be split.

[0240] Example 2: The first parameter includes a channel frequency domain quality indicator.

[0241] In this example, the correspondence may include the correspondence between an AI model and a channel frequency domain quality evaluation indicator. In other words, different values ​​of the channel frequency domain quality indicator correspond to different AI models. Thus, the corresponding AI model can be learned based on channel frequency domain quality indicator values ​​obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0242] Specifically, the first network element performs channel measurements based on a reference signal to obtain channel measurement results, and obtains a channel frequency domain quality indicator through calculations based on the channel measurement results. For example, the first network element performs channel measurements based on a reference signal to obtain a CIR, and obtains a channel frequency domain quality indicator through calculations based on the CIR. In another example, the first network element performs channel measurements based on a reference signal to obtain a channel frequency domain response (CFR), and obtains a channel frequency domain quality indicator through calculations based on the CFR. In embodiments of this application, when the first parameter is calculated based on the CFR, the bandwidth, subbands, and number of ports of the CFR are not limited. In addition, the CFR may be replaced with a normalized CFR, details of which are not described again below.

[0243] The channel frequency domain quality indicator may be, for example, a Doppler frequency shift or a Doppler frequency. When the channel frequency domain quality indicator is a Doppler frequency shift or a Doppler frequency, the Doppler frequency shift or Doppler frequency can reflect changes in the channel environment, allowing the AI ​​model to be adaptively adjusted based on changes in the channel environment. In addition, because the Doppler frequency shift or Doppler frequency can accurately reflect changes in the channel environment, using the Doppler frequency shift or Doppler frequency provides high accuracy in determining whether the AI ​​model needs to be adjusted.

[0244] In addition, the Doppler shift can reflect the movement speed of the first network element. Therefore, different movement speeds can correspond to different AI models. For example, a corresponding movement speed can be learned based on the Doppler frequency shift obtained through measurement, and a corresponding AI model can be learned based on that movement speed. For example, the unit of movement speed may be kilometers per hour (km / h), and the value of movement speed may be, for example, 0, 10, 30, 60, 90, or 120. Different values ​​of movement speed correspond to different AI models.

[0245] For an example of the correspondence when the first parameter is a channel frequency domain quality indicator, please refer to Table 2 or Table 3. Further details will not be explained again here.

[0246] Example 3: The first parameter includes the number of paths whose energy is greater than k times the energy of the first path in the CIR, where k is a number greater than 0. For example, k is a number greater than 0 and less than or equal to 1. The number of paths whose energy is greater than k times the energy of the first path in the CIR can reflect changes in the channel environment, so the AI ​​model can be adaptively adjusted based on changes in the channel environment. In addition, when the measurement result of the first parameter is transmitted between network elements, the overhead is low if the measurement result of the first parameter is the number of paths whose energy is greater than k times the energy of the first path in the CIR.

[0247] For the sake of brevity, below we will refer to the number of paths in the CIR whose energy is greater than k times the energy of the first path as the number of paths #A.

[0248] In this example, the correspondence may include a correspondence between the AI ​​model and the number of passes #A. In other words, different values ​​of the number of passes #A correspond to different AI models. Thus, the corresponding AI model can be learned based on the value of the number of passes #A obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0249] Specifically, the first network element performs channel measurement based on a reference signal to obtain the channel measurement result, and then obtains the number of paths #A through calculations based on the channel measurement result. For example, the first network element performs channel measurement based on a reference signal to obtain the CIR, and then obtains the number of paths #A through calculations based on the CIR.

[0250] The number of paths #A represents the number of sampled paths whose energy is greater than k times the energy of the first path. The first path refers to the first path through which the signal reaches the receiver, for example, a path detected based on an algorithm and considered to be the first path through which the signal reaches the receiver. The specific method of algorithmic detection is not limited. For example, an algorithmic detection method may be as follows: First, based on the strongest path threshold and the strongest path search window, the location of the strongest path among multiple paths is found in the time-domain channel response, and based on the location of the strongest path found, and based on the first path threshold and the first path search window, the location of the first path is searched. As another example, an algorithmic detection method may be as follows: The time-domain channel response is searched for the location of the first path directly based on the first path threshold and the first path search window.

[0251] For example, Table 4 shows an example of the correspondence between the AI ​​model and the number of passes #A. [Table 4]

[0252] Table 4 shows that different pass counts #A correspond to different AI models. For example, when the value of pass count #A is 128, the corresponding AI model is AI model #6. Therefore, the corresponding AI model is determined based on pass count #A obtained through actual measurements, and the corresponding AI model is compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0253] It should be understood that Table 4 is an example for illustrative purposes only. This is not limited to this. For example, the specific values ​​for the number of paths in Table 4 are merely examples. This is not limited to this. The values ​​for the number of paths may be replaced with other values ​​(for example, a specific value or a range of values).

[0254] Example 4: The first parameter includes the average power of multiple sampling points. Since the average power of multiple sampling points can reflect changes in the channel environment, the AI ​​model can be adaptively adjusted based on changes in the channel environment. In addition, calculating the average power of multiple sampling points is simple and easy to implement.

[0255] In this example, the correspondence may include the correspondence between an AI model and the average power of multiple sampling points. In other words, different average powers of multiple sampling points correspond to different AI models. Multiple sampling points may include all sampling points. Thus, the corresponding AI model can be learned based on the average power of multiple sampling points obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0256] Specifically, the first network element performs channel measurements based on a reference signal to obtain channel measurement results, and then obtains the average power of multiple sampling points through calculations based on the channel measurement results. For example, the first network element performs channel measurements based on a reference signal to obtain CIR, and then obtains the average power of multiple sampling points through calculations based on CIR.

[0257] The average power of multiple sampling points represents the average power of multiple sampling points in the received reference signal. The average power of multiple sampling points may be replaced by the average power of some sampling points or the average power of all sampling points; in other words, it may be understood that decisions may be made based on the average power of some sampling points or the average power of all sampling points. The number of sampling points is not limited.

[0258] For example, Table 5 shows an example of the correspondence between an AI model and the average power of multiple sampling points. [Table 5]

[0259] Table 5 shows that different average power values ​​from multiple sampling points correspond to different AI models. For example, when the average power value from multiple sampling points is P6, the corresponding AI model is AI model #6. Therefore, the corresponding AI model is determined based on the average power value from multiple sampling points obtained through actual measurements, and the corresponding AI model is compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0260] It should be understood that Table 5 is an example for illustrative purposes only. This is not limited to this. For example, the specific values ​​of average power for multiple sampling points in Table 5 are merely examples. This is not limited to this. The average power values ​​for multiple sampling points may be replaced with other values ​​(e.g., specific values ​​or ranges of values).

[0261] Example 5: The first parameter includes the LOS probability. Since the LOS probability can reflect changes in the channel environment, the AI ​​model can be adaptively adjusted based on changes in the channel environment. In addition, when transmitting the measurement result of the first parameter between network elements, the overhead is low if the measurement result of the first parameter is the LOS probability. In this example, the correspondence may include the correspondence between the AI ​​model and the LOS probability. In other words, different LOS probabilities correspond to different AI models. Thus, the corresponding AI model can be learned based on the LOS probability obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0262] For example, Table 6 shows an example of the correspondence between AI models and LOS probabilities. [Table 6]

[0263] Table 6 shows that different LOS probabilities correspond to different AI models. For example, when the LOS probability is 1 / 4, the corresponding AI model is AI model #2. Therefore, the corresponding AI model can be determined based on the LOS probability obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0264] It should be understood that Table 6 is an example for illustrative purposes only. This is not limited. For example, the specific values ​​of LOS probabilities in Table 6 are merely examples. This is not limited. The LOS probability values ​​may be replaced with other values ​​(e.g., specific values ​​or ranges of values).

[0265] Example 6: The first parameter includes SINR. Since SINR can reflect changes in the channel environment, the AI ​​model can be adaptively adjusted based on changes in the channel environment. In addition, when the measurement result of the first parameter is transmitted between network elements, the overhead is low if the measurement result of the first parameter is SINR.

[0266] In this example, the correspondence may include the correspondence between an AI model and a SINR. In other words, different SINRs correspond to different AI models. Thus, the corresponding AI model can be learned based on SINRs obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0267] Specifically, the first network element performs channel measurements based on a reference signal to obtain the channel measurement results, and then obtains the SINR through calculations based on the channel measurement results. For example, the first network element performs channel measurements based on a reference signal to obtain the CFR, and then obtains the SINR of the full band or subband of the CFR through calculations based on the CFR.

[0268] For example, the value of SINR may be, for example, -30, -20, -10, 0, 10, 20, or 30, and different values of SINR correspond to different AI models. For examples of the correspondence when the first parameter is SINR, refer to Tables 2 to 5. Details will not be explained again here.

[0269] Example 7: The first parameter includes RSRP. Since RSRP can reflect changes in the channel environment, the AI model can be adaptively adjusted based on changes in the channel environment. In addition, when transmitting the measurement result of the first parameter between network elements, if the measurement result of the first parameter is RSRP, the overhead is low.

[0270] In this example, the correspondence may include the correspondence between the AI model and RSRP. In other words, different RSRPs correspond to different AI models. Thus, the corresponding AI model can be learned based on the RSRP obtained through actual measurement, and by comparing the corresponding AI model with the current AI model, it is necessary to determine whether the current AI model needs to be adjusted. RSRP may be replaced with an interference level. In other words, different RSRPs correspond to different interference levels.

[0271] Specifically, the first network element performs channel measurement based on a reference signal to obtain a channel measurement result, and obtains RSRP through calculations based on the channel measurement result. For example, the first network element performs channel measurement based on a reference signal to obtain CFR, and obtains the RSRP of the full band or sub-band of CFR through calculations based on CFR.

[0272] For example, the value of RSRP may be, for example, -30, -20, -10, 0, 10, 20, or 30, and different values of RSRP correspond to different AI models. For examples of the correspondence when the first parameter is RSRP, refer to Tables 2 to 5. Details will not be explained again here.

[0273] Example 8: The first parameter includes RSSI. Since RSSI can reflect changes in the channel environment, the AI ​​model can be adaptively adjusted based on changes in the channel environment. In addition, when transmitting the measurement results of the first parameter between network elements, the overhead is low if the measurement results of the first parameter are RSSI.

[0274] In this example, the correspondence may include the correspondence between an AI model and an RSSI. In other words, different RSSIs correspond to different AI models. Thus, the corresponding AI model can be learned based on RSSIs obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0275] Specifically, the first network element performs channel measurements based on a reference signal to obtain the channel measurement results, and then obtains the RSSI through calculations based on the channel measurement results. For example, the first network element performs channel measurements based on a reference signal to obtain the CFR, and then obtains the RSSI of the full band or subband of the CFR through calculations based on the CFR.

[0276] For example, the RSSI value can be -30, -20, -10, 0, 10, 20, or 30, and different RSSI values ​​correspond to different AI models. See Tables 2 to 5 for examples of correspondences when the first parameter is RSSI. Further details will not be explained here.

