Method for determining CSI input into ai model, terminal, network device, system, and medium

Abstract

The present disclosure relates to a method for determining a CSI input into an AI model, a terminal, a network device, a system, and a medium. The method comprises: determining that a first artificial intelligence (AI) model of a terminal does not have a first channel state information (CSI) input, wherein the first CSI input refers to a CSI input of a first data transmission layer into the first AI model at a first moment; and determining that second CSI output by the first AI model is the first CSI input. Thus, when a CSI input into an AI model is not obtained, another CSI output from the AI model is used as the CSI input into the AI model, which clarifies a method for determining the CSI input into the AI model, and ensures the reliability and accuracy of AI model inference during a CSI compression feedback process in a communication system.

Description

Method, terminal, network device, system and medium for determining CSI input of AI model Technical Field

[0001] The present disclosure relates to the field of communication technologies, and in particular to a method, terminal, network device, system, and medium for determining CSI input of an AI model. Background Art

[0002] The use of AI (Artificial Intelligence) technology in communication systems can reduce terminal information feedback overhead and improve the accuracy of CSI (Channel State Information) feedback. Related technologies, a bilateral model based on an encoder and decoder, can achieve compressed feedback of space-frequency-time domain CSI, which offers performance gains for communication system data transmission compared to Type II feedback codebooks.

[0003] Summary of the Invention

[0004] In order to overcome the technical problem of unreliable CSI input of AI models in related technologies, the present disclosure provides a method, terminal, network device, system and medium for determining CSI input of an AI model.

[0005] According to a first aspect of an embodiment of the present disclosure, a method for determining a CSI input of an AI model is proposed, which is executed by a terminal. The method includes:

[0006] Determining that a first channel state information (CSI) input does not exist for a first artificial intelligence (AI) model of the terminal, where the first CSI input is a CSI input of a first data transmission layer of the first AI model at a first moment;

[0007] Determine the second CSI output by the first AI model as the first CSI input.

[0008] According to a second aspect of an embodiment of the present disclosure, a method for determining a CSI input of an AI model is provided, the method being performed by a network device. The method includes:

[0009] Determining that a second AI model of the network device does not have a second CSI input, where the second CSI input is a CSI input of the second AI model at the first time to the first data transmission layer;

[0010] Determine the third CSI output by the second AI model as the second CSI input.

[0011] According to a third aspect of an embodiment of the present disclosure, a terminal is provided, including:

[0012] A first processing module is configured to determine that a first AI model of the terminal does not have a first CSI input, where the first CSI input is a CSI input of the first AI model at a first data transmission layer at a first moment;

[0013] The second processing module is configured to determine the second CSI output by the first AI model as the first CSI input.

[0014] a third processing module configured to determine that the second AI model of the network device does not have a second CSI input, where the second CSI input is a CSI input of the second AI model at the first time to the first data transmission layer;

[0015] The fourth processing module is configured to determine the third CSI output by the second AI model as the second CSI input.

[0016] According to a fifth aspect of an embodiment of the present disclosure, a terminal is provided, including:

[0017] one or more processors;

[0018] The terminal is used to execute the method for determining the AI ​​model CSI input described in any one of the first aspects of this disclosure.

[0019] According to a sixth aspect of an embodiment of the present disclosure, a network device is provided, including:

[0020] one or more processors;

[0021] The network device is used to execute the method for determining the AI ​​model CSI input described in any one of the second aspects of this disclosure.

[0022] According to the seventh aspect of an embodiment of the present disclosure, a communication system is provided, including a terminal and a network device, wherein the terminal is configured to implement the method for determining the AI ​​model CSI input described in any one of the first aspect of the present disclosure, and the network device is configured to implement the method for determining the AI ​​model CSI input described in any one of the second aspect of the present disclosure.

[0023] According to an eighth aspect of an embodiment of the present disclosure, a storage medium is provided, which stores instructions. When the instructions are executed on a communication device, the communication device executes the method for determining the AI ​​model CSI input as described in any one of the first aspect of the present disclosure, or causes the communication device to execute the method for determining the AI ​​model CSI input as described in any one of the second aspect of the present disclosure.

[0024] According to a ninth aspect of an embodiment of the present disclosure, a computer program product is provided, comprising a computer program and / or instructions, which, when executed by a communication device, implement the method for determining the AI ​​model CSI input as described in any one of the first aspects of the present disclosure, or implement the method for determining the AI ​​model CSI input as described in any one of the second aspects of the present disclosure when the computer program and / or instructions are executed by a communication device.

[0025] In the above solution, it is determined that the first channel state information (CSI) input to the first artificial intelligence (AI) model of the terminal does not exist, and the first CSI input is the CSI input of the first AI model at the first data transmission layer at the first moment. The second CSI output by the first AI model is determined as the first CSI input. Therefore, when the CSI input to the AI ​​model is not available, the other CSI outputs of the AI ​​model are used as the CSI input to the AI ​​model. This clarifies the method for determining the CSI input to the AI ​​model and ensures the reliability and accuracy of AI model reasoning during the CSI compression feedback process in the communication system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following drawings required for describing the embodiments are introduced. The following drawings are merely some embodiments of the present disclosure and do not impose specific limitations on the protection scope of the present disclosure.

[0027] FIG1a is a schematic diagram showing the architecture of a communication system according to an embodiment of the present disclosure.

[0028] FIG1 b is a schematic diagram showing CSI compression feedback according to an embodiment of the present disclosure.

[0029] FIG2 is an interactive schematic diagram of a method for determining CSI input of an AI model according to an embodiment of the present disclosure.

[0030] FIG3 is a flow chart of a method for determining CSI input of an AI model according to an embodiment of the present disclosure.

[0031] FIG4 is a flow chart of a method for determining CSI input of an AI model according to an embodiment of the present disclosure.

[0032] FIG5 is a schematic flow chart of a method for determining historical CSI input based on CSI compression according to an embodiment of the present disclosure.

[0033] FIG6 is a schematic structural diagram of a terminal proposed in an embodiment of the present disclosure.

[0034] FIG7 is a schematic diagram of the structure of a network device proposed in an embodiment of the present disclosure.

[0035] FIG8 is a schematic structural diagram of a communication device 8100 according to an embodiment of the present disclosure.

[0036] FIG9 is a schematic structural diagram of a chip 8200 according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0037] The embodiments of the present disclosure provide a method, terminal, network device, system, and medium for determining CSI input of an AI model.

[0038] In a first aspect, an embodiment of the present disclosure provides a method for determining CSI input of an AI model, which is executed by a terminal. The method includes:

[0039] Determining that a first channel state information (CSI) input does not exist for a first artificial intelligence (AI) model of the terminal, where the first CSI input is a CSI input of a first data transmission layer of the first AI model at a first moment;

[0040] Determine the second CSI output by the first AI model as the first CSI input.

[0041] In combination with some embodiments of the first aspect, the second CSI is the CSI output by the first AI model at a second moment, the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

[0042] In combination with some embodiments of the first aspect, the second moment is the moment corresponding to the first moment minus N times the set period, where N is a positive integer.

[0043] In combination with some embodiments of the first aspect, the CSI transmission at the first moment is multi-layer data transmission, the second CSI is the CSI output by the first AI model at the second data transmission layer at the second moment, the second moment is the last output moment before the first moment, the output moment is the moment when the first AI model outputs the CSI, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0044] In combination with some embodiments of the first aspect, the CSI transmission at the first moment is multi-layer data transmission, the second CSI is the CSI output by the first AI model at the second data transmission layer at the first moment, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0045] In combination with some embodiments of the first aspect, the second CSI includes initial value information of the first AI model.

[0046] In combination with some embodiments of the first aspect, the second CSI is the CSI output by the first AI model at a third moment, the third moment is the last output moment within a set period and before the first moment, the output moment is the moment when the first AI model outputs the CSI, and the set period is the period for the terminal to report CSI.

[0047] In conjunction with some embodiments of the first aspect, the method further includes:

[0048] receiving first indication information sent by a network device through signaling, where the first indication information is used to instruct the terminal to report CSI based on the first indication information;

[0049] Determine the second CSI according to the first indication information, where the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

[0050] In combination with some embodiments of the first aspect, the signaling includes at least one of the following: a radio resource control RRC message, a medium access control layer-control element MAC-CE information, and downlink control information DCI.

[0051] In conjunction with some embodiments of the first aspect, the method further includes:

[0052] receiving second indication information sent by the network device;

[0053] Re-report CSI according to the second indication information.

[0054] In combination with some embodiments of the first aspect, the second indication information includes one or more non-periodic channel state reference signal CSI-RS resource information, and there is a time interval between the transmission order of the multiple CSI-RS resource information.

[0055] In combination with some embodiments of the first aspect, the generation time of the second CSI is the same as the generation time of the third CSI, and the third CSI is the CSI input of the network device corresponding to the second AI model.

[0056] In combination with some embodiments of the first aspect, the second CSI is a first CSI initial value output by the first AI model, and the third CSI is a second CSI initial value output by the second AI model.

[0057] In a second aspect, an embodiment of the present disclosure provides a method for determining CSI input of an AI model, which is performed by a network device. The method includes:

[0058] Determining that a second AI model of the network device does not have a second CSI input, where the second CSI input is a CSI input of the second AI model at the first time to the first data transmission layer;

[0059] Determine the third CSI output by the second AI model as the second CSI input.

[0060] In combination with some embodiments of the second aspect, the third CSI is the CSI output by the second AI model at a second moment, the second moment is the last output moment before the first moment, and the output moment is the moment when the second AI model outputs the CSI.

[0061] In combination with some embodiments of the second aspect, the second moment is the moment corresponding to the first moment minus N times the set period, where N is a positive integer.

[0062] In combination with some embodiments of the second aspect, the CSI transmission at the first moment is multi-layer data transmission, the third CSI is the CSI output by the second AI model at the second data transmission layer at the second moment, the second moment is the last output moment before the first moment, the output moment is the moment when the second AI model outputs the CSI, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0063] In combination with some embodiments of the second aspect, the CSI transmission at the first moment is multi-layer data transmission, the third CSI is the CSI output by the second AI model at the second data transmission layer at the first moment, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0064] In combination with some embodiments of the second aspect, the third CSI includes initial value information of the second AI model.

[0065] In combination with some embodiments of the second aspect, the third CSI is the CSI output by the second AI model at a third moment, the third moment is the last output moment within a set period and before the first moment, the output moment is the moment when the second AI model outputs the CSI, and the set period is the period for the terminal to report CSI.

[0066] In conjunction with some embodiments of the second aspect, the method further includes:

[0067] First indication information is sent to the terminal via signaling, where the first indication information is used to instruct the terminal to determine second CSI of a first AI model of the terminal based on the first indication information, where the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

[0068] In combination with some embodiments of the second aspect, the signaling includes at least one of the following: radio resource control RRC message, medium access control layer-control unit MAC-CE information and downlink control information DCI.

[0069] In conjunction with some embodiments of the second aspect, the method further includes:

[0070] Send second indication information to the terminal, where the second indication information is used to instruct the terminal to determine a first CSI input of the first AI model based on the second indication information, where the first CSI input is the CSI input of the first AI model at the first moment and the first data transmission layer.

[0071] In combination with some embodiments of the second aspect, the second indication information includes one or more non-periodic channel state reference signal CSI-RS resource information, and there is a time interval between the transmission order of the multiple CSI-RS resource information.

[0072] In combination with some embodiments of the second aspect, the generation time of the third CSI is the same as the generation time of the second CSI, and the second CSI is the CSI input of the terminal corresponding to the first AI model.

[0073] In combination with some embodiments of the second aspect, the second CSI is the first CSI initial value output by the first AI model, and the third CSI is the second CSI initial value output by the second AI model.

[0074] In a third aspect, an embodiment of the present disclosure provides a terminal, including:

[0075] A first processing module is configured to determine that a first AI model of the terminal does not have a first CSI input, where the first CSI input is a CSI input of the first AI model at a first data transmission layer at a first moment;

[0076] The second processing module is configured to determine the second CSI output by the first AI model as the first CSI input.

[0077] In a fourth aspect, an embodiment of the present disclosure provides a network device, including:

[0078] a third processing module configured to determine that the second AI model of the network device does not have a second CSI input, where the second CSI input is a CSI input of the second AI model at the first time to the first data transmission layer;

[0079] The fourth processing module is configured to determine the third CSI output by the second AI model as the second CSI input.

[0080] In a fifth aspect, an embodiment of the present disclosure provides a terminal, including:

[0081] one or more processors;

[0082] The terminal is used to execute the method for determining the AI ​​model CSI input described in any one of the first aspects of this disclosure.

[0083] In a sixth aspect, an embodiment of the present disclosure provides a network device, including:

[0084] one or more processors;

[0085] The network device is used to execute the method for determining the AI ​​model CSI input described in any one of the second aspects of this disclosure.

[0086] In the seventh aspect, an embodiment of the present disclosure proposes a communication system, comprising a terminal and a network device, wherein the terminal is configured to implement the method for determining the CSI input of the AI ​​model described in any one of the first aspect of the present disclosure, and the network device is configured to implement the method for determining the CSI input of the AI ​​model described in any one of the second aspect of the present disclosure.

[0087] In an eighth aspect, an embodiment of the present disclosure proposes a storage medium storing instructions. When the instructions are executed on a communication device, the communication device executes the method for determining the AI ​​model CSI input as described in any one of the first aspects of the present disclosure, or the communication device executes the method for determining the AI ​​model CSI input as described in any one of the second aspects of the present disclosure.

[0088] In the ninth aspect, an embodiment of the present disclosure proposes a computer program product, including a computer program and / or instructions, which, when executed by a communication device, implements the method for determining the CSI input of the AI ​​model as described in any one of the first aspects of the present disclosure, or implements the method for determining the CSI input of the AI ​​model as described in any one of the second aspects of the present disclosure when the computer program and / or instructions are executed by a communication device.

[0089] Through the above method, it is determined that the first channel state information (CSI) input to the first artificial intelligence (AI) model of the terminal does not exist, the first CSI input is the CSI input of the first AI model at the first data transmission layer at the first moment, and the second CSI output by the first AI model is determined as the first CSI input. Therefore, when the CSI input of the AI ​​model is not obtained, the other CSI outputs of the AI ​​model are used as the CSI input of the AI ​​model. This clarifies the method for determining the CSI input of the AI ​​model and ensures the reliability and accuracy of AI model reasoning during the CSI compression feedback process in the communication system.

[0090] The embodiments of the present disclosure propose a method, terminal, network device, system and medium for determining the CSI input of an AI model. In some embodiments, terms such as the method for determining the CSI input of an AI model and the communication method can be replaced with each other, terms such as the terminal and the information processing device and the communication device can be replaced with each other, and terms such as the information processing system and the communication system can be replaced with each other.

[0091] The embodiments of the present disclosure are not exhaustive and are merely illustrative of some embodiments, and are not intended to be a specific limitation on the scope of protection of the present disclosure. In the absence of contradiction, each step in a certain embodiment can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a certain embodiment can also be implemented as an independent embodiment, and the order of the steps in a certain embodiment can be arbitrarily exchanged. In addition, the optional implementation methods in a certain embodiment can be arbitrarily combined; in addition, the embodiments can be arbitrarily combined. For example, some or all steps of different embodiments can be arbitrarily combined, and a certain embodiment can be arbitrarily combined with the optional implementation methods of other embodiments.

[0092] In each embodiment of the present disclosure, unless otherwise specified or provided for by logic, the terms and / or descriptions between the embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form a new embodiment based on their inherent logical relationships.

[0093] The terms used in the embodiments of the present disclosure are only for the purpose of describing specific embodiments and are not intended to limit the present disclosure.

[0094] In the embodiments of the present disclosure, unless otherwise specified, elements expressed in the singular, such as "a", "an", "the", "above", "said", "the", "the", etc., may mean "one and only one", or "one or more", "at least one", etc. For example, when using articles such as "a", "an", "the" in English in translation, the noun following the article may be understood as a singular expression or a plural expression.

[0095] In the embodiments of the present disclosure, “plurality” refers to two or more.

[0096] In some embodiments, the terms "at least one," "one or more," "a plurality of," "multiple," etc. may be used interchangeably.

[0097] In some embodiments, descriptions such as "at least one of A and B," "A and / or B," "A in one case, B in another case," or "in response to one case A, in response to another case B" may include the following technical solutions depending on the situation: in some embodiments, A (A is executed independently of B); in some embodiments, B (B is executed independently of A); in some embodiments, execution is selected from A and B (A and B are selectively executed); and in some embodiments, A and B (both A and B are executed). The above is also applicable when there are more branches such as A, B, and C.

[0098] In some embodiments, "A or B" and other descriptions may include the following technical solutions depending on the situation: in some embodiments, A (A is executed independently of B); in some embodiments, B (B is executed independently of A); in some embodiments, execution is selected from A and B (A and B are selectively executed). The above is also applicable when there are more branches such as A, B, C, etc.

[0099] The prefixes such as "first" and "second" in the embodiments of the present disclosure are only used to distinguish different description objects and do not constitute any restriction on the position, order, priority, quantity or content of the description objects. For the statement of the description object, please refer to the description in the context of the claims or embodiments, and the use of prefixes should not constitute unnecessary restrictions. For example, if the description object is a "field", the ordinal number before the "field" in the "first field" and the "second field" does not limit the position or order between the "fields". "First" and "second" do not limit whether the "fields" they modify are in the same message, nor do they limit the order of the "first field" and the "second field". For another example, if the description object is a "level", the ordinal number before the "level" in the "first level" and the "second level" does not limit the priority between the "levels". For another example, the number of description objects is not limited by the ordinal number and can be one or more. Taking "first device" as an example, the number of "devices" can be one or more. In addition, the objects modified by different prefixes can be the same or different. For example, if the description object is "device", then the "first device" and the "second device" can be the same device or different devices, and their types can be the same or different; for another example, if the description object is "information", then the "first information" and the "second information" can be the same information or different information, and their contents can be the same or different.

[0100] In some embodiments, “including A,” “comprising A,” “used to indicate A,” and “carrying A” can be interpreted as directly carrying A or indirectly indicating A.