[0277] Example 9: The first parameter includes the Rice coefficient. Since the Rice coefficient can reflect changes in the channel environment, the AI ​​model can be adaptively adjusted based on changes in the channel environment. In addition, when transmitting the measurement results of the first parameter between network elements, the overhead is low if the measurement result of the first parameter is the Rice coefficient.

[0278] In this example, the correspondence may include the correspondence between an AI model and a Rice coefficient. In other words, different Rice coefficients correspond to different AI models. Thus, the corresponding AI model can be learned based on Rice coefficients obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0279] Specifically, the first network element performs channel measurement based on a reference signal to obtain the channel measurement result, and then obtains the Rice coefficient through calculations based on the channel measurement result. For example, the first network element performs channel measurement based on a reference signal to obtain the CIR, and then obtains the Rice coefficient through calculations based on the CIR. In another example, the first network element performs channel measurement based on a reference signal to obtain the CFR, and then obtains the Rice coefficient through calculations based on the CFR.

[0280] The Rice coefficient represents the ratio of power in a LOS path to power in a NLOS path across multiple paths. The Rice coefficient is related to the geometric location on the map. Specifically, if the first network element moves to a different location, the Rice coefficient also changes. In this case, the power of the multiple paths also changes depending on the location.

[0281] Generally, channel propagation conditions are classified into LOS and NLOS. An LOS scenario includes at least one LOS path, i.e., a pure LOS, and X additional NLOS paths, where X is a non-negative integer. In an LOS scenario, the energy of the LOS paths is higher than the total energy of the NLOS paths. Generally, the Rice coefficient is used to define the ratio of the power of LOS paths to the power of NLOS paths. The power of an NLOS path represents the sum of the powers of all NLOS paths. An NLOS scenario includes NLOS paths but does not include LOS paths. From the above explanation, it can be seen that, generally, the larger the Rice coefficient, the higher the energy of the LOS paths relative to the energy of NLOS, and the lower the degree of NLOS.

[0282] For example, Table 7 shows an example of the correspondence between AI models and the Rice coefficient. [Table 7]

[0283] Table 7 shows that different Rice coefficients correspond to different AI models. For example, when the Rice coefficient is 5, the corresponding AI model is AI model #6. Therefore, the corresponding AI model can be learned based on the Rice coefficient obtained through actual measurements, and the corresponding AI model can be compared to the current AI model to determine whether the current AI model needs to be adjusted.

[0284] It should be understood that Table 7 is an example for illustrative purposes only. This is not limited to this. For example, in Table 7, more Rice coefficients may be divided.

[0285] The above explains each piece of information. The first parameter may alternatively include at least two of the above pieces of information. A brief example is listed below for illustrative purposes.

[0286] Example 10: The first parameter includes the number of passes #A and the inter-site synchronization error.

[0287] In this example, the correspondence could be between the AI ​​model, the number of passes #A, and the inter-site synchronization error.

[0288] For example, Table 8 shows an example of the correspondence between the AI ​​model, the number of passes #A, and the inter-site synchronization error.

[0289] Table 8 is used as an example. For example, if the AI model currently used for positioning is AI model #1, the value of the number of paths #A obtained through actual measurement belongs to the second value range, and the inter-site synchronization error is at level 2, this indicates that it is necessary to adjust the AI model. Specifically, it is necessary to adjust AI model #1 to AI model #2. In another example, if the AI model currently used for positioning is AI model #1, the value of the number of paths #A obtained through actual measurement belongs to the second value range, and the inter-site synchronization error is at level 5, this indicates that it is not necessary to adjust the AI model.

Table 8

[0290] It can be understood that Table 8 is an example for explanation. This is not limiting.

[0291] Optionally, the first network element receives first indication information, where the first indication information indicates that a first parameter is used to determine whether to adjust the AI model. Based on this, the first network element may learn that it is necessary to measure the first parameter based on the first indication information, and may determine whether to adjust the AI model by using the measurement result of the first parameter.

[0292] The first indication information indicating that a first parameter is used to determine whether to adjust the AI model may be replaced with first indication information instructing to measure the first parameter, or may be replaced with first indication information instructing that the first parameter reflects the current channel environment (for example, the degree of NLOS of the current channel).

[0293] It can be understood that the above description is an exemplary description. This is not limiting. For example, the first network element may also determine a specific type of the first parameter to be measured based on settings predefined or pre-stored in the protocol.

[0294] Optionally, the first network element may further transmit parameters that can be measured by the first network element to the second network element. For example, the second network element may select one or more parameters that can be measured by the first network element as first parameters and, by using first instruction information, instruct that the first parameters be used to determine whether to tune the AI ​​model. For example, the parameters that can be measured by the first network element include the number of passes #A and the LOS probability. The first network element transmits information to the second network element, which indicates that the parameters that can be measured by the first network element include the number of passes #A and the LOS probability. Based on this information, the second network element transmits first instruction information to the first network element, which indicates that the number of passes #A and / or the LOS probability be used to determine whether to tune the AI ​​model.

[0295] For example, in this embodiment of the present application, the decision of whether or not to adjust the AI ​​model includes the following two solutions:

[0296] Solution 1: The first network element determines whether to adjust the AI ​​model based on the measurement results and correspondence of the first parameter.

[0297] The method by which the first network element obtains the correspondence is not limited.

[0298] For example, the first network element may store the correspondence locally.

[0299] As another example, the first network element receives a correspondence. For example, the first network element receives a correspondence from another network element (for example, the second network element). For example, the other network element maintains the correspondence and actively transmits it to the first network element. In another example, the other network element maintains the correspondence and, after receiving a request from the first network element, transmits the correspondence to the first network element.

[0300] Let's use Table 1 as an example. For example, if the current AI model is AI model #2, and the measurement result of the first parameter obtained by the first network element based on the channel measurement result is N3, then the first network element will determine that the AI ​​model needs to be adjusted, specifically that AI model #2 needs to be switched to AI model #3. In another example, if the current AI model is AI model #4, and the measurement result of the first parameter obtained by the first network element based on the channel measurement result is N6, then the AI ​​model library does not have an AI model corresponding to N6, so the first network element will determine that AI model #4 needs to be updated.

[0301] According to Solution 1, when the first network element optionally determines that the AI ​​model needs to be adjusted, the first network element obtains the adjusted AI model.

[0302] In the first possible case, if the first network element determines, based on the measurement results and correspondence of the first parameter, to adjust the first type of AI model used for positioning from the first AI model to the second AI model, for example, the first network element sends request information to the second network element, where the request information is used to request the second AI model; the first network element receives a response to the request information, where the response to the request information includes information about the second AI model. In this case, the channel measurement results are the inputs to the first type of AI model, and the correspondences include the correspondence between the first type of AI model and the value of the first parameter. Specifically, the first network element determines, based on the measurement results of the first parameter and the correspondence between the first type of AI model and the value of the first parameter, to adjust the first type of AI model from the first AI model to the second AI model.

[0303] In this case, the second network element may represent a network element that stores the second AI model, or it may represent a network element that can retrieve the second AI model. For example, requesting the second AI model using request information includes requesting that the AI ​​model be updated using request information. The updated AI model is the second AI model.

[0304] Information regarding the second AI model may include, for example, information that can identify the second AI model or information that can be used to obtain the second AI model. For example, information regarding the second AI model may include the identifier of the second AI model and / or the neural network parameters of the second AI model. The neural network parameters may include at least one of the following: namely, the number of layers in the neural network, the width of the neural network, the weights of the neurons, or all or some of the parameters in the activation function of the neurons. Further details are not described below.

[0305] If the first network element stores the second AI model, it can be understood that the first network element may directly adjust the first AI model to the second AI model, in other words, it does not need to send request information to the second network element.

[0306] In a second possible case, if the first network element determines, based on the measurement results and correspondence of the first parameter, to adjust the second type of AI model used for positioning from the third AI model to the fourth AI model, for example, the first network element transmits second instruction information to the second network element, where the second instruction information instructs to adjust the second type of AI model from the third AI model to the fourth AI model.

[0307] In this case, the first network element may learn that the current AI model of type 2 is the third AI model. For example, based on the fact that the current AI model of type 1 is the first AI model, and based on the match or correspondence between the second type AI model and the first type AI model, the first network element may learn that the AI ​​model that matches the first AI model is the third AI model. In another example, the second network element informs the first network element that the current AI model of type 2 is the third AI model.

[0308] In a possible implementation, the first network element directly determines, based on the measurement results and correspondence of the first parameter, to adjust the second type AI model used for positioning from the third AI model to the fourth AI model, in which case the channel measurement results are the inputs to the first type AI model, the outputs of the first type AI model are the inputs to the second type AI model, and the correspondences include the correspondence between the second type AI model and the value of the first parameter. In this case, if the AI ​​model configured within the first network element is the first type AI model, and the first type AI model and the second type AI model are matched for use, the first network element may optionally determine to perform positioning by using a fifth AI model that matches the fourth AI model, based on the match or correspondence between the second type AI model and the first type AI model. The above description is illustrative and not limited thereto. For example, the first network element may alternatively determine whether it is necessary to adjust the first type AI model based on the correspondence between the first type AI model and the value of the first parameter.

[0309] In another possible implementation, the first network element determines, based on the measurement results and correspondence of the first parameter, to adjust the first type of AI model used for positioning from the first AI model to the fifth AI model, where the correspondence includes the correspondence between the first type of AI model used for positioning and the value of the first parameter. Furthermore, based on the match or correspondence between the second type of AI model and the first type of AI model, the first network element learns that the second type of AI model that matches the fifth AI model among the first type of AI models is the fourth AI model. Thus, the first network element determines to adjust the second type of AI model from the third AI model to the fourth AI model.

[0310] Solution 2: The first network element sends the measurement results of the first parameter to the second network element, and the second network element decides whether to adjust the AI ​​model based on the measurement results of the first parameter and their corresponding relationships.

[0311] The method by which the second network element obtains the correspondence is not limited.

[0312] For example, the second network element may store the correspondence locally.

[0313] As another example, the second network element receives a correspondence. For example, the second network element receives a correspondence from another network element (for example, the first network element). For example, the other network element maintains the correspondence and actively transmits the correspondence to the second network element. As yet another example, the other network element maintains the correspondence and, after receiving a request from the second network element, transmits the correspondence to the second network element.

[0314] For example, after receiving the measurement result of the first parameter transmitted by the first network element, the second network element determines whether the AI ​​model needs to be adjusted based on the correspondence and the AI ​​model currently being used by the first network element. Using Table 1 as an example, if the AI ​​model currently being used by the first network element is AI model #1, and the measurement result of the first parameter received by the second network element from the first network element is N3, then the second network element will determine that the AI ​​model needs to be adjusted, specifically that AI model #1 needs to be switched to AI model #3.

[0315] The method by which the second network element learns the AI ​​model currently used by the first network element is not limited. For example, the first network element sends instruction information to the second network element, where the instruction information indicates the AI ​​model currently used by the first network element. The instruction information and the measurement results of the first parameter may be carried by one signaling or by different signaling, but this is not limited. As another example, the second network element knows the AI ​​model currently used by the first network element. For example, the AI ​​model currently used by the first network element is sent to the first network element by the second network element.