[0101] In some embodiments, terms such as "time / frequency" and "time / frequency domain" refer to the time domain and / or the frequency domain.

[0102] In some embodiments, terms such as "in response to...", "in response to determining...", "in the case of...", "at the time of...", "when...", "if...", "if...", etc. can be used interchangeably.

[0103] In some embodiments, terms such as "greater than", "greater than or equal to", "not less than", "more than", "more than or equal to", "not less than", "higher than", "higher than or equal to", "not less than", and "above" can be replaced with each other, and terms such as "less than", "less than or equal to", "not greater than", "less than", "less than or equal to", "not more than", "lower than", "lower than or equal to", "not higher than", and "below" can be replaced with each other.

[0104] In some embodiments, devices, etc. can be interpreted as physical or virtual, and their names are not limited to the names recorded in the embodiments. Terms such as "device", "equipment", "device", "circuit", "network element", "node", "function", "unit", "section", "system", "network", "chip", "chip system", "entity", and "subject" can be used interchangeably.

[0105] In some embodiments, "network" can be interpreted as devices included in the network (eg, access network equipment, core network equipment, etc.).

[0106] In some embodiments, the terms "access network device (AN device)", "radio access network device (RAN device)", "base station (BS)", "radio base station" "fixed station", "node", "access point", "transmission point (TP)", "reception point (RP)", "transmission / reception point (TRP)" "panel", "antenna panel", "antenna array", "cell", "macro cell", "small cell", "femto cell", "pico cell", "sector", "cell group", "serving cell", "carrier", "component carrier", "bandwidth part (BWP)" and the like may be used interchangeably.

[0107] In some embodiments, the terms "terminal", "terminal device", "user equipment (UE)", "user terminal", "mobile station (MS)", "mobile terminal (MT)", subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, etc. can be used interchangeably.

[0108] In some embodiments, the access network device, the core network device, or the network device can be replaced by a terminal. For example, the various embodiments of the present disclosure can also be applied to a structure in which the communication between the access network device, the core network device, or the network device and the terminal is replaced by communication between multiple terminals (for example, device-to-device (D2D), vehicle-to-everything (V2X), etc.). In this case, it is also possible to set the structure in which the terminal has all or part of the functions of the access network device. In addition, terms such as "uplink" and "downlink" can also be replaced by terms corresponding to communication between terminals (for example, "side"). For example, uplink channels, downlink channels, etc. can be replaced by side channels, and uplinks, downlinks, etc. can be replaced by side links.

[0109] In some embodiments, the terminal may be replaced by an access network device, a core network device, or a network device. In this case, the access network device, the core network device, or the network device may have a structure that has all or part of the functions of the terminal.

[0110] In some embodiments, obtaining data, information, etc. may comply with the laws and regulations of the country where the data is obtained.

[0111] In some embodiments, data, information, etc. may be obtained with the user's consent.

[0112] In addition, each element, each row, or each column in the table of the embodiment of the present disclosure can be implemented as an independent embodiment, and the combination of any elements, any rows, and any columns can also be implemented as an independent embodiment.

[0113] FIG1a is a schematic diagram illustrating the architecture of a communication system according to an embodiment of the present disclosure. As shown in FIG1a , a communication system 100 includes a terminal 101 and a network device 102 .

[0114] In some embodiments, the terminal 101 includes, for example, a mobile phone, a wearable device, an Internet of Things device, a car with communication function, a smart car, a tablet computer, a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal device in industrial control, a wireless terminal device in self-driving, a wireless terminal device in remote medical surgery, a wireless terminal device in a smart grid, a wireless terminal device in transportation safety, a wireless terminal device in a smart city, and at least one of a wireless terminal device in a smart home, but is not limited thereto.

[0115] In some embodiments, the network device 102 is, for example, a node or device that accesses the terminal to a wireless network. The access network device may include an evolved NodeB (eNB), a next generation evolved NodeB (ng-eNB), a next generation NodeB (gNB), a node B (NB), a home node B (HNB), a home evolved nodeB (HeNB), a wireless backhaul device, a radio network controller (RNC), a base station controller (BSC), a base transceiver station (BTS), a base band unit (BBU), a mobile switching center, a base station in a 6G communication system, an open base station (Open RAN), a cloud base station (Cloud RAN), a base station in other communication systems, and at least one of an access node in a Wi-Fi system, but is not limited thereto.

[0116] In some embodiments, the technical solution of the present disclosure can be applied to the Open RAN architecture. In this case, the interfaces between or within the access network devices involved in the embodiments of the present disclosure can be transformed into internal interfaces of the Open RAN, and the processes and information interactions between these internal interfaces can be implemented through software or programs.

[0117] In some embodiments, the access network device can be composed of a centralized unit (CU) and a distributed unit (DU), where the CU can also be called a control unit. The CU-DU structure can be used to split the protocol layer of the access network device, with the functions of some protocol layers centrally controlled by the CU, and the functions of the remaining part or all of the protocol layers distributed in the DU, which is centrally controlled by the CU, but is not limited to this.

[0118] It can be understood that the communication system described in the embodiment of the present disclosure is for the purpose of more clearly illustrating the technical solution of the embodiment of the present disclosure, and does not constitute a limitation on the technical solution proposed in the embodiment of the present disclosure. Ordinary technicians in this field can know that with the evolution of the system architecture and the emergence of new business scenarios, the technical solution proposed in the embodiment of the present disclosure is also applicable to similar technical problems.

[0119] The following embodiments of the present disclosure may be applied to the communication system 100 shown in FIG1 , or a portion thereof, but are not limited thereto. The entities shown in FIG1 are illustrative only. The communication system may include all or part of the entities shown in FIG1 , or may include other entities outside of FIG1 . The number and form of the entities are arbitrary, and the entities may be physical or virtual. The connection relationships between the entities are illustrative only. The entities may be connected or disconnected, and the connection may be in any manner, including direct or indirect, wired or wireless.

[0120] The embodiments of the present disclosure can be applied to Long Term Evolution (LTE), LTE-Advanced (LTE-A), LTE-Beyond (LTE-B), SUPER 3G, IMT-Advanced, 4th generation mobile communication system (4G), 5th generation mobile communication system (5G), 5G new radio (NR), future radio access (FRA), new radio access technology (RAT), new radio (NR), new radio access (NX), future generation radio access (FX), Global System for Mobile communications (GSM (registered trademark)), CDMA2000, Ultra Mobile Broadband (UMB), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, Ultra-WideBand (UWB), Bluetooth (registered trademark), Public Land Mobile Network (PLMN) networks, Device-to-Device (D2D) systems, Machine-to-Machine (M2M) systems, Internet of Things (IoT) systems, Vehicle-to-Everything (V2X), systems utilizing other communication methods, and next-generation systems based on and extending these methods. Furthermore, multiple systems may be combined (for example, a combination of LTE or LTE-A with 5G).

[0121] In some embodiments, in the space-frequency-time domain CSI compression feedback process of the communication system, in order to improve the CSI compression performance, the correlation of the time domain channel can be utilized to implement space-frequency-time domain CSI compression based on the bilateral model of the encoder and decoder. For example, FIG1b is a schematic diagram of CSI compression feedback according to an embodiment of the present disclosure. As shown in FIG1b, the encoder part model on the UE (User Equipment) side includes, in addition to the input of the channel information to be compressed H, the historical CSI input a t , the historical CSI input a t The CSI output by the encoder in the previous CSI reporting cycle before the current moment. Correspondingly, the decoder model on the NW (Network side) not only inputs the quantized binary bit stream information, but also includes the historical CSI input b t Similarly, the historical CSI input b t The CSI output by the decoder in the previous CSI reporting cycle before the current moment. That is, in the bilateral model composed of the encoder and decoder, both the encoder and decoder need to use the CSI output in the previous CSI reporting cycle as the historical CSI input for the next moment, and then recover the historical downlink information approximation H' of the corresponding moment through the encoder or decoder inference. The binary bit stream information is obtained by inputting the historical CSI into a t The codeword is input into the encoder and compressed and encoded on the channel information H, and then the codeword is output after bit quantization.

[0122] In some embodiments, for the Lth layer data transmission at time K in the CSI compression feedback process of multi-layer data streams, if at time KT (T is the reporting period of binary bit stream information) before time K, there is no historical CSI input a of the encoder in the bilateral model of the encoder and decoder. t and / or the historical CSI input b of the decoder t , then the UE and NW need to define the historical CSI input a of the Encoder during the L-th layer data transmission at time K t , and the historical CSI input b of the Decoder t , and ensure that the Encoder's historical CSI input a t With the historical CSI input b of the Decoder t Correspondingly, that is, a t and b tThe CSI output of the encoder and the CSI output of the decoder at the same time are used, so that the UE encoder and the NW decoder can obtain reliable historical CSI input, ensuring the reliability and consistency of model inference during the CSI compression feedback process.

[0123] Figure 2 is an interactive diagram of a method for determining CSI input of an AI model according to an embodiment of the present disclosure. As shown in Figure 2, the embodiment of the present disclosure relates to a method for determining CSI input of an AI model, which is executed by a terminal and a network device. The method includes:

[0124] In step S2101, the network device determines that the second AI model does not have a second CSI input, and sends first indication information to the terminal via signaling.

[0125] In some embodiments, the second CSI input is the CSI input of the first data transmission layer of the second AI model at the first moment.

[0126] In some embodiments, the first moment may be the current moment or a moment in the future.

[0127] In some embodiments, the first indication information is used to instruct the terminal to re-measure and report the CSI based on the first indication information.

[0128] For example, the first indication information is used to instruct the terminal to remeasure and report the first CSI input of the first data transmission layer at the first moment in the first AI model.

[0129] In some embodiments, the name of the first indication information is not limited, and it may be, for example: CSI reporting indication information, CSI-RS information, CSI resource transmission information, etc.

[0130] For example, assuming that when the terminal reports the binary bits stream information compressed by the first AI model at the first moment, due to insufficient uplink transmission resources or a conflict between the currently reported binary bits stream information and other CSI reports, the reported binary bits stream information will be incomplete. The second AI model of the network device is unable to obtain the second CSI input of the current data transmission layer at the current moment based on the binary bits stream information. When the second AI model of the network device infers the binary bits stream information, it may cause the output CSI to be inaccurate or the corresponding CSI to be unable to be output. At this time, the network device can send a non-periodic CSI-RS information to the terminal, so that the terminal re-measures and reports the corresponding binary bits stream information based on the CSI-RS information, so that the second AI model in the network device infers the approximate value H' of the original downlink information based on the re-reported binary bits stream information and the historical CSI input at the first moment.

[0131] The second CSI input corresponding to the second AI model is the CSI output by the second AI model in the previous CSI reporting period before the current moment. Typically, the second AI model uses the CSI output at the current moment as the historical CSI input for the next moment. However, if the current network environment does not require the second AI model to output the CSI for the current data transmission layer at the current moment, historical CSI input for the current data transmission layer may not exist at the next moment. Therefore, it is necessary to indicate the historical CSI input for the next moment.

[0132] In some embodiments, the signaling includes at least one of the following: RRC information, MAC-CE information, and DCI information.

[0133] For example, the first indication information can be carried by at least one of RRC information, MAC-CE information and DCI information, and the network device indicates the first indication information to the terminal through RRC information, MAC-CE information and DCI information.

[0134] In step S2102, the terminal determines the first CSI input and channel information of the first AI model based on the first indication information, generates CSI according to the first CSI input and channel information, and sends it to the network device.

[0135] For example, in this embodiment, the first CSI input is the historical CSI input of the data transmission layer of the first AI model at the current moment. Based on the first indication information, the terminal remeasures the channel state, determines the channel information and the historical CSI input of the first AI model, encodes and compresses the channel information using the historical CSI input as input to the first AI model, and generates a binary bits stream that is sent to the network device.

[0136] It should be noted that in this embodiment, the first indication information is used to trigger the terminal to re-measure and report. The terminal can determine the historical CSI input of the first AI model based on a predefined protocol. For example, the CSI output by the first AI model at the previous moment can be used as the historical CSI input at the current moment, where the previous moment is the output moment in the previous CSI reporting cycle before the current moment, and the output moment is the moment when the first AI model outputs CSI. Optionally, if the terminal did not output historical CSI by the first AI model at the previous moment, the CSI output by the first AI model at the most recent historical moment can be used as the historical CSI input. The initial value of the first AI model can also be used as the historical CSI input. The historical CSI of the second data layer output by the first AI model at the most recent historical moment can also be used as the historical CSI input for the first data transmission layer at the first moment. This embodiment does not limit the manner in which the terminal determines the historical CSI input.

[0137] In the process of encoding by the first AI model of the terminal and decoding by the second AI model of the network device, the reliability and consistency of the first AI model and the second AI model need to be guaranteed. Therefore, it is necessary to keep the historical CSI input a of the first AI model. t With the historical CSI input b of the second AI model t The consistency between them corresponds to the determination of the historical CSI input a based on the first AI model in the above embodiment. t The second AI model determines the historical CSI input b t For example, when the terminal is connected to the most recent historical moment t r , the CSI output by the first AI model is used as the historical CSI input a t When the network device also sends the most recent historical moment t r , the CSI output by the second AI model is used as the historical CSI input b t Optionally, the terminal uses the first CSI initial value of the first AI model as the historical CSI input a t , the network device uses the second CSI initial value of the second AI model as the historical CSI input b t .

[0138] In step S2103 , the network device determines, based on the CSI, the third CSI of the second AI model as the second CSI input, and generates a historical CSI output according to the CSI and the third CSI.

[0139] For example, in this embodiment, the first CSI input of the first AI model is the historical CSI input a of the first AI model. t , the second CSI input of the second AI model is the historical CSI input b of the second AI model t The network device determines the third CSI of the first data transmission layer of the second AI model at the first moment as the second CSI input based on the binary bit stream information re-fed back by the terminal, wherein the network device determines the second CSI input b t The method is the same as that in the above embodiment where the terminal determines the first CSI input a t The method is the same as that of step S2102 above, which will not be repeated here. The second AI model of the network device generates a historical CSI output based on the CSI and the historical CSI input. The historical CSI output can be used as the historical CSI input of the second AI model at the next moment for subsequent second AI model reasoning.

[0140] Through the above method, when the network device cannot infer the corresponding CSI based on the CSI reported by the terminal or the inferred CSI is inaccurate, the terminal can be instructed to re-measure and report the CSI by means of an instruction report, so that the first AI model of the terminal and the second AI model of the network device can both output the historical CSI output at the corresponding moment, which serves as the historical CSI input of the AI ​​model at the next moment for subsequent AI model inference, ensuring the reliability and consistency of the CSI compression feedback process of the communication system.

[0141] Figure 3 is a flow chart of a method for determining CSI input of an AI model according to an embodiment of the present disclosure. As shown in Figure 3, the embodiment of the present disclosure relates to a method for determining CSI input of an AI model, which is executed by a terminal. The method includes:

[0142] Step S3101: Determine whether the first AI model of the terminal has a first CSI input.

[0143] In some embodiments, the first CSI input is the CSI input of the first AI model at the first data transmission layer at the first moment.

[0144] For example, this embodiment is applied to realize compressed feedback of space-frequency-time domain CSI based on a bilateral model of encoding and decoding. A first AI model for encoding is configured on the terminal side, and a second AI model for decoding is configured on the network device side. The terminal measures the channel state at the current moment according to the signaling sent by the network device or the set communication protocol in the communication system, and generates the channel information H to be compressed. The channel information H is input into the first AI model for encoding and compression. In addition to the channel information H to be compressed, the first AI model on the terminal side also needs to input the historical CSI output by the first AI model in the previous reporting cycle as the historical CSI input a of the first AI model at the current moment. t , input the historical CSI into a t The first AI model encodes and compresses the channel information H to be compressed, and then generates binary bits stream information and sends it to the network device side. Correspondingly, the second AI model on the network device side is a decoding model. In addition to the input of the quantized binary bits stream information, it also includes the historical CSI output by the second AI model in the previous reporting cycle as the historical CSI input b of the second AI model at the current moment. t The second AI model recovers the original downlink information approximation H' at the corresponding moment through reasoning based on the above input information.

[0145] In some embodiments, the above embodiments need to ensure that the historical CSI input a of the first AI model input t The historical CSI input b is input to the second AI model tThe consistency between them, that is, the historical CSI input a t With historical CSI input b t The CSI input is of the same nature at the same time, so as to ensure that the channel information H reported by the terminal is consistent with the downlink information H' inferred and recovered by the network equipment.

[0146] In some embodiments, the CSI compression feedback between the terminal and the network device is multi-layer data stream compression feedback, and the process of performing the CSI compression feedback includes CSI compression feedback of multiple layers of data.

[0147] For example, the terminal will a t During the encoding and compression process of H input to the first AI model, due to various reasons such as insufficient uplink transmission resources, a conflict between the CSI currently required to be reported and other CSI reports, or the current communication system interaction environment not requiring the terminal to report the CSI of the corresponding data layer, the historical CSI input for the corresponding data layer at the next moment may not be generated by the first AI model when the binary bits stream information is reported at the current moment. When the CSI of the corresponding data layer needs to be encoded and compressed by the first AI model at the next moment, the historical CSI input for the corresponding data layer cannot be obtained. Therefore, the historical CSI input for the next moment needs to be set. In addition, to ensure that the historical CSI input of the corresponding data layer input by the first AI model of the terminal and the second AI model of the network device at the same moment is consistent, when the historical CSI input of the corresponding data layer cannot be obtained, the terminal and the network device determine the historical CSI input in the same manner.

[0148] In some embodiments, the terminal needs to input the first CSI input of the first data transmission layer at the first moment corresponding to the previous historical moment into the first AI model based on the transmission requirements of the current network environment. When the first CSI input does not exist in the first AI model, in order to avoid performance degradation of CSI compression transmission, the historical CSI input of the first data transmission layer at the first moment can be set.