[0316] For example, the transmission of the measurement result of the first parameter from the first network element to the second network element may include the following possible implementations:

[0317] In possible implementations, the transmission of the measurement result of a first parameter from a first network element to a second network element includes: when the measurement result of the first parameter satisfies a specified condition, the first network element transmits the measurement result of the first parameter to the second network element. The specified condition may, but is not limited to, a preset value or a range of preset values. The specified condition may, but is not limited to, be predefined, pre-configured, or configured.

[0318] In another possible implementation, the transmission of the measurement result of the first parameter from the first network element to the second network element includes the periodic transmission of the measurement result of the first parameter from the first network element to the second network element. The periodicity value of the transmission of the measurement result of the first parameter by the first network element is not limited. For example, the periodicity value may be predefined or preconfigured.

[0319] The two implementations described above are examples for illustrative purposes only. This is not an limitation. For example, after obtaining the measurement result of the first parameter, the first network element may each time send the measurement result of the first parameter to the second network element.

[0320] According to Solution 2, when it is determined that the AI ​​model needs to be adjusted, the second network element instructs the first network element to provide the adjusted AI model.

[0321] In the first possible case, if the second network element determines, based on the measurement results and correspondence of the first parameter, to adjust the first type of AI model used for positioning from the first AI model to the second AI model, for example, the second network element transmits third instruction information to the first network element, where the third instruction information instructs to adjust the first type of AI model used for positioning from the first AI model to the second AI model. Optionally, the third instruction information includes information about the second AI model.

[0322] In possible implementations, the second network element directly determines, based on the measurement results and correspondences of the first parameter, whether to adjust the first type AI model used for positioning from the first AI model to the second AI model, where the channel measurement results are the inputs to the first type AI model and the correspondences include the correspondence between the first type AI model and the values ​​of the first parameter. In this case, if the AI ​​model configured within the second network element is the second type AI model, and the first type AI model and the second type AI model are matched for use, the second network element optionally uses the AI ​​model that matches the second AI model to perform positioning, based on the match or correspondence between the second type AI model and the first type AI model.

[0323] In another possible implementation, the second network element determines, based on the measurement results and correspondence of the first parameter, to adjust the second type AI model used for positioning from the third AI model to the fourth AI model, where the channel measurement results are the inputs to the first type AI model, the outputs of the first type AI model are the inputs to the second type AI model, and the correspondences include the correspondence between the second type AI model and the values ​​of the first parameter. Furthermore, the second network element learns that the first type AI model that matches the fourth AI model among the second type AI models is the second AI model, based on the match or correspondence between the second type AI model and the first type AI model. Thus, the second network element determines to adjust the first type AI model used for positioning from the first AI model to the second AI model.

[0324] In the second possible case, if the second network element determines, based on the measurement results and correspondence of the first parameter, to adjust the second type of AI model used for positioning from the third AI model to the fourth AI model, for example, the second network element transmits fourth instruction information to the first network element, where the fourth instruction information instructs to adjust the second type of AI model used for positioning from the third AI model to the fourth AI model, the channel measurement results are the input to the first type of AI model, the output of the first type of AI model is the input to the second type of AI model, and the correspondence includes the correspondence between the second type of AI model and the value of the first parameter.

[0325] For example, if the fourth instruction instructs the system to switch or update the second type AI model used for positioning from the third AI model to the fourth AI model, this can be replaced with the fourth instruction instructing the system to switch or update the second type AI model used for positioning.

[0326] Optionally, in response to the fourth instruction information, the first network element performs positioning by using the fifth AI model. The fifth AI model is an AI model used for matching with the fourth AI model.

[0327] The above describes two solutions, which are variations of the aforementioned solutions and are applicable to embodiments of this application. For example, both the first and second network elements determine whether to adjust the AI ​​model based on the measurement results and correspondences of the first parameter. If the determination results are consistent, they decide to adjust the AI ​​model; otherwise, they decide not to adjust the AI ​​model. The decision executor may be the first network element, the second network element, or another network element different from the first and second network elements. The decision executor may obtain one or more correspondence information used for the determination from the other end or another end, such as information about the AI ​​model to switch to, or information about whether to switch or update. As another example, the first network element determines whether to adjust a first type of AI model based on the measurement results and correspondences of the first parameter, and the second network element determines whether to adjust a second type of AI model based on the measurement results and correspondences of the first parameter. As yet another example, another network element determines whether to adjust the AI ​​model based on the measurement results and correspondences of the first parameter and notifies the first and / or second network elements of the determination result. The other network element may be, for example, a different network element from the first and second network elements, such as a core network element, a network device, a third party, or an AI network element.

[0328] Optionally, in embodiments of this application, the measurement result of the first parameter includes a measurement result obtained at least at one point in time and / or a measurement result provided by at least one first network element.

[0329] In a possible implementation, the measurement result for the first parameter includes measurement results obtained at least at two time points.

[0330] For example, the first network element receives a reference signal at time point 1, performs a channel measurement based on the reference signal to obtain the channel measurement result, and obtains the measurement result for the first parameter (denoted as measurement result #1 for distinction) based on the channel measurement result; the first network element receives a reference signal at time point 2, performs a channel measurement based on the reference signal to obtain the channel measurement result, and obtains the measurement result for the first parameter (denoted as measurement result #2 for distinction) based on the channel measurement result. Measurement result #1, measurement result #2 and their correspondence are used to determine whether to adjust the AI ​​model.

[0331] Let's use Solution 1 above as an example. The first network element determines whether to adjust the AI ​​model based on measurement result #1, measurement result #2, and the correspondence. For example, the first network element performs an averaging process on measurement result #1 and measurement result #2, and then determines whether to adjust the AI ​​model based on the result of the averaging process and the correspondence.

[0332] Let's use Solution 2 above as an example. The second network element decides whether to adjust the AI ​​model based on measurement result #1, measurement result #2, and the correspondence. For example, the first network element performs an averaging process on measurement result #1 and measurement result #2 and sends the result of the averaging process to the second network element. The second network element decides whether to adjust the AI ​​model based on the result of the averaging process and the correspondence. As another example, the first network element sends measurement result #1 and measurement result #2 to the second network element, the second network element performs an averaging process on measurement result #1 and measurement result #2, and the second network element decides whether to adjust the AI ​​model based on the result of the averaging process and the correspondence.

[0333] In another possible implementation, the measurement result of the first parameter includes the measurement results provided by at least two first network elements.

[0334] For example, the first network element receives a reference signal, performs a channel measurement based on the reference signal to obtain the channel measurement result, and obtains the measurement result of the first parameter (referred to as measurement result #3 for distinction) based on the channel measurement result. The first network element #2 receives a reference signal, performs a channel measurement based on the reference signal to obtain the channel measurement result, and obtains the measurement result of the first parameter (referred to as measurement result #4 for distinction) based on the channel measurement result. The first network element #3 receives a reference signal, performs a channel measurement based on the reference signal to obtain the channel measurement result, and obtains the measurement result of the first parameter (referred to as measurement result #5 for distinction) based on the channel measurement result. Measurement result #3, measurement result #4, measurement result #5 and their corresponding relationships are used to determine whether to adjust the AI ​​model.

[0335] Let's use Solution 2 above as an example. The first network element sends measurement result #3 to the second network element, the first network element #2 sends measurement result #4 to the second network element, the first network element #3 sends measurement result #5 to the second network element, and the second network element decides whether to adjust the AI ​​model based on measurement result #3, measurement result #4, measurement result #5 and the correspondence.

[0336] In one example, the second network element determines that the AI ​​model needs to be adjusted based on at least P measurement results; otherwise, it determines that the AI ​​model does not need to be adjusted. P is an integer greater than or equal to 1. The value of P may be predefined or preconfigured, but is not limited to this.

[0337] For example, the second network element determines whether to adjust the AI ​​model based on measurement result #3 and its correspondence, the second network element determines whether to adjust the AI ​​model based on measurement result #4 and its correspondence, and the second network element determines whether to adjust the AI ​​model based on measurement result #5 and its correspondence. If the second network element determines that the AI ​​model needs to be adjusted based on at least two measurement results, it may decide that the AI ​​model needs to be adjusted; otherwise, it decides that the AI ​​model does not need to be adjusted. In this example, the value of P is 2.

[0338] It should be understood that the above description is illustrative and not limited to. For example, the second network element may alternatively perform an averaging process on measurement results #3, #4, and #5, and then decide whether to adjust the AI ​​model based on the results of the averaging process and the corresponding relationships. Alternatively, the first network element may alternatively provide the second network element with channel measurement results, and the second network element may obtain measurement results for the first parameter based on the channel measurement results and then make a decision.

[0339] To facilitate understanding, the following describes the procedures applicable to Solutions 1 and 2 described above, using the interaction between the first and second network elements as an example, with reference to Figures 9 and 10. For details not described below, please refer to the description of Method 800. Further details are not provided below.

[0340] Figure 9 is a schematic flowchart of a communication method 900 according to an embodiment of the present application. Method 900 is applicable to Solution 1 described above and is applicable to a scenario in which AI model inference is performed on the first network element side and storage of the AI ​​model library and / or training of the AI ​​model is performed on the second network element side. Method 900 shown in Figure 9 may include the following steps:

[0341] 901: The first network element performs channel measurement based on a reference signal and obtains the channel measurement result.

[0342] In one example, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal. In another example, the first network element is a network device or a positioning device, or a component of a network device or a positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0343] For example, if the first network element is a terminal device and the second network element is a network device or a positioning device, the terminal device may receive a reference signal (e.g., downlink PRS) from the network device, perform channel measurements based on the reference signal, and obtain channel measurement results.

[0344] In another example, if the first network element is a network device and the second network element is a terminal device, the network device may receive a reference signal (e.g., an uplink positioning SRS) from the terminal device, perform channel measurements based on the reference signal, and obtain channel measurement results.

[0345] In another example, if both the first and second network elements are terminal devices, the first network element may receive a reference signal (e.g., SL-PRS) from the second network element, perform a channel measurement based on the reference signal, and obtain the channel measurement result.

[0346] Optionally, method 900 further includes step 902.

[0347] 902: The first network element inputs the channel measurement results into AI model #1 to perform positioning.

[0348] In this embodiment of the present application, the AI ​​model currently used by the first network element is assumed to be AI model #1.

[0349] AI model #1 may be used for direct positioning or for auxiliary positioning. See the description in Method 800 for details.

[0350] Optionally, method 900 further includes step 903.

[0351] 903: The second network element transmits the first instruction information and / or correspondence to the first network element.

[0352] For example, the second network element is a network device, a core network device, a positioning device, or a server or cloud device in an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device in an OTT system, such as a chip or circuit.

[0353] For the solutions and corresponding relationships related to the first instruction information, please refer to the explanation in Method 800.

[0354] 904: The first network element determines the measurement result of the first parameter based on the channel measurement result, and decides whether to adjust AI model #1 based on the measurement result of the first parameter and its correspondence.