[0149] For example, when setting the historical CSI input of the terminal, in order to ensure that the network device can successfully parse the binary bits stream information to obtain the corresponding downlink information approximation H', it is necessary to ensure the consistency between the historical CSI input in the first AI model and the historical CSI input in the second AI model of the network device. That is, the setting rules of the historical CSI input of the first AI model in the terminal are consistent with the setting rules of the historical CSI input of the second AI model in the network device.

[0150] Step S3102: Receive first indication information sent by the network device through signaling, where the first indication information is used to instruct the terminal to report CSI based on the first indication information.

[0151] Step S3103: Determine the second CSI according to the first indication information.

[0152] In some embodiments, the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

[0153] For example, in this embodiment, before a first moment, a network device sends first indication information to a terminal via signaling. The first indication information is used to instruct the terminal to report CSI of a first data transmission layer corresponding to a second moment before the first moment. The terminal determines, based on the first indication information, second CSI at the second moment as the first CSI input.

[0154] Optionally, the signaling includes at least one of the following: RRC (Radio Resource Control, radio resource control protocol) message, MAC-CE (MAC-Control Element, media access control layer-control unit) information and DCI (Downlink Control Information, downlink control information).

[0155] For example, the first indication information may be carried by at least one of an RRC message, MAC-CE information, and DCI.

[0156] In some embodiments, the method further comprises:

[0157] receiving second indication information sent by the network device;

[0158] Re-report CSI according to the second indication information.

[0159] For example, in this embodiment, the terminal has a historical CSI input in the first AI model output at the previous moment relative to the first moment, but due to insufficient uplink transmission resources or other CSI reporting conflicts during the CSI reporting of the first data transmission layer at the first moment, the network device cannot infer the corresponding CSI based on the reported binary bits stream information, or the determined CSI is inaccurate. The network device can send a second indication message to instruct the terminal to re-measure and report. Correspondingly, the second CSI determined by the terminal based on the second indication information is the historical CSI input. The network device can determine the historical CSI input of the second AI model based on the re-measured and reported binary bits stream information, using the method of the terminal determining the historical CSI input.

[0160] Optionally, in some embodiments, the second indication information includes one or more non-periodic channel state reference signal CSI-RS resource information.

[0161] In some embodiments, there is a time interval between the transmission sequences of multiple CSI-RS resource information.

[0162] For example, the second indication information includes one or more aperiodic CSI-RS resources. The terminal can re-measure and report based on the second indication information, and carry binary bits stream information based on the one or more aperiodic CSI-RS resources. This enables the first AI model of the terminal and the second AI model of the network device to both output historical CSI input at the corresponding time for subsequent inference by the first and second AI models.

[0163] Step S3104: Determine the second CSI output by the first AI model as the first CSI input.

[0164] For example, if it is determined that the terminal does not have the first CSI input for the first data transmission layer of the first AI model at the first moment, the second CSI input to the first AI model before the first moment can be used as the first CSI input for the first data transmission layer at the current first moment, or the second CSI output from the first AI model at the first moment for another data transmission layer can be used as the first CSI input, where the first CSI input is the historical CSI input for the first data transmission layer of the first AI model at the first moment. Correspondingly, if the terminal does not have the first CSI input in the network device, the second AI model of the corresponding network device also does not have the second CSI input for the first data transmission layer at the first moment. Therefore, the network device can also use the third CSI output from the second AI model before the first moment as the historical CSI input for the first data transmission layer at the current first moment, or the second CSI output from the second AI model at the first moment for another data transmission layer as the second CSI input. The second CSI of the first AI model and the third CSI of the second AI model are generated at the same time and in the same manner.

[0165] In some embodiments, the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

[0166] For example, in this embodiment, when it is determined that the first CSI input does not exist, the second CSI output by the first data transmission layer at the last output moment before the first moment can be used as the historical CSI input for the first data transmission layer at the first moment. Correspondingly, on the network device side, the third CSI output by the first data transmission layer at the most recent second moment is used as the historical CSI input for the first data transmission layer of the second AI model at the first moment.

[0167] Optionally, in some embodiments, the above step S2102 includes:

[0168] The second moment is the moment corresponding to the first moment minus N times the set period, where N is a positive integer.

[0169] For example, in the communication system of this embodiment, the terminal can be required to perform periodic or semi-continuous CSI reporting by setting a protocol, so the corresponding second moment can be the first moment minus N times the moment corresponding to the set period, where N is a positive integer, and the set period is the period for the terminal to perform CSI reporting specified by the protocol.

[0170] It should be noted that this embodiment is applicable to the situation where CSI reporting is performed periodically or semi-continuously, and the corresponding second moment is the moment corresponding to the first moment minus N times the set period. However, for the case of non-periodic CSI reporting, the second moment is the CSI reporting moment closest to the first moment, which is the non-periodic reporting moment.

[0171] In some embodiments, the CSI transmission at the first moment is multi-layer data transmission, the second CSI is the CSI output by the first AI model at the second data transmission layer at the second moment, the second moment is the last output moment before the first moment, the output moment is the moment when the first AI model outputs the CSI, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0172] For example, when it is determined that the CSI transmission at the first moment is multi-layer data transmission, that is, when the Rank of the CSI transmission is greater than 1, the second CSI of the second data transmission layer output by the first AI model at the second moment closest to the first moment is used as the first CSI input, and the first CSI input can be used as the historical CSI input of the first data transmission layer of the first AI model at the first moment. The index of the second data transmission layer is lower than the index of the first data transmission layer. For example, if the index of the first data transmission layer is 4, the second CSI output by the first AI model corresponding to the second moment of the first layer, the second layer, or the third layer can be used as the historical CSI input. For example, if the CSI transmission of the current terminal at the first moment is 4-layer data transmission, the corresponding Rank level is 4, and the index of the first data transmission layer is 3, the second CSI of the second data transmission layer at the second moment closest to the first moment can be used as the historical CSI input. The index of the second data transmission layer is lower than the index of the first data transmission layer. For example, the index of the second data transmission layer can be 2 or 1.

[0173] In some embodiments, the CSI transmission at the first moment is multi-layer data transmission, the second CSI is the CSI output by the first AI model at the second data transmission layer at the first moment, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0174] For example, in this embodiment, the second CSI of the second data transmission layer at the first moment can also be used as the historical CSI input of the first data transmission layer at the first moment to the first AI model. For example, if it is determined that the terminal's CSI transmission is multi-layer data transmission and the first AI model cannot obtain the first CSI input at the fourth layer at the first moment, the second CSI of the third or second layer output by the first AI model at the first moment can be used as the historical CSI input of the fourth layer at the first moment.

[0175] In some embodiments, the second CSI includes initial value information of the first AI model.

[0176] For example, the second CSI can be assigned as the initial assignment information of the first AI model, and the corresponding historical CSI input of the second AI model on the network device side can also be assigned as the initial assignment information of the second AI model.

[0177] In some embodiments, the second CSI is the CSI output by the first AI model at a third moment, where the third moment is the last output moment within a set period and before the first moment, the output moment is the moment when the first AI model outputs the CSI, and the set period is the period for the terminal to report the CSI.

[0178] For example, the terminal sets a CSI reporting period so that the terminal reports the CSI corresponding to the first data transmission layer at least once within a period before the first moment, that is, the terminal reports CSI based on the set period. In this embodiment, the terminal determines the previous historical set period before the first moment, and determines, within the historical set period, the second CSI of the first data transmission layer output by the first AI model closest to the first moment as the historical CSI input of the first AI model.

[0179] In some embodiments, the generation time of the second historical CSI input information is the same as the generation time of the third historical CSI input information of the second AI model of the network device, and the third historical CSI input information is the target historical CSI input information of the first data transmission layer of the second AI model at the first moment.

[0180] For example, in this embodiment, during the determination of the historical CSI input, it is necessary to ensure the consistency between the second CSI of the first AI model of the terminal and the third CSI of the second AI model of the network device, wherein the third CSI is the historical CSI input of the second AI model at the first data transmission layer at the first moment. Correspondingly, the generation time of the second CSI is the same as the generation time of the third historical CSI. For example, the first AI model will be the most recent t r The second CSI output at the moment is used as the historical CSI input Then the second AI model also sets t rThe third CSI output is used as the historical CSI input

[0181] Optionally, in some embodiments, the second historical CSI input information is a first CSI initial value output by the first AI model, and the third CSI is a second CSI initial value output by the second AI model.

[0182] For example, in this embodiment, when it is determined that the first AI model cannot obtain the first CSI input output at the previous moment, or the second AI model cannot obtain the second CSI input output at the previous moment, the first CSI initial value output by the first AI model can be used as the historical CSI input, and the second CSI initial value of the second AI model can be used as the historical CSI input.

[0183] Through the above method, it is determined that the first channel state information (CSI) input to the first artificial intelligence (AI) model of the terminal does not exist, the first CSI input is the CSI input of the first AI model at the first data transmission layer at the first moment, and the second CSI output by the first AI model is determined as the first CSI input. Therefore, when the CSI input of the AI ​​model is not obtained, the other CSI outputs of the AI ​​model are used as the CSI input of the AI ​​model. This clarifies the method for determining the CSI input of the AI ​​model and ensures the reliability and accuracy of AI model reasoning during the CSI compression feedback process in the communication system.