[0355] For example, if method 900 includes step 903, in which step 903 a second network element transmits first instruction information and correspondence to the first network element, the first network element may learn, based on the first instruction information, that the measurement result of the first parameter is used to determine whether to adjust AI model #1. Thus, the first network element determines the measurement result of the first parameter based on the channel measurement result and determines whether to adjust AI model #1 based on the measurement result of the first parameter and the correspondence received in step 903.

[0356] As another example, if method 900 includes step 903, in which step 903 a second network element transmits first instruction information to the first network element, the first network element may learn, based on the first instruction information, that the measurement result of the first parameter is used to determine whether to adjust AI model #1. Thus, the first network element determines the measurement result of the first parameter based on the channel measurement result and determines whether to adjust AI model #1 based on the measurement result of the first parameter and the correspondence. The correspondence may be acquired in advance, for example, predefined, or preconfigured.

[0357] As another example, if method 900 includes step 903, in which step 903 a second network element transmits a correspondence to the first network element, the first network element determines whether to adjust AI model #1 based on the measurement result of the first parameter and the correspondence received in step 903. By default, the first network element may determine, based on a setting predefined or pre-stored in the protocol, whether the measurement result of the first parameter is used to determine whether to adjust AI model #1.

[0358] For a specific method of determining the first parameter and whether the first network element adjusts the AI ​​model based on the measurement results of the first parameter and their corresponding relationships, please refer to the explanation in Method 800.

[0359] In step 904, if the first network element determines to adjust AI model #1 based on the measurement results and correspondence of the first parameter, for example, to adjust AI model #1 to AI model #2, then, in one example, method 900 further includes steps 905 to 907.

[0360] 905: The first network element sends request information to the second network element, which is used to request AI model #2.

[0361] 906: The second network element sends information about AI model #2 to the first network element.

[0362] For example, the second network element sends AI model #2 to the first network element. In another example, the second network element sends the index of AI model #2 to the first network element, and the first network element can then identify AI model #2 based on its index.

[0363] 907: The first network element performs positioning by using AI model #2.

[0364] If, in step 904, the first network element determines that it does not need to adjust AI model #1 based on the measurement results and correspondence of the first parameter, method 900 further includes step 908, but does not include steps 905 to 907.

[0365] 908: The first network element continues to perform positioning by using AI model #1.

[0366] According to the above technical solution, the first network element can determine whether the AI ​​model needs to be adjusted based on the input features of the AI ​​model. In particular, after performing channel measurements based on a reference signal, the first network element can obtain the channel measurement results, obtain the measurement results of the first parameter based on the channel measurement results, and determine whether the AI ​​model needs to be adjusted based on the measurement results of the first parameter. It is conceivable that changes in the channel environment may affect positioning accuracy. For example, when using a single AI model, changes in the channel environment may cause a deterioration in positioning performance. Since the measurement results of the first parameter are obtained based on the channel measurement results, the measurement results of the first parameter can reflect the current channel environment. Thus, according to the above solution, the AI ​​model configured on the first network element side can be adjusted in a timely manner based on changes in the channel state, thereby improving the positioning accuracy of the AI ​​model.

[0367] Figure 10 is a schematic flowchart of communication method 1000 according to another embodiment of the present application. Method 1000 is applicable to solution 2 described above and is applicable to a scenario in which AI model inference is performed on the first network element side and storage of the AI ​​model library and / or training of the AI ​​model is performed on the second network element side. Method 1000 shown in Figure 10 may include the following steps:

[0368] 1001: The first network element performs channel measurement based on a reference signal and obtains the channel measurement result.

[0369] In one example, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal. In another example, the first network element is a network device or a positioning device, or a component of a network device or a positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0370] Optionally, method 1000 further includes step 1002.

[0371] 1002: The first network element inputs the channel measurement results into AI model #1 to perform positioning.

[0372] For steps 1001 and 1002, please refer to steps 901 and 902.

[0373] Optionally, method 1000 further includes step 1003.

[0374] 1003: The second network element transmits the first instruction information to the first network element.

[0375] For example, the second network element is a network device, a core network device, a positioning device, or a server or cloud device in an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device in an OTT system, such as a chip or circuit.

[0376] For solutions related to the first instruction information, please refer to the description of Method 800.

[0377] 1004: The first network element transmits the measurement result of the first parameter to the second network element, where the measurement result of the first parameter is obtained based on the channel measurement result.

[0378] 1005: The second network element determines whether to adjust AI model #1 based on the measurement results and correspondence of the first parameter.

[0379] For a specific method of determining whether the second network element adjusts the AI ​​model based on the measurement results and correspondence of the first parameter, please refer to the explanation in Method 800.

[0380] In step 1005, if the second network element determines to adjust AI model #1 based on the measurement results and correspondence of the first parameter, for example, to adjust AI model #1 to AI model #2, then, in one example, method 1000 further includes steps 1006 and 1007.

[0381] 1006: The second network element transmits information about AI model #2 to the first network element.

[0382] 1007: The first network element performs positioning by using AI model #2.

[0383] In step 1005, if the second network element determines that it does not need to adjust AI model #1 based on the measurement results and correspondence of the first parameter, then, for example, method 1000 further includes steps 1008 and 1009, but does not include steps 1006 and 1007 described above.

[0384] 1008: The second network element sends notification information to the first network element to inform it to continue performing positioning using AI model #1.

[0385] 1009: The first network element continues to perform positioning by using AI model #1.

[0386] It should be understood that the above explanation is illustrative and not limited to. For example, if the second network element determines that it does not need to adjust AI model #1 based on the measurement results and correspondence of the first parameter, the second network element does not send a response to the first network element for the measurement results of the first parameter. If the first network element does not receive a response from the second network element within a certain period (referred to as period #1 for distinction), the first network element continues to perform positioning by default using AI model #1. For example, the start of period #1 may be the time when the first network element sends the measurement results of the first parameter, and the duration of period #1 may be predefined or estimated based on historical circumstances. This is not limited to. For example, period #1 may be implemented using a timer.

[0387] According to the above technical solution, the second network element can determine whether the AI ​​model needs to be adjusted based on the input features of the AI ​​model. In particular, after performing channel measurements based on a reference signal, the first network element obtains the measurement results of the first parameter based on the channel measurement results and transmits the measurement results of the first parameter to the second network element. In this way, the second network element can determine whether the AI ​​model needs to be adjusted based on the measurement results of the first parameter. It is conceivable that changes in the channel environment may affect positioning accuracy. For example, when using a single AI model, changes in the channel environment may cause a deterioration in positioning performance. Since the measurement results of the first parameter are obtained based on the channel measurement results, the measurement results of the first parameter can reflect the current channel environment. Thus, according to the above solution, the AI ​​model configured on the first network element side can be adjusted in a timely manner based on changes in the channel state, thereby improving the positioning accuracy of the AI ​​model.

[0388] For ease of understanding, the following describes the procedure applicable to embodiments of this application, using an example of interaction between a first network element and a second network element, where the first parameter is the number of paths #A (i.e., the number of paths whose energy is greater than k times the energy of the first path in the CIR), with reference to Figures 11 and 12. For details not described below, please refer to the description of Method 800. Details are not described below.

[0389] Figure 11 is a schematic flowchart of a communication method 1100 according to another embodiment of the present application. Method 1100 is applicable to a scenario in which AI model inference is performed on the first network element side and storage of an AI model library and / or training of an AI model is performed on the second network element side. Method 1100 shown in Figure 11 may include the following steps:

[0390] 1101: The first network element performs channel measurement based on a reference signal and obtains CIR.

[0391] In one example, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal. In another example, the first network element is a network device or a positioning device, or a component of a network device or a positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0392] Optionally, method 1100 further includes step 1102.

[0393] 1102: The first network element inputs the CIR to AI model #1 to perform positioning.

[0394] Optionally, method 1100 further includes step 1103.

[0395] 1103: The second network element transmits the first instruction information and / or correspondence to the first network element.

[0396] The first instruction indicates that the number of passes #A will be used to determine whether to adjust the AI ​​model, or that the number of passes #A should be calculated, or that the number of passes #A should reflect the current channel environment (e.g., the degree of NLOS of the current channel).

[0397] For example, the second network element is a network device, a core network device, a positioning device, or a server or cloud device in an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device in an OTT system, such as a chip or circuit.

[0398] For steps 1101-1103, please refer to the explanation for steps 901-903.

[0399] 1104: The first network element calculates the number of paths #A based on CIR.

[0400] Optionally, method 1100 further includes step 1105.

[0401] 1105: The first network element decides to adjust AI model #1 to AI model #2 based on the number of paths #A and the correspondence.

[0402] The correspondence is the relationship between the AI ​​model and the number of passes #A.

[0403] Table 3 of Method 800 is used as an example. If the number of paths #A calculated by the first network element is 8, the first network element decides to adjust AI model #1 to AI model #2.

[0404] In this embodiment of the present application, for example, it is necessary to adjust AI model #1 to AI model #2.

[0405] 1106: The first network element sends request information to the second network element, which is used to request path number #A or AI model #2.

[0406] Where possible, if method 1100 includes step 1105, in step 1106, the first network element sends request information to the second network element, where the request information is used to request AI model #2.

[0407] In another possible case, if method 1100 does not include step 1105, in step 1106, the first network element sends a pass number #A to the second network element, where the pass number #A may be used by the second network element to determine whether to adjust AI model #1.

[0408] 1107: The second network element sends information about AI model #2 to the first network element.

[0409] If possible, in step 1106, when the first network element sends the number of passes #A to the second network element, the second network element determines whether to adjust AI model #1 based on the number of passes #A and the correspondence. If the second network element determines, based on the number of passes #A and the correspondence, that AI model #1 needs to be adjusted to AI model #2, it sends AI model #2 to the first network element.

[0410] In another possible case, if in step 1106 the first network element sends request information to the second network element that is used to request AI model #2, the second network element responds to the request information by sending AI model #2 to the first network element.

[0411] 1108: The first network element performs positioning by using AI model #2.

[0412] According to the above technical solution, whether or not the AI ​​model needs to be adjusted can be determined based on the measurement results of channel measurements, such as the number of paths #A, and their corresponding relationships. It is conceivable that changes in the channel environment may affect positioning accuracy. For example, when using a single AI model, changes in the channel environment may cause a deterioration in positioning performance. The number of paths #A can reflect the current channel environment, such as the degree of NLOS. Thus, according to the above solution, the AI ​​model configured on the first network element side can be adjusted in a timely manner based on changes in the channel state, thereby improving the positioning accuracy of the AI ​​model.

[0413] Figure 12 is a schematic flowchart of communication method 1200 according to another embodiment of the present application. Method 1200 is applicable to scenarios in which AI model inference, storage of an AI model library, and / or training of an AI model are deployed on both the first and second network element sides. In Method 1200, a first type of AI model is deployed on the first network element, a second type of AI model is deployed on the second network element, and the first and second type of AI models are used together to perform positioning. In other words, the channel measurement result is the input to the first type of AI model, and the output of the first type of AI model is the input to the second type of AI model. Method 1200 shown in Figure 12 may include the following steps:

[0414] 1201: The first network element performs channel measurement based on a reference signal and obtains CIR.