[0184] In some embodiments, the names of information, etc. are not limited to the names described in the embodiments, and terms such as "information", "message", "signal", "signaling", "report", "configuration", "indication", "instruction", "command", "channel", "parameter", "domain", "field", "symbol", "symbol", "codeword", "codebook", "codeword", "codepoint", "bit", "data", "program", and "chip" can be used interchangeably.

[0185] In some embodiments, the terms "codebook," "codeword," and "precoding matrix" may be used interchangeably. For example, a codebook may be a collection of one or more codewords / precoding matrices.

[0186] In some embodiments, terms such as "uplink", "uplink", "physical uplink" can be interchangeable with each other, and terms such as "downlink", "downlink", "physical downlink" can be interchangeable with each other, and terms such as "side", "sidelink", "side communication", "sidelink communication", "direct connection", "direct link", "direct communication", "direct link communication" can be interchangeable with each other.

[0187] In some embodiments, the terms "downlink control information (DCI)", "downlink (DL) assignment", "DL DCI", "uplink (UL) grant", "UL DCI" and the like may be used interchangeably.

[0188] In some embodiments, terms such as "physical downlink shared channel (PDSCH)" and "DL data" can be used interchangeably, and terms such as "physical uplink shared channel (PUSCH)" and "UL data" can be used interchangeably.

[0189] In some embodiments, the terms "radio", "wireless", "radio access network (RAN)", "access network (AN)", "RAN-based" and the like may be used interchangeably.

[0190] In some embodiments, terms such as "synchronization signal (SS)", "synchronization signal block (SSB)", "reference signal (RS)", "pilot", and "pilot signal" can be used interchangeably.

[0191] In some embodiments, terms such as "moment", "time point", "time", and "time position" can be replaced with each other, and terms such as "duration", "period", "time window", "window", and "time" can be replaced with each other.

[0192] In some embodiments, the terms "precoding", "precoder", "weight", "precoding weight", "quasi-co-location (QCL)", "transmission configuration indication (TCI) state", "spatial relation", "spatial domain filter", "transmission power", "phase rotation", "antenna port", "antenna port group", "layer", "the number of layers", "rank", "resource", "resource set", "resource group", "beam", "beam width", "beam angular degree", "antenna", "antenna element", "panel" and the like can be used interchangeably.

[0193] In some embodiments, terms such as "frame", "radio frame", "subframe", "slot", "sub-slot", "mini-slot", "symbol", "symbol", and "transmission time interval (TTI)" can be used interchangeably.

[0194] In some embodiments, "obtain", "get", "get", "receive", "transmit", "bidirectional transmission", "send and / or receive" can be interchangeable, and can be interpreted as receiving from other entities, obtaining from protocols, obtaining from higher layers, obtaining by self-processing, autonomous implementation, etc.

[0195] In some embodiments, terms such as "send", "transmit", "report", "download", "transmit", "bidirectional transmission", "send and / or receive" can be used interchangeably.

[0196] In some embodiments, terms such as "certain", "preset", "preset", "setting", "indicated", "a certain", "any", and "first" can be interchangeable. "Specific A", "preset A", "preset A", "setting A", "indicated A", "a certain A", "any A", and "first A" can be interpreted as A pre-specified in a protocol, etc., or as A obtained through setting, configuration, or indication, etc., or as specific A, a certain A, any A, or first A, etc., but not limited to this.

[0197] In some embodiments, the determination or judgment can be performed by a value represented by 1 bit (0 or 1), or by a true or false value (Boolean value) represented by true or false, or by comparison of numerical values ​​(for example, comparison with a predetermined value), but is not limited thereto.

[0198] In some embodiments, "not expecting to receive" can be interpreted as not receiving on time domain resources and / or frequency domain resources, or as not performing subsequent processing on the data after receiving it; "not expecting to send" can be interpreted as not sending, or as sending but not expecting the recipient to respond to the content sent.

[0199] The communication method involved in the embodiment of the present disclosure may include at least one of steps S3101 to S3104. For example, step S3104 may be implemented as an independent embodiment, and steps S3101+S3104 may be implemented as independent embodiments, but are not limited thereto.

[0200] In some embodiments, steps S3101 , S3102 , and S3103 may be executed in an interchangeable order or simultaneously.

[0201] In some embodiments, step S3102 and step S3103 are optional, and one or more of these steps may be omitted or replaced in different embodiments.

[0202] In some embodiments, reference may be made to other optional implementations described before or after the description corresponding to FIG. 3 .

[0203] FIG4 is a flow chart of a method for determining CSI input of an AI model according to an embodiment of the present disclosure. As shown in FIG4 , an embodiment of the present disclosure relates to a method for determining CSI input of an AI model, which is executed by a network device and includes:

[0204] Step S4101: Determine whether the second AI model of the network device has no second CSI input.

[0205] In some embodiments, the second CSI input is the CSI input of the first data transmission layer of the second AI model at the first moment.

[0206] For example, the historical CSI input of the network device in this embodiment corresponds to the historical CSI input of the terminal in the above embodiment. The second AI model in the network device is used to decode the binary bits stream information compressed and encoded by the first AI model and reported by the terminal to output the approximate value H' of the original downlink information at the corresponding moment. When the first AI model of the terminal cannot obtain the first historical CSI input of the first data transmission layer at the first moment, the second AI model on the corresponding network device side is also unable to obtain the first historical CSI input of the first data transmission layer at the first moment. Therefore, it is necessary to set the historical CSI input to ensure the stability of the AI ​​model during the CSI data transmission process. At the same time, the setting method of the historical CSI input in the first AI model is the same as the setting method of the historical CSI input in the second AI model, so as to ensure the consistency of the AI ​​model during the CSI data transmission process.

[0207] Step S4102: Send first indication information to the terminal via signaling.

[0208] In some embodiments, the first indication information is used to instruct the terminal to determine the second CSI of the first AI model of the terminal based on the first indication information, where the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

[0209] In some embodiments, the signaling includes at least one of the following: an RRC message, MAC-CE information, and DCI.

[0210] The optional implementation of step S4102 can refer to the optional implementation of step S3102 and step S103 in FIG3 and other related parts in the embodiment involved in FIG3, which will not be repeated here.

[0211] Step S4103: Determine the third CSI output by the second AI model as the second CSI input.

[0212] In some embodiments, the third CSI is the CSI output by the second AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the second AI model outputs the CSI.

[0213] In some embodiments, the second moment is the moment corresponding to the first moment minus N times the set period, where N is a positive integer.

[0214] In some embodiments, the CSI transmission at the first moment is multi-layer data transmission, the third CSI is the CSI output by the second AI model at the second data transmission layer at the second moment, the second moment is the last output moment before the first moment, the output moment is the moment when the second AI model outputs the CSI, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0215] In some embodiments, the CSI transmission at the first moment is multi-layer data transmission, the third CSI is the CSI output by the second AI model at the second data transmission layer at the first moment, and the index of the second data transmission layer is less than the index of the first data transmission layer.

[0216] In some embodiments, the third CSI includes initial value information of the second AI model.

[0217] In some embodiments, the third CSI is the CSI output by the second AI model at a third moment, where the third moment is the last output moment within a set period and before the first moment, the output moment is the moment when the second AI model outputs the CSI, and the set period is the period for the terminal to report the CSI.

[0218] Optionally, in some embodiments, the method further comprises:

[0219] Send second indication information to the terminal, where the second indication information is used to instruct the terminal to determine a first CSI input of the first AI model based on the second indication information, where the first CSI input is the CSI input of the first data transmission layer of the first AI model at the first moment.

[0220] In some embodiments, the second indication information includes one or more non-periodic channel state reference signal CSI-RS resource information, and there is a time interval between the transmission sequence of multiple CSI-RS resource information.

[0221] In some embodiments, the generation time of the third CSI is the same as the generation time of the second CSI, and the second CSI is the CSI input of the terminal corresponding to the first AI model.

[0222] In some embodiments, the second CSI is a first CSI initial value output by the first AI model, and the third CSI is a second CSI initial value output by the second AI model.

[0223] The optional implementation of step S4103 can refer to the optional implementation of step S3104 in Figure 3 and other related parts in the embodiment involved in Figure 3, which will not be repeated here.

[0224] Through the above method, it is determined that the second AI model of the network device does not have a second CSI input, that the second CSI input is the CSI input of the second AI model at the first data transmission layer at the first moment, and that the third CSI output by the second AI model is determined as the second CSI input. Therefore, when the AI ​​model's CSI input is not obtained, the AI ​​model's other CSI outputs are used as the AI ​​model's CSI input. This clarifies the method for determining the AI ​​model's CSI input and ensures the reliability and accuracy of AI model reasoning during the CSI compression feedback process in the communication system.

[0225] The communication method involved in the embodiments of the present disclosure may include at least one of steps S4101 to S4103. For example, step S4103 may be implemented as an independent embodiment, steps S4101+S4103 may be implemented as an independent embodiment, and steps S4102+S4103 may be implemented as an independent embodiment, but are not limited thereto.

[0226] In some embodiments, steps S4101 , S4102 , and S4103 may be executed in an interchangeable order or simultaneously.

[0227] In some embodiments, step S4102 is optional, and one or more of these steps may be omitted or replaced in different embodiments.

[0228] In some embodiments, reference may be made to other optional implementations described before or after the description corresponding to FIG. 4 .

[0229] Figure 5 is a flow chart of a method for determining historical CSI input based on CSI compression according to an embodiment of the present disclosure. As shown in Figure 5, an embodiment of the present disclosure relates to a method for determining historical CSI input based on CSI compression, which is executed by a terminal or a network device. The method includes:

[0230] Step S5101: Determine whether there is any historical CSI input of the AI ​​model at the historical moment before the K moment.

[0231] For example, this embodiment is applicable to space-frequency-time domain three-dimensional CSI compression based on AI. During the space-frequency-time domain CSI compression transmission process, an AI model Encoder (encoder) for encoding is configured in the terminal, and an AI model Decoder (decoder) for decoding is configured in the network device. For the data transmission of the Lth layer at time K, if the terminal's Encoder does not have a historical CSI input a before time K, t , and / or the decoder of the network device does not have historical CSI input b t , it is necessary to redefine the historical CSI input of the Encoder and Decoder models during the data transmission of the Lth layer at time K.

[0232] Step S5102: Obtain the CSI output of the AI ​​model as the historical CSI input of the AI ​​model.

[0233] In some embodiments, the historical CSI inputs of the encoder and decoder can be the historical time t closest to the current K time. r , the CSI of the Lth layer output by the Encoder, and the CSI of the Lth layer output by the Decoder.

[0234] In some embodiments, the historical moment t r The time corresponding to the period T can be set as K time minus N times, that is, t r =K-NT, where N is a positive integer and T is the reporting period of binary bit stream information during the space-frequency-time domain CSI compression process.

[0235] For example, in this embodiment, the input information of the encoder is the eigenvector of the downlink channel information H after SVD (Singular Value Decomposition). At time K, the UE estimates the downlink channel H based on the received pilot signal, and determines after channel measurement that two-layer data transmission is required, that is, the corresponding rank V=2. The UE performs CSI compression feedback on the two eigenvectors corresponding to the two singular values ​​obtained after SVD decomposition of the downlink channel information H. If the currently adopted model is a layer-common model, that is, all layers use the same AI model for data inference. When compression is performed according to the AI ​​model shown in Figure 1b, the CSI output at the previous moment needs to be used as the historical CSI input. When compressing the eigenvector corresponding to the first layer, there is a corresponding historical CSI output at the previous moment (such as KT moment), but there is no historical CSI input output during the second layer inference at KT moment. At this time, the UE needs to look forward, for example, to the historical CSI input output during the second layer inference at K-2T moment, and so on, step by step forward until t is obtained. r The historical CSI output during the second-layer reasoning at time K is used as the historical CSI input for the second-layer reasoning at the current time K. Among them, for t r There is no time limit and it can be any historical moment before time K when the AI ​​model outputs CSI.

[0236] In some embodiments, the space-frequency-time domain CSI bit information transmission at time K is determined to be a multi-layer data stream transmission, that is, when the data transmission Rank (level) is greater than 1, the historical CSI inputs of the encoder and decoder are the historical time t closest to the current time K. r, output by the Lith layer and Where i = 1, 2, 3, ..., the Lith layer is any data transmission layer before the Lth layer, that is, the historical CSI input of the Lth layer at time K in this embodiment is the AI ​​model t r The output of the Lith layer at the moment and For example, if the output transmission Rank is determined to be 4 and L is 4, then the CSI output by the third layer at time K can be t r Historical CSI output of layer 1, layer 2, or layer 3 at the moment and

[0237] For example, in this embodiment, the input information of the encoder is the eigenvector of the downlink channel information H after SVD (Singular Value Decomposition). At time K, the UE estimates the downlink channel H based on the received pilot signal, and determines after channel measurement that two-layer data transmission is required, that is, the corresponding rank V=2. The UE performs CSI compression feedback on the two eigenvectors corresponding to the two singular values ​​obtained after SVD decomposition of the downlink channel information H. If the currently adopted model is a layer-common model, that is, all layers use the same AI model for data inference. When compression is performed according to the AI ​​model shown in Figure 1b, the historical CSI output at the previous moment needs to be used. When compressing the eigenvector corresponding to the first layer, there is a corresponding historical CSI output at the previous moment (such as KT moment), but there is no historical CSI input output during the second layer inference at KT moment. The UE can use the output historical CSIa of the first layer output by the encoder at the previous KT moment. k-T , as the input history CSIa during the second-layer Encoder inference k-T .

[0238] Optionally, when the decoder on the network device side cannot directly obtain the input historical CSI of the second layer inference at the previous moment at time K, the output historical CSI of the first layer decoder at time KT can also be used as the historical CSI input of the second layer inference at the current time K. k-T .

[0239] In some embodiments, the historical CSI inputs of the encoder and decoder are respectively assigned as the CSI initialization value output by the encoder and the CSI initialization value output by the decoder.

[0240] For example, when the terminal's encoder does not obtain the CSI output by the Lth layer at the previous moment at time K, the encoder's CSI initialization value can be used as the historical CSI input for the Lth layer at the current moment K. Correspondingly, when the network device's decoder does not obtain the CSI output by the Lth layer at the previous moment at time K, the decoder's initialization value can be used as the historical CSI input for the Lth layer at the current moment K.

[0241] In some embodiments, a reporting period T for binary bit stream information during space-frequency-time domain CSI compression can be predefined. d , in the reporting period T d The UE reports the binary bit stream corresponding to the Lth layer at least once, and determines the historical reporting period T before time K. d , the historical CSI inputs of Encoder and Decoder are respectively a reporting period T before time K d The t closest to time K r At this moment, the output of the Lth layer and

[0242] For example, in a communication system composed of a terminal and a network device, a period T can be predefined by a communication protocol. d , so that the terminal is based on the period T d Report the binary bit stream. d The terminal reports the binary bit stream corresponding to the Lth layer at least once within a period of time. When the terminal does not obtain the historical CSI input information of the encoder at the Lth layer at the previous moment at time K, it determines the previous historical period T before time K. d The t closest to time K r At this moment, the output of the Lth layer As the historical CSI input of the terminal encoder at the Lth layer at time K. When the corresponding network device does not obtain the historical CSI input of the decoder at the previous moment at time K, it determines the previous historical period T before time K. d The t closest to time K r At this moment, the output of the Lth layer Serves as the historical CSI input of the Lth layer of the network device decoder at time K.

[0243] In some embodiments, before time K, the NW sends a signaling instruction to the UE to report t r The binary bit stream corresponding to the Lth layer at time t, the historical CSI inputs of the Encoder and Decoder are the t before time K and the closest to time K. rAt this moment, the output of the Lth layer and

[0244] Optionally, the signaling may be one or more of RRC information, MAC-CE information and DCI information.

[0245] For example, the network device can send a non-periodic CSI-RS information or CSI-RS burst (a burst of multiple CSI-RS resources) information to the terminal. The burst information includes multiple CSI-RS resource transmissions, which are used for the CSI-RS to obtain the historical CSI input information of the encoder when the L-th layer inference is performed.

[0246] For example, when the network device decompresses the binary bit stream information uploaded by the terminal based on the Decoder, it needs to input the historical CSI input information of the Lth layer at time K to the Decoder. When the historical CSI input information of the Lth layer at time K does not exist in the network device, the network device sends non-periodic CSI-RS information or CSI-RS burst information to the terminal. The CSI-RS information or CSI-RS burst information is used to instruct the terminal to re-measure and report the CSI, and determine the historical CSI input of the Lth layer encoder at time K according to the set protocol, input the determined historical CSI to the encoder, compress the channel information determined by the re-measurement, generate binary bits stream information, and report the binary bits stream information to the network device.

[0247] For example, if a terminal reports the binary bits stream compressed by the encoder at time K, and there are insufficient uplink transmission resources or a conflict with other CSI reports, the decoder's inference of the binary bits stream may result in inaccurate or unavailable CSI. In this case, the network device can send aperiodic CSI-RS information to the terminal, enabling it to remeasure and report the corresponding binary bits stream based on the CSI-RS information.

[0248] Optionally, if the terminal fails to report successfully after multiple attempts, causing the network device's decoder to be unable to parse the binary bits stream information, the network device can send burst information containing multiple CSI-RS resources to the terminal, instructing the terminal to re-measure and report CSI based on this information. This allows both the terminal's encoder and the network device's decoder to output the CSI at the historical moment corresponding to time K, which is used for subsequent encoder / decoder reasoning.

[0249] Through the above method, when the historical CSI input output at the previous moment is not available at moment K, a method for determining the historical CSI input at moment K is proposed. This enables the encoder and decoder models to obtain reliable historical CSI input, ensuring the reliability and consistency of the CSI compression feedback process of the communication system.

[0250] Figure 6 is a structural diagram of the terminal proposed in an embodiment of the present disclosure. As shown in Figure 6, the terminal 6100 may include: a first processing module 6101 and a second processing module 6102. In some embodiments, the first processing module 6101 is configured to determine that there is no first CSI input for the first AI model of the terminal, the first CSI input is the CSI input of the first data transmission layer of the first AI model at the first moment, and the processing module 6102 is configured to determine that the second CSI output by the first AI model is the first CSI input. Optionally, the first processing module 6101 and the second module 6102 are used to perform at least one of the communication steps such as determination and / or acquisition performed by the terminal 101 in any of the above methods, which will not be repeated here.