[0415] In one example, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal. In another example, the first network element is a network device or a positioning device, or a component of a network device or a positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0416] Optionally, method 1200 further includes steps 1102 to 1204.

[0417] 1202: The first network element inputs CIR to AI model #1 to perform auxiliary positioning and obtain auxiliary positioning information.

[0418] For information regarding auxiliary positioning, please refer to the explanation in Method 800.

[0419] 1203: The first network element transmits auxiliary positioning information to the second network element.

[0420] For example, the second network element is a network device, a core network device, a positioning device, or a server or cloud device in an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device in an OTT system, such as a chip or circuit.

[0421] 1204: The second network element inputs auxiliary positioning information into AI model #3 to perform direct positioning.

[0422] AI model #3 may be an AI model used for matching with AI model #1. In AI model #3, the input is the output of AI model #1, and the output is the position information of the first network element.

[0423] Optionally, method 1200 further includes step 1205.

[0424] 1205: The second network element transmits the first instruction information and / or correspondence to the first network element.

[0425] The correspondence may be between the first type of AI model and the number of paths #A. The second network element provides a correspondence to the first network element, and as a result, 1 Based on their correspondences, the network elements can determine whether or not the first AI model needs to be adjusted.

[0426] 1206: The first network element calculates the number of paths #A based on CIR.

[0427] Optionally, method 1200 further includes step 1207.

[0428] 1207: The first network element decides to adjust AI model #1 to AI model #2 based on the number of paths #A and the correspondence.

[0429] The correspondence may be the correspondence between the first type AI model and the value of path number #A. In particular, in step 1207, the first network element decides to adjust the first type AI model from AI model #1 to AI model #2 based on the measurement result of path number #A and the correspondence between the first type AI model and the value of path number #A.

[0430] The first network element may locally store the correspondence between the first type of AI model and the value of the number of passes #A, or it may receive the correspondence between the first type of AI model and the value of the number of passes #A from the second network element (e.g., step 1205). This is not limited to this.

[0431] 1208: The first network element sends request information to the second network element, which is used to request path number #A or AI model #2.

[0432] In a possible implementation, if in step 1207 the first network element determines to adjust AI model #1 to AI model #2 based on the number of paths #A and the correspondence, then in step 1208 the first network element sends the number of paths #A to the second network element.

[0433] In another possible implementation, if in step 1207 the first network element determines to adjust AI model #1 to AI model #2 based on the number of paths #A and the correspondence, then in step 1208 the first network element sends request information to the second network element to be used to request AI model #2.

[0434] 1209: The second network element adjusts AI model #3 to AI model #4 and performs direct positioning using AI model #4.

[0435] In a possible implementation, in step 1208, the first network element sends the number of paths #A to the second network element, and the second network element decides to adjust AI model #3 to AI model #4 based on the correspondence between the second type of AI model and the value of the number of paths #A. The second network element may store the correspondence between the second type of AI model and the value of the number of paths #A locally, or it may receive the correspondence between the first type of AI model and the value of the number of paths #A from another network element (e.g., the first network element). This is not limited to this.

[0436] In another possible implementation, in step 1208, the first network element transmits request information for AI model #2 to the second network element, and the second network element adjusts AI model #3, which matches AI model #1, to AI model #4, which matches AI model #2, based on the request information and the matching or correspondence relationship between the AI ​​models configured in the first network element and the AI ​​models configured in the second network element.

[0437] Optionally, method 1200 further includes step 1210.

[0438] 1210: The second network element transmits the fourth instruction information to the first network element.

[0439] The fourth instruction indicates that the AI ​​model composed of the second network elements will be adjusted, or the fourth instruction indicates that the AI ​​model composed of the second network elements will be adjusted from AI model #3 to AI model #4.

[0440] 1211: The first network element performs auxiliary positioning by using AI model #2.

[0441] In a possible implementation, after receiving fourth instruction information from the first network element, the first network element performs auxiliary positioning by using AI model #2.

[0442] In another possible implementation, after step 1207, which determines to adjust AI model #1 to AI model #2 based on the number of passes #A and the correspondence, the first network element performs auxiliary positioning by using AI model #2.

[0443] According to the above technical solution, whether or not the AI ​​model needs to be adjusted can be determined based on the measurement results of channel measurements, such as the number of passes #A, and their corresponding relationships. It is conceivable that changes in the channel environment may affect positioning accuracy. The number of passes #A can reflect the current channel environment, for example, the degree of NLOS. Thus, according to the above solution, the AI ​​model configured on the first network element side and the AI ​​model configured on the second network element side can be adjusted in a timely and cooperative manner based on changes in the channel state, thereby improving the positioning accuracy of the AI ​​model.

[0444] Method 1200 above primarily uses, for illustrative purposes, an example in which a second network element determines whether to adjust the AI ​​model configured within the second network element based on the measurement result of a first parameter (e.g., number of passes #A) provided by the first network element. This is not limited to this example. In other possible implementations, the second network element may determine whether to adjust the AI ​​model configured within the second network element based on the measurement result of a second parameter. The measurement result of the second parameter may be auxiliary positioning information. Based on this implementation, Method 1200 may not include steps 1205 to 1210. Alternatively, and even more optionally, based on this implementation, if the second network element determines that AI model #3 needs to be adjusted to AI model #4 based on the measurement result of the second parameter and its correspondence, the second network element may further transmit fourth instruction information to the first network element.

[0445] The second parameter may be, for example, one or more of the following: LOS probability, AOA, AoD, TOA, or the arrival time difference of the reference signal. There is a correspondence between the value of the second parameter and the AI ​​model. Thus, the second network element can determine whether to adjust the AI ​​model configured within the second network element based on the measurement result of the second parameter and the correspondence between the value of the second parameter and the AI ​​model within the second network element. For the correspondence between the value of the second parameter and the AI ​​model within the second network element, please refer to Table 1; in other words, replace the first parameter in Table 1 with the second parameter. Further details will not be explained here.

[0446] In the above embodiments, examples were primarily used for illustrative purposes in which a decision is made based on the input characteristics of the AI ​​model to switch or update the AI ​​model. In this application, it is considered that the intermediate performance indicator of the AI ​​model can also reflect the final network performance, such as network throughput performance, and a solution is further proposed for monitoring the intermediate performance indicator of the AI ​​model, and for deciding whether to switch or update the AI ​​model based on the monitoring results of the intermediate performance indicator. The intermediate performance indicator of the AI ​​model represents, for example, the output of the AI ​​model used for auxiliary positioning. For matters not described in detail below, please refer to the description of the embodiments above.

[0447] Figure 13 is a diagram of a communication method 1300 according to an embodiment of the present application. The method 1300 shown in Figure 13 may include the following steps.

[0448] 1310: The first network element receives the reference signal.

[0449] 1320: A first network element performs channel measurements based on a reference signal and obtains channel measurement results, which are inputs to a first AI model used for positioning. The output accuracy of the first AI model, a preset threshold, and a correspondence are used to determine whether to switch or update the first AI model, and the correspondence includes the complexity of the AI ​​model used for positioning and the correspondence between the AI ​​model used for positioning.

[0450] A pre-configured threshold may indicate the expected output accuracy. This pre-configured threshold may be transmitted to the first network element by another network element, pre-configured, or pre-defined. This is not limited to these pre-configured thresholds.

[0451] For example, the first network element may determine whether the first AI model needs to be adjusted based on the output accuracy of the first AI model and a preset threshold. For instance, if the output accuracy of the first AI model is lower than the preset threshold, it may be determined that the first AI model should be adjusted to a more complex AI model. Specifically, the particular AI model to which the first AI model should be adjusted may be determined based on a correspondence. For example, if the output accuracy of the first AI model is greater than or equal to the preset threshold, it may be determined that the first AI model does not need to be adjusted.

[0452] Optionally, the first AI model is an AI model configured within a first network element and used for LOS identification, or the first AI model is an AI model configured within a first network element and used for auxiliary positioning. In this embodiment of the present application, positioning may be performed by jointly using the first AI model with another AI model (e.g., an AI model configured within the first network element, or an AI model configured within another network element). For example, the output of the first AI model may be used as input to another AI model, the output of which may be position information of the first network element. In this embodiment of the present application, the output accuracy of the first AI model affects the final positioning accuracy, and the output accuracy of the first AI model is considered to be related to complexity. For example, a higher output accuracy of the first AI model indicates a higher complexity of the first AI model, and a more complex AI model is more adaptive to complex channels. Therefore, whether it is necessary to switch or update the first AI model is determined by monitoring the output accuracy of the first AI model, or the first AI model may be switched or updated adaptively in a timely manner based on changes in the channel state to ensure final positioning accuracy performance.

[0453] Optionally, the output of the first AI model includes at least one of the following: LOS probability, LOS hard decision result, angle, and time. For example, the angle may be AOA. For example, time may be TDoA or TOA. For example, the output of the AI ​​model is TOA, and the output accuracy of the AI ​​model is TOA accuracy. For example, the value of TOA accuracy may be, for example, 5.0m, 4.0m, 3.0m, 2.0m, 1.0m, or 0.5m. As another example, the output of the AI ​​model may be, for example, LOS probability or LOS hard decision result, and the output accuracy of the AI ​​model is LOS discrimination accuracy. For example, the value of LOS discrimination accuracy may be, for example, 50%, 70%, 80%, 90%, 95%, or 99%, etc.

[0454] In one example, the first AI model is an AI model configured within the first network element and used for LOS identification. In this case, for example, the output of the first AI model may be the LOS identification result (e.g., LOS probability, LOS hardness determination result).

[0455] In another example, the first AI model is an AI model configured within a first network element and used for auxiliary positioning. In this case, for example, the output of the first AI model may include, but is not limited to, at least one of the following: angle and time. For example, the angle may be AOA. For example, the time may be TDoA or TOA.

[0456] Similarly, for the sake of clarity, in the following explanation, "adjusting the AI ​​model" means "switching or updating the AI ​​model." In other words, for example, "adjusting the AI ​​model" as described below may be replaced with "switching the AI ​​model" or "updating the AI ​​model."

[0457] The correspondence includes the relationship between the complexity of the AI ​​model used for positioning and the AI ​​model used for positioning. Generally, AI models have different accuracies or accuracy ranges due to their different accuracies, and AI models with higher complexity are more adaptable to complex channels. Therefore, if the accuracy of the current AI model does not meet a preset threshold, this indicates that the current AI model cannot adapt to the current channel environment. In this case, a model with higher complexity can be determined based on the relationship between the complexity of the AI ​​model used for positioning and the AI ​​model used for positioning, in order to adapt to the current channel environment.

[0458] For example, a correspondence can exist in the form of a table, function, text, or string, and can be stored or transmitted. Table 9 is an example of presenting a correspondence in tabular form.

[0459] Table 9 is used as an example. Assume that complexity increases from level 1 to level 6. The first AI model is AI model #1. If the output accuracy of AI model #1 is below a predetermined threshold, AI model #1 may be adjusted to an AI model with a complexity level higher than level 1, such as AI model #2, or to any of the AI ​​models from AI model #3 to AI model #6. [Table 9]

[0460] Optionally, the first network element receives sixth instruction information, which indicates that the output accuracy of the first AI model is used to determine whether to adjust the AI ​​model used for positioning. Based on this, the first network element learns that it needs to monitor the output accuracy of the first AI model based on the sixth instruction information, and then may decide whether to adjust the AI ​​model based on the output of the first AI model. Furthermore, optionally, the sixth instruction information includes a preset threshold.

[0461] For example, in this embodiment of the present application, determining whether to adjust the first AI model involves the following two solutions.

[0462] Solution 1: The first network element determines whether to adjust the first AI model based on the output accuracy of the first AI model, a preset threshold, and the correspondence.

[0463] For further details, please refer to the relevant explanation in Solution 1 of Method 800. Further details will not be explained again here.

[0464] Solution 2: The first network element sends the output of the first AI model to the second network element, which then decides whether to adjust the first AI model based on the output accuracy of the first AI model, a preset threshold, and the correspondence.

[0465] For further details, please refer to the relevant explanation in Solution 2 of Method 800. Further details will not be explained again here.

[0466] For ease of understanding, the following describes the procedures applicable to the embodiments of this application, using the interaction between a first network element and a second network element as an example, with reference to Figure 14. For details not described below, please refer to the description of Method 1300. Details are not described below.

[0467] Figure 14 is a schematic flowchart of communication method 1400 according to another embodiment of the present application. Method 1400 is applicable to a scenario in which AI model inference is performed on the first network element side and storage of the AI ​​model library and / or training of the AI ​​model is performed on the second network element side. Method 1400 shown in Figure 14 may include the following steps:

[0468] 1401: The first network element performs channel measurements based on a reference signal and obtains the CIR.

[0469] In one example, the first network element is a terminal device or a component of a terminal device, such as a chip or circuit, and the reference signal is a downlink reference signal. In another example, the first network element is a network device or a positioning device, or a component of a network device or a positioning device, such as a chip or circuit, and the reference signal is an uplink reference signal.

[0470] Optionally, method 1400 further includes step 1402.

[0471] 1402: The first network element inputs the CIR to the first AI model to perform LOS identification or auxiliary positioning.

[0472] For example, if the first network element inputs CIR into the first AI model to perform LOS identification, the first network element outputs an LOS probability or LOS hardness determination result. In another example, if the first network element inputs CIR into the first AI model to perform auxiliary positioning, the first network element outputs an angle or time. For example, the angle may be AOA. For example, the time may be TDoA or TOA.

[0473] Optionally, method 1400 further includes step 1403.

[0474] 1403: The second network element transmits the sixth instruction information and / or correspondence to the first network element.

[0475] The sixth instruction indicates that the output accuracy of the first AI model is used to determine whether to adjust the AI ​​model used for positioning.

[0476] Optionally, the sixth instruction information includes a preset threshold. In this case, the instruction that the output accuracy of the first AI model is used to determine whether to adjust the AI ​​model used for positioning may be replaced with the instruction that the AI ​​model used for positioning should be adjusted when the output accuracy of the first AI model is less than the preset threshold.

[0477] For example, the second network element is a network device, a core network device, a positioning device, or a server or cloud device in an OTT system; or a component of a network device, a core network device, a positioning device, or a server or cloud device in an OTT system, such as a chip or circuit.

[0478] 1404: The first network element monitors the output of the first AI model.

[0479] Optionally, method 1400 further includes step 1405.

[0480] 1405: The first network element determines to adjust the first AI model to the second AI model based on the output accuracy of the first AI model, a preset threshold, and the correspondence relationship.

[0481] In this embodiment of the present application, we assume, for example, that it is necessary to adjust the first AI model to the second AI model.

[0482] 1406: The first network element sends request information to the second network element, which is used to request the output of the first AI model or the second AI model.

[0483] Where possible, if method 1400 includes step 1405, in step 1406, the first network element transmits request information to the second network element, where the request information is used to request a second AI model.

[0484] In another possible case, if method 1400 does not include step 1405, in step 1406, the first network element transmits the output of the first AI model to the second network element, where the output of the first AI model may be used by the second network element to determine whether to adjust the first AI model.

[0485] 1407: The second network element transmits information about the second AI model to the first network element.

[0486] If possible, in step 1406, when the first network element sends the output of the first AI model to the second network element, the second network element decides whether to adjust the first AI model based on the output accuracy of the first AI model, the preset thresholds, and the correspondence. Assume that the second network element decides, based on the output accuracy of the first AI model, the preset thresholds, and the correspondence, that the first AI model needs to be adjusted to the second AI model. In this case, the second network element sends the second AI model to the first network element.

[0487] In another possible case, if in step 1406 the first network element sends request information to the second network element used to request the second AI model, the second network element responds to the request information by sending the second AI model to the first network element.

[0488] 1408: The first network element performs LOS identification or auxiliary positioning by using the second AI model.

[0489] It can be understood that Figure 14 is primarily illustrated by using an example of tuning a first AI model configured within a first network element. For example, when tuning the first AI model to a second AI model, another AI model used in conjunction with the first AI model to perform positioning may also be tuned to match the second AI model. See the embodiment shown in Figure 12 for details. Further details will not be explained again here.

[0490] According to the technical solutions described above, it is possible to determine whether the AI ​​model needs to be adjusted based on the output and correspondence of the AI ​​model used for LOS identification or the AI ​​model used for auxiliary positioning, or to switch or update the AI ​​model as needed to adapt to changes in the environment in which the AI ​​model is used. This helps to mitigate or avoid the impact on the performance of the network in which the AI ​​model is deployed.

[0491] It can be understood that some optional features in embodiments of this application may be independent of other features in some scenarios, or may be combined with other features in some scenarios. This is not limited to this.

[0492] It may be further understood that in some of the embodiments described above, the act of transmitting information is mentioned multiple times. For example, A transmits information to B. A transmitting information to B may include A transmitting information to B, or it may include A transmitting information to B via another device or network element. This is not limited to these actions.

[0493] It may be further understood that in some of the embodiments described above, adjusting the AI ​​model may be replaced by switching or updating the AI ​​model.

[0494] It may be further understood that in some of the embodiments described above, the AI ​​model used is primarily for illustrative purposes to illustrate an example where the AI ​​model is used for positioning. It may also be understood that the AI ​​model may be used for other purposes.

[0495] The solutions in the embodiments of this application may be appropriately combined for use, and it can be further understood that the definitions or descriptions of terms in the embodiments may be mutually referenced or described in the embodiments. This is not limited to these definitions.

[0496] In embodiments of the methods described above, the methods and operations implemented by the network element may be alternatively implemented by components of the network element (such as chips or circuits). This is not limited to these.

[0497] The above describes in detail the method provided in the embodiments of this application with reference to Figures 8 to 14. Below, the apparatus provided in the embodiments of this application will be described in detail with reference to Figures 8 to 14. Please understand that the description of the apparatus embodiments corresponds to the description of the method embodiments. Therefore, for matters not described in detail, please refer to the method embodiments described above. For the sake of brevity, details will not be described again here.

[0498] Figure 15 is a diagram of a communication device 1500 according to an embodiment of the present application. The device 1500 includes a transceiver unit 1510 and a processing unit 1520. The transceiver unit 1510 may be configured to implement corresponding communication functions. The transceiver unit 1510 may also be referred to as a communication interface or communication unit. The processing unit 1520 may be configured to perform processing such as determining whether to adjust an AI model.

[0499] Optionally, the device 1500 may further include a storage unit. The storage unit may be configured to store instructions and / or data. The processing unit 1520 may read instructions and / or data from the storage unit so that the device implements embodiments of the method described above.

[0500] In the design, the device 1500 is configured to perform steps or procedures performed by the first network element in the embodiment of the method described above, the transmitting / receiving unit 1510 is configured to perform operations related to receiving and transmitting on the first network element side in the embodiment of the method described above, and the processing unit 1520 is configured to perform operations related to processing on the first network element side in the embodiment of the method described above.

[0501] In possible implementations, the device 1500 is configured to perform steps or procedures performed by the first network element in the embodiments shown in Figures 8 to 12. Optionally, a transmit / receive unit 1510 is configured to receive a reference signal, and a processing unit 1520 is configured to perform channel measurements based on the reference signal and obtain channel measurement results, where the channel measurement results, or auxiliary positioning information based on the channel measurement results, are inputs to an artificial intelligence (AI) model used for positioning, and the measurement results and correspondences of the first parameter are used to determine whether to switch or update the AI ​​model, where the measurement results of the first parameter are obtained based on the channel measurement results, and the correspondences include the correspondence between the AI ​​model used for positioning and the value of the first parameter.

[0502] In another possible implementation, the device 1500 is configured to perform steps or procedures performed by a second network element in the embodiments shown in Figures 8, 10, 11, and 12. Optionally, a transmit / receive unit 1510 is configured to receive measurement results of a first parameter, where the measurement results of the first parameter are based on channel measurement results, and the channel measurement results or auxiliary positioning information based on channel measurement results are inputs to an artificial intelligence (AI) model used for positioning. A processing unit 1520 is configured to determine whether to switch or update the AI ​​model based on the measurement results of the first parameter and a correspondence, where the correspondence includes a correspondence between the AI ​​model used for positioning and the value of the first parameter.

[0503] In another possible implementation, the device 1500 is configured to perform steps or procedures performed by a second network element in the embodiments shown in Figures 8, 9, 11, and 12. Optionally, the transmit / receive unit 1510 is configured to transmit first instruction information and / or correspondence, where the correspondence includes a correspondence between an artificial intelligence (AI) model used for positioning and a value of a first parameter, the measurement result of the first parameter is obtained based on a channel measurement result, the channel measurement result or auxiliary positioning information based on the channel measurement result is an input to the AI ​​model, and the first instruction information indicates that the first parameter is used to determine whether to switch or update the AI ​​model used for positioning.

[0504] In another possible implementation, the device 1500 is configured to perform steps or procedures performed by the first network element in the embodiments shown in Figures 13 and 14. Optionally, the transmit / receive unit 1510 is configured to receive a reference signal, and the processing unit 1520 is configured to perform channel measurements based on the reference signal and obtain channel measurement results, where the channel measurement results or auxiliary positioning information based on the channel measurement results are inputs to a first AI model used for positioning, and the output accuracy of the first AI model, a preset threshold, and a correspondence are used to determine whether to switch or update the first AI model, the correspondence includes the complexity of the AI ​​model used for positioning and the correspondence between the AI ​​model used for positioning.

[0505] In another possible implementation, the device 1500 is configured to perform steps or procedures performed by a second network element in the embodiments shown in Figures 13 and 14. Optionally, a transmit / receive unit 1510 is configured to receive the output of a first AI model, where the input of the first AI model is a channel measurement result. A processing unit 1520 is configured to determine whether to switch or update the AI ​​model based on the output accuracy of the first AI model, a preset threshold, and a correspondence, where the correspondence includes a correspondence between the complexity of the AI ​​model used for positioning and the AI ​​model used for positioning.

[0506] The specific process by which the unit performs the corresponding steps described above is described in detail in the embodiments of the method described above, and for the sake of brevity, please understand that the details will not be described here.

[0507] It should be understood that the apparatus 1500 described herein is implemented in the form of a functional unit. The term “unit” as used herein may mean an application-specific integrated circuit (ASIC), an electronic circuit, a processor configured to run one or more software or firmware programs (e.g., a shared processor, a dedicated processor, or a group processor), memory, a merged logic circuit, and / or another suitable component that supports the described function. In any example, those skilled in the art will understand that the apparatus 1500 may specifically be a network element in the embodiments described above (e.g., a first network element or a second network element) and may be configured to perform procedures and / or steps corresponding to the network element in the embodiments of the method described above. To avoid repetition, further details are not described here.

[0508] Each of the above solutions, the apparatus 1500, has the function of implementing the corresponding steps performed by a network element (e.g., a first network element or a second network element) in the method described above. This function may be implemented by hardware, or by the hardware running corresponding software. The hardware or software includes one or more modules corresponding to the above function. For example, a transceiver unit may be replaced by a transceiver (e.g., a transmitting unit in a transceiver unit may be replaced by a transmitting machine, and a receiving unit in a transceiver unit may be replaced by a receiving machine), and another unit, such as a processing unit, may be replaced by a processor to separately perform the transceiver operation and associated processing operation in the embodiment of the method.

[0509] In addition, the transmitting / receiving unit 1510 may alternatively be a transmitting / receiving circuit (which may include, for example, a receiving circuit and a transmitting circuit), and the processing unit may be a processing circuit.

[0510] The device in Figure 15 may be a network element in the above embodiment, or it may be a chip or chip system, such as a system on a chip (SoC). The transceiver unit may be an input / output circuit or a communication interface. The processing unit is a processor, microprocessor, or integrated circuit on a chip. This is not limited to the foregoing.

[0511] Figure 16 is a diagram of another communication device 1600 according to an embodiment of the present application. The device 1600 includes a processor 1610. The processor 1610 is coupled to a memory 1620. The memory 1620 is configured to store computer programs or instructions and / or data. The processor 1610 is configured to execute computer programs or instructions stored in the memory 1620, or to read data stored in the memory 1620, and to perform the method in the embodiment of the method described above.

[0512] There may be one or more processors 1610.

[0513] There is arbitrarily one or more memory units of 1620.

[0514] Optionally, the memory 1620 may be integrated with the processor 1610, or the memory 1620 and the processor 1610 may be arranged separately.

[0515] Optionally, as shown in Figure 16, the device 1600 may further include a transceiver 1630, which is configured to receive and / or transmit signals. For example, a processor 1610 is configured to control the transceiver 1630 to receive and / or transmit signals.

[0516] For example, the processor 1610 may have the functions of the processing unit 1520 shown in Figure 15, the memory 1620 may have the functions of a storage unit, and the transceiver 1630 may have the functions of the transmitting / receiving unit 1510 shown in Figure 15.

[0517] In one solution, the device 1600 is configured to implement the operations performed by the network elements (e.g., a first network element or a second network element) in the embodiments of the method described above.

[0518] For example, the processor 1610 is configured to execute a computer program or instruction stored in the memory 1620 to implement the relevant operations of the network elements (e.g., the first network element or the second network element) in the embodiment of the method described above.

[0519] It should be understood that the processor referred to in the embodiments of this application may be a central processing unit (CPU), or another general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field programmable gate array (FPGA), or another programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general-purpose processor may be a microprocessor, or the processor may be any conventional processor, etc.

[0520] It should be further understood that the memory referred to in the embodiments of this application may be volatile memory and / or non-volatile memory. Non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may be random access memory (RAM). For example, RAM may be used as an external cache. As an example, rather than an exhaustive list, RAM includes multiple forms such as static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchlink dynamic random access memory (synchlink DRAM, SLDRAM), and direct rambus random access memory (direct rambus RAM, DR RAM).

[0521] Note that when the processor is a general-purpose processor, DSP, ASIC, FPGA, or other programmable logic device, individual gate or transistor logic device, or individual hardware component, the memory (storage module) may be integrated into the processor.

[0522] It should be further noted that the memories described herein are intended to include, but are not limited to, these and any other suitable types of memory.

[0523] Figure 17 is a diagram of a chip system 1700 according to an embodiment of the present application. The chip system 1700 (sometimes referred to as a processing system) includes a logic circuit 1710 and an input / output interface 1720.

[0524] The logic circuit 1710 may be a processing circuit within the chip system 1700. The logic circuit 1710 may be coupled to and connected to a memory unit to call instructions within the memory unit, enabling the chip system 1700 to implement the methods and functions of the embodiments of this application. The input / output interface 1720 may be an input / output circuit within the chip system 1700 that outputs information processed by the chip system 1700 or inputs data or signal information to be processed into the chip system 1700 for processing.

[0525] Specifically, for example, if the chip system 1700 is installed in a first network element, the logic circuit 1710 is coupled to an input / output interface 1720, which inputs a reference signal to the logic circuit 1710 for processing, and can, for example, perform channel measurements on the reference signal to obtain the measurement results of the first parameter. In another example, if the chip system 1700 is installed in a second network element, the logic circuit 1710 is coupled to an input / output interface 1720, which inputs the measurement results of the first parameter from the first network element to the logic circuit 1710 for processing.

[0526] In one solution, the chip system 1700 is configured to implement the operations performed by the network elements (e.g., a first network element or a second network element) in the embodiments of the method described above.

[0527] For example, the logic circuit 1710 is configured to implement processing-related operations performed by a network element (e.g., a first network element or a second network element) in an embodiment of the method described above, and the input / output interface 1720 is configured to implement transmission-related operations and / or reception-related operations performed by a network element (e.g., a first network element or a second network element) in an embodiment of the method described above.

[0528] Embodiments of this application further provide a computer-readable storage medium. The computer-readable storage medium stores computer instructions for implementing a method executed by a network element (e.g., a first network element or a second network element) in embodiments of the above method.

[0529] For example, when a computer program is executed by a computer, the computer can implement a method that is executed by a network element (e.g., a first network element or a second network element) in an embodiment of the above method.

[0530] Embodiments of this application further provide a computer program product including instructions. When the instructions are executed by a computer, an embodiment of the above method implements a method that is executed by a network element (e.g., a first network element or a second network element).

[0531] Embodiments of this application further provide a communication system. The communication system includes the first network element and / or the second network element in the embodiments described above. For example, the system includes the first network element and / or the second network element in the embodiment shown in Figure 8. As another example, the system includes the first network element and / or the second network element in the embodiments shown in Figures 8 to 12.

[0532] For a description of the relevant aspects and effects of any of the devices provided above, please refer to the corresponding embodiments of the methods provided above. Further details will not be explained here.

[0533] In some embodiments provided in this application, it should be understood that the disclosed apparatus and methods may be implemented in other ways. For example, the embodiments of the apparatus described are merely examples. For example, the division into units is merely a logical functional division, and other divisions may be used in actual implementations. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not performed. In addition, the mutual coupling, direct coupling, or communication connection shown or discussed may be implemented through some interfaces. Indirect coupling or communication connection between apparatus or units may be implemented electronically, mechanically, or in other forms.

[0534] All or part of the embodiments described above may be implemented using software, hardware, firmware, or any combination thereof. When an embodiment is implemented using software, all or part of the embodiment may be implemented in the form of a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded into a computer and executed, all or part of the procedures or functions according to the embodiments of this application are generated. The computer may be a general-purpose computer, a dedicated computer, a computer network, or another programmable device. For example, the computer may be a personal computer, a server, or a network device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (e.g., coaxial cable, optical fiber, or digital subscriber line (DSL)) or wireless (e.g., infrared, radio, or microwave). Computer-readable storage media may be any usable media accessible by a computer, or a data storage device that integrates one or more usable media, such as a server or data center. Usable media may be magnetic media (e.g., floppy disks, hard disks, or magnetic tapes), optical media (e.g., DVDs), semiconductor media (e.g., solid-state drives (SSDs)), etc. Usable media may include, but are not limited to, any media capable of storing program code, such as USB flash drives, removable hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0535] The foregoing description is merely a specific implementation of the present application and is not intended to limit the scope of protection of this application. Any modification or substitution readily understood by a person skilled in the art within the technical scope disclosed in this application shall be included within the scope of protection of this application. Accordingly, the scope of protection of this application shall be subject to the scope of protection of the claims.

Claims

1. A method of communication, The first network element receives a reference signal, A step of performing channel measurement based on the reference signal using the first network element and obtaining channel measurement results, wherein the channel measurement results or auxiliary positioning information based on the channel measurement results are inputs to an artificial intelligence (AI) model used for positioning, and the measurement results and correspondence of a first parameter are used to determine whether to switch or update the AI ​​model, the measurement results of the first parameter are obtained based on the channel measurement results, and the correspondence includes a correspondence between the AI ​​model used for positioning and the value of the first parameter. A communication method that includes this.

2. This method is A step of receiving first instruction information and / or the correspondence by the first network element, wherein the first instruction information indicates that the first parameter is used to determine whether to switch or update the AI ​​model used for positioning. The method according to claim 1, further comprising:

3. This method is The first network element determines whether to switch or update the AI ​​model based on the measurement results of the first parameter and the corresponding relationship. The method according to claim 1 or 2, further comprising:

4. The first network element determines whether to switch or update the AI ​​model based on the measurement results of the first parameter and the corresponding relationship, The first network element determines, based on the measurement results of the first parameter and the correspondence, whether to switch or update the first type of AI model used for positioning from the first AI model to the second AI model, wherein the channel measurement results are inputs to the first type of AI model used for positioning, and the correspondence includes the correspondence between the first type of AI model used for positioning and the value of the first parameter. This method is The first network element transmits request information, the request information being used to request the second AI model, and the steps are as follows: A step of receiving a response to the request information by the first network element, wherein the response to the request information includes information about the second AI model, The first network element performs positioning by using the second AI model, The method according to claim 3, further comprising:

5. The first network element determines whether to switch or update the AI ​​model based on the measurement results of the first parameter and the corresponding relationship, A step in which the first network element determines, based on the measurement result of the first parameter and the correspondence, to switch or update the second type AI model used for positioning from a third AI model to a fourth AI model, wherein the channel measurement result is an input to the first type AI model used for positioning, the auxiliary positioning information is an output to the first type AI model, the first type AI model is switched or updated from a first AI model corresponding to the third AI model to a second AI model corresponding to the fourth AI model, the output of the first type AI model used for positioning is an input to the second type AI model used for positioning, and the correspondence includes a correspondence between the second type AI model used for positioning and the value of the first parameter. The first network element transmits second instruction information, the second instruction information instructs to switch or update the second type of AI model used for positioning from the third AI model to the fourth AI model. The method according to claim 3, including the method described in claim 3.

6. This method is The first network element transmits the measurement result of the first parameter. The method according to claim 1 or 2, further comprising:

7. The step of transmitting the measurement result of the first parameter by the first network element is: The process includes the step of transmitting the measurement result of the first parameter by the first network element when the measurement result of the first parameter satisfies a specified condition, The method according to claim 6.

8. This method is The first network element receives third instruction information, the third instruction information instructs to switch or update the first type of AI model used for positioning from the first AI model to the second AI model, The first network element performs positioning by using the second AI model, The method according to claim 6 or 7, further comprising:

9. The third instruction information includes information relating to the second AI model, The method according to claim 8.

10. This method is A step of receiving fourth instruction information by the first network element, wherein the fourth instruction information instructs to switch or update the second type of AI model used for positioning from the third AI model to the fourth AI model. The method according to claim 6 or 7, further comprising:

11. The first parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or the moving speed of the first network element, where k is a number greater than 0. The method according to any one of claims 1 to 10.

12. The first type of AI model used for positioning is an AI model used for positioning that is configured within the first network element. The method according to claim 4, 5, 8, or 9.

13. The second type of AI model used for positioning is an AI model used for positioning, configured within a second network element. The method according to claim 5 or 10.

14. A method of communication, A step of receiving a measurement result of a first parameter by a second network element, wherein the measurement result of the first parameter is based on a channel measurement result, and the measurement result of the first parameter or auxiliary positioning information based on the measurement result of the first parameter is input to an artificial intelligence (AI) model used for positioning. A step of determining whether to switch or update the AI ​​model based on the correspondence between the measurement result and the first parameter using the second network element, wherein the correspondence includes the correspondence between the AI ​​model used for positioning and the value of the first parameter. A communication method that includes this.

15. The second network element determines whether to switch or update the AI ​​model based on the correspondence between the measurement results and the first parameter, The second network element determines, based on the measurement results and correspondence of the first parameter, whether to switch or update the first type AI model used for positioning from the first AI model to the second AI model, wherein the channel measurement results are inputs to the first type AI model used for positioning, and the correspondence includes the correspondence between the first type AI model used for positioning and the value of the first parameter. This method is A step of transmitting third instruction information by the second network element, wherein the third instruction information instructs to switch or update the first type AI model used for positioning from the first AI model to the second AI model. The method according to claim 14, further comprising:

16. The second network element determines whether to switch or update the AI ​​model based on the correspondence between the measurement results and the first parameter, The second network element determines, based on the measurement results of the first parameter and the correspondence, whether to switch or update the second type AI model used for positioning from a third AI model to a fourth AI model, wherein the channel measurement results are inputs to the first type AI model used for positioning, the auxiliary positioning information is an output of the first type AI model, the first type AI model is switched or updated from a first AI model corresponding to the third AI model to a second AI model corresponding to the fourth AI model, the output of the first type AI model used for positioning is an input to the second type AI model used for positioning, and the correspondence includes a correspondence between the second type AI model used for positioning and the value of the first parameter. The second network element performs positioning by using the fourth AI model, The method according to claim 14, including the method described in claim 14.

17. This method is A step of transmitting fourth instruction information by the second network element, wherein the fourth instruction information instructs to switch or update the second type AI model used for positioning from the third AI model to the fourth AI model. The method according to claim 16, further comprising:

18. This method is A step of transmitting first instruction information by the second network element, wherein the first instruction information indicates that the first parameter is used to determine whether to switch or update the AI ​​model used for positioning. The method according to any one of claims 14 to 17, further comprising:

19. The first parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or moving velocity of the first network element, where k is a number greater than 0. The method according to claim 18.

20. The first type of AI model used for positioning is an AI model used for positioning that is configured within the first network element. The method according to any one of claims 15 to 17.

21. The second type of AI model used for positioning is an AI model used for positioning that is configured within the second network element. The method according to any one of claims 15 to 17 or 20.

22. A method of communication, A step of transmitting first instruction information and / or correspondence by a second network element, wherein the correspondence includes a correspondence between an artificial intelligence (AI) model used for positioning and a value of a first parameter, the measurement result of the first parameter is based on a channel measurement result, the measurement result of the first parameter or auxiliary positioning information based on the measurement result of the first parameter is an input to the AI ​​model, and the first instruction information indicates that the first parameter is used to determine whether to switch or update the AI ​​model used for positioning. A communication method that includes this.

23. This method is The steps include: receiving request information by the second network element, wherein the request information is used to request a second AI model; A step of transmitting a response to the request information by the second network element, wherein the response to the request information includes information about the second AI model. The method according to claim 22, further comprising:

24. This method is A step of receiving second instruction information by the second network element, wherein the second instruction information instructs to switch or update the second type AI model used for positioning from a third AI model to a fourth AI model, the channel measurement result is an input to the first type AI model used for positioning, the auxiliary positioning information is an output to the first type AI model, the first type AI model is switched or updated from a first AI model corresponding to the third AI model to a second AI model corresponding to the fourth AI model, the output of the first type AI model used for positioning is an input to the second type AI model used for positioning, and the correspondence includes a correspondence between the second type AI model used for positioning and the value of the first parameter. The second network element performs positioning based on the second instruction information by using the fourth AI model, The method according to claim 22, further comprising:

25. A method of communication, A step of receiving measurement results of a second parameter from a first network element via a second network element, wherein the measurement results of the second parameter include measurement results obtained based on the output of an artificial intelligence (AI) model configured within the first network element. The second network element determines whether to switch or update the AI ​​model used for positioning, which is configured within the second network element, based on the correspondence between the measurement result of the second parameter, wherein the correspondence includes the correspondence between the AI ​​model used for positioning and the value of the second parameter. A communication method that includes this.

26. The second parameter includes at least one of the following: timing error group, inter-site synchronization error, delay spread, number of paths whose energy is greater than k times the energy of the first path in the channel impulse response, average power of multiple sampling points, line-of-sight probability, signal-to-interference plus noise ratio, reference signal received power, Rice coefficient, Doppler frequency, Doppler shift, or the moving speed of the first network element, where k is a number greater than 0. The method according to claim 25.

27. The step of determining whether to switch or update the AI ​​model used for positioning, which is configured within the second network element, based on the correspondence between the measurement results of the second parameter and the second parameter, is as follows: The second network element determines, based on the measurement results of the second parameter and the corresponding relationship, whether to switch or update the AI ​​model used for positioning, which is configured within the second network element, from the third AI model to the fourth AI model. A step of transmitting a fifth instruction information by the second network element, wherein the fifth instruction information instructs to switch or update the AI ​​model used for positioning, which is configured within the second network element, from the third AI model to the fourth AI model. The method according to claim 25 or 26, including the method described in claim 25 or 26.

28. A method of communication, The first network element receives a reference signal, A step of performing channel measurement based on the reference signal using the first network element and obtaining channel measurement results, wherein the channel measurement results are input to a first artificial intelligence (AI) model used for positioning, and the output accuracy, preset thresholds, and correspondences of the first AI model are used to determine whether to switch or update the first AI model, the correspondences include a correspondence between the complexity of the AI ​​model used for positioning and the AI ​​model used for positioning. A communication method that includes this.

29. The first AI model is an AI model configured within the first network element and used for auxiliary positioning, or the first AI model is an AI model configured within the first network element and used for LOS identification. The method according to claim 28.

30. This method is A step of receiving sixth instruction information and / or the correspondence by the first network element, wherein the sixth instruction information indicates that the output accuracy of the first AI model is used to determine whether to switch or update the AI ​​model used for positioning. The method according to claim 28 or 29, further comprising:

31. This method is The first network element determines whether to switch or update the first AI model based on the output accuracy of the first AI model, the preset threshold, and the correspondence relationship. The method according to claim 28 or 29, further comprising:

32. The first network element determines whether to switch or update the first AI model based on the output accuracy of the first AI model, the preset threshold, and the correspondence relationship, The first network element includes the step of determining whether to switch or update the first AI model to the second AI model based on the output accuracy of the first AI model, the preset threshold, and the correspondence relationship, This method is The first network element transmits request information, the request information being used to request the second AI model, and the steps are as follows: A step of receiving a response to the request information by the first network element, wherein the response to the request information includes information about the second AI model, The first network element performs positioning by using the second AI model, The method according to claim 31, further comprising:

33. This method is The first network element transmits the output of the first AI model. The method according to any one of claims 28 to 32, further comprising:

34. The step of transmitting the output of the first AI model by the first network element is: The process includes the step of transmitting the output of the first AI model by the first network element when the output of the first AI model satisfies a specified condition. The method according to claim 33.

35. The output of the first AI model includes at least one of the following: line of sight probability, line of sight hardness determination result, angle, and time. The method according to claim 33 or 34.

36. A method of communication, A second network element receives the output of a first artificial intelligence (AI) model used for positioning, wherein the input to the first AI model is a channel measurement result. The second network element determines whether to switch or update the AI ​​model based on the output accuracy of the first AI model, a preset threshold, and a correspondence relationship, wherein the correspondence relationship includes a correspondence between the complexity of the AI ​​model used for positioning and the AI ​​model used for positioning. A communication method that includes this.

37. The first AI model is an AI model used for positioning and / or an AI model used for auxiliary positioning, configured within the first network element. The method according to claim 36.

38. The second network element determines whether to switch or update the AI ​​model based on the output accuracy of the first AI model, a preset threshold, and the correspondence relationship. The second network element includes the step of determining whether to switch or update the first AI model to the second AI model based on the output accuracy of the first AI model, a preset threshold, and the correspondence relationship. This method is A step of transmitting third instruction information by the second network element, wherein the third instruction information instructs to switch or update the first AI model to the second AI model. The method according to claim 36 or 37, further comprising:

39. The third instruction information includes information relating to the second AI model, The method according to claim 38.

40. This method is A step of transmitting sixth instruction information by the second network element, wherein the sixth instruction information indicates that the output accuracy of the first AI model is used to determine whether to switch or update the AI ​​model used for positioning. The method according to any one of claims 36 to 39, further comprising:

41. The output of the first AI model includes at least one of the following: line of sight probability, line of sight hardness determination result, angle, and time. The method according to any one of claims 36 to 40.

42. A communication device comprising a module or unit configured to perform the method described in any one of claims 1 to 41.

43. A communication device comprising a processor, wherein the processor is configured to execute a computer program or instruction stored in memory, thereby enabling the device to perform the method described in any one of claims 1 to 41.

44. The device further comprises the memory and / or communication interface, the communication interface being coupled to the processor, The aforementioned communication interface is configured to input and / or output information. The apparatus according to claim 43.

45. The device in question is a communication device, circuit, or chip. The apparatus according to any one of claims 42 to 44.

46. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program or instruction, and when the computer program or instruction is executed by a communication device, the communication device becomes capable of performing the method according to any one of claims 1 to 41.

47. A computer program product comprising a computer program or instructions for performing the method described in any one of claims 1 to 41.

48. A communication system comprising a first network element and a second network element, The first network element is configured to perform the method described in any one of claims 1 to 13, A communication system in which the second network element is configured to perform the method described in any one of claims 14 to 21, or the second network element is configured to perform the method described in any one of claims 22 to 24, or the second network element is configured to perform the method described in any one of claims 25 to 27.

49. A communication system comprising a first network element and a second network element, The first network element is configured to perform the method described in any one of claims 28 to 35, A communication system wherein the second network element is configured to perform the method described in any one of claims 36 to 41.

Images