[0251] In some embodiments, the first processing module 6101 and the second processing module 6102 may include an execution module and an acquisition module, which may be separate or integrated. Optionally, the execution module and the executor may be interchangeable.

[0252] Figure 7 is a structural diagram of a network device proposed in an embodiment of the present disclosure. As shown in Figure 7, the network device 7100 may include: a third processing module 7101 and a fourth processing module 7102. In some embodiments, the third transceiver module 7101 is configured to determine that there is no second CSI input for the second AI model of the network device, and the second CSI input is the CSI input of the first data transmission layer of the second AI model at the first moment, and the fourth transceiver module 7102 is configured to determine that the third CSI output by the second AI model is the second CSI input. Optionally, the first processing module 7101 and the second module 7102 are used to perform at least one of the communication steps such as determination and / or acquisition performed by the terminal 101 in any of the above methods, which will not be repeated here.

[0253] In some embodiments, the first processing module 7101 and the second processing module 7102 may include an execution module and an acquisition module, which may be separate or integrated. Optionally, the execution module and the executor may be interchangeable.

[0254] Figure 8 is a schematic diagram of the structure of a communication device 8100 according to an embodiment of the present disclosure. Communication device 8100 can be a network device (e.g., an access network device, a core network device, etc.), a terminal (e.g., a user equipment, etc.), a chip, a chip system, or a processor that supports a network device to implement any of the above methods, or a chip, a chip system, or a processor that supports a terminal to implement any of the above methods. Communication device 8100 can be used to implement the methods described in the above method embodiments. For details, please refer to the description of the above method embodiments.

[0255] As shown in Figure 8, the communication device 8100 includes one or more third processors 8101. The third processor 8101 can be a general-purpose processor or a dedicated processor, for example, a baseband processor or a central processing unit. The baseband processor can be used to process the communication protocol and communication data, and the central processing unit can be used to control the communication device (such as a base station, a baseband chip, a terminal device, a terminal device chip, a DU or a CU, etc.), execute programs, and process program data. Optionally, the communication device 8100 is used to perform any of the above methods. Optionally, one or more third processors 8101 are used to call instructions to cause the communication device 8100 to perform any of the above methods.

[0256] In some embodiments, the communication device 8100 further includes one or more third transceivers 8102. When the communication device 8100 includes one or more third transceivers 8102, the third transceiver 8102 performs at least one of the communication steps, such as sending and / or receiving, in the above method, and the third processor 8101 performs at least one of the other steps. In an optional embodiment, the transceiver may include a receiver and / or a transmitter, and the receiver and transmitter may be separate or integrated. Optionally, the terms transceiver, transceiver unit, transceiver, transceiver circuit, interface circuit, and interface may be interchangeable, the terms transmitter, transmitting unit, transmitter, and transmitting circuit may be interchangeable, and the terms receiver, receiving unit, receiver, and receiving circuit may be interchangeable.

[0257] In some embodiments, the communication device 8100 further includes one or more third memories 8103 for storing data. Alternatively, all or part of the third memories 8103 may be located outside the communication device 8100. In an alternative embodiment, the communication device 8100 may include one or more first interface circuits 8104. Optionally, the first interface circuit 8104 is connected to the third memories 8103. The first interface circuit 8104 may be configured to receive data from the third memories 8103 or other devices, and to send data to the third processor 8101 or other devices. For example, the first interface circuit 8104 may read data stored in the third memories 8103 and send the data to the third processor 8101.

[0258] The communication device 8100 described in the above embodiment may be a network device or a terminal, but the scope of the communication device 8100 described in the present disclosure is not limited thereto, and the structure of the communication device 8100 may not be limited by FIG8 . The communication device may be an independent device or may be part of a larger device. For example, the communication device may be: 1) an independent integrated circuit IC, or a chip, or a chip system or subsystem; (2) a collection of one or more ICs, optionally, the above IC collection may also include a storage component for storing data and programs; (3) an ASIC, such as a modem; (4) a module that can be embedded in other devices; (5) a receiver, a terminal device, an intelligent terminal device, a cellular phone, a wireless device, a handheld device, a mobile unit, an in-vehicle device, a network device, a cloud device, an artificial intelligence device, etc.; (6) others, etc.

[0259] FIG9 is a schematic diagram of the structure of a chip 8200 according to an embodiment of the present disclosure. If the communication device 8100 can be a chip or a chip system, please refer to the schematic diagram of the structure of the chip 8200 shown in FIG9 , but the present invention is not limited thereto.

[0260] The chip 8200 includes one or more fourth processors 8201. The chip 8200 is configured to execute any one of the above methods.

[0261] In some embodiments, the chip 8200 further includes one or more second interface circuits 8202. The terms interface circuit, interface, and transceiver pins are optionally interchangeable. In some embodiments, the chip 8200 further includes one or more fourth memories 8203 for storing data. Optionally, all or part of the fourth memories 8203 may be located external to the chip 8200. Optionally, the second interface circuit 8202 is connected to the fourth memory 8203. The second interface circuit 8202 can be used to receive data from the fourth memory 8203 or other devices, or to send data to the fourth memory 8203 or other devices. For example, the second interface circuit 8202 can read data stored in the fourth memory 8203 and send the data to the fourth processor 8201.

[0262] In some embodiments, the second interface circuit 8202 performs at least one of the communication steps, such as sending and / or receiving, in the above-described method. For example, the second interface circuit 8202 performing the communication steps, such as sending and / or receiving, in the above-described method means that the second interface circuit 8202 performs data exchange between the fourth processor 8201, the chip 8200, the fourth memory 8203, or the transceiver device. In some embodiments, the fourth processor 8201 performs at least one of the other steps.

[0263] The modules and / or devices described in various embodiments, such as virtual devices, physical devices, and chips, can be arbitrarily combined or separated according to circumstances. Optionally, some or all steps can also be performed collaboratively by multiple modules and / or devices, which is not limited here.

[0264] The present disclosure also proposes a storage medium having instructions stored thereon, which, when executed on the communication device 8100, causes the communication device 8100 to execute any of the above methods. Optionally, the storage medium is an electronic storage medium. Optionally, the storage medium is a computer-readable storage medium, but is not limited thereto, and may also be a storage medium readable by other devices. Optionally, the storage medium may be a non-transitory storage medium, but is not limited thereto, and may also be a temporary storage medium.

[0265] The present disclosure also provides a program product, which, when executed by the communication device 8100, enables the communication device 8100 to perform any of the above methods. Optionally, the program product is a computer program product.

[0266] The present disclosure also proposes a computer program, which, when executed on a computer, causes the computer to perform any one of the above methods.

Claims

1. A method for determining CSI input of an AI model, characterized in that: Executed by a terminal, the method includes: Determining that a first channel state information (CSI) input does not exist for a first artificial intelligence (AI) model of the terminal, where the first CSI input is a CSI input of a first data transmission layer of the first AI model at a first moment; Determine the second CSI output by the first AI model as the first CSI input.

2. The method according to claim 1, characterized in that The second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

3. The method according to claim 2, characterized in that The second moment is the moment corresponding to the first moment minus N times the set period, where N is a positive integer.

4. The method according to claim 1, wherein The CSI transmission at the first moment is multi-layer data transmission, the second CSI is the CSI output by the first AI model at the second data transmission layer at the second moment, the second moment is the last output moment before the first moment, the output moment is the moment when the first AI model outputs the CSI, and the index of the second data transmission layer is less than the index of the first data transmission layer.

5. The method according to claim 1, wherein The CSI transmission at the first moment is multi-layer data transmission, and the second CSI is the CSI output by the first AI model at the second data transmission layer at the first moment, and the index of the second data transmission layer is smaller than the index of the first data transmission layer.

6. The method according to claim 1, characterized in that The second CSI includes initial value information of the first AI model.

7. The method according to claim 1, characterized in that The second CSI is the CSI output by the first AI model at a third moment, where the third moment is the last output moment within a set period and before the first moment, the output moment is the moment when the first AI model outputs the CSI, and the set period is a period for the terminal to report CSI.

8. The method according to claim 1, characterized in that The method further comprises: receiving first indication information sent by a network device through signaling, where the first indication information is used to instruct the terminal to report CSI based on the first indication information; Determine the second CSI according to the first indication information, where the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

9. The method according to claim 8, characterized in that The signaling includes at least one of the following: a radio resource control RRC message, a medium access control layer-control element MAC-CE information and downlink control information DCI.

10. The method according to claim 1, characterized in that The method further comprises: receiving second indication information sent by the network device; Re-report CSI according to the second indication information.

11. The method according to claim 10, characterized in that The second indication information includes one or more non-periodic channel state reference signal CSI-RS resource information, and there is a time interval between the transmission order of the multiple CSI-RS resource information.

12. The method according to claim 1, characterized in that The generation time of the second CSI is the same as the generation time of the third CSI, and the third CSI is the CSI input of the network device corresponding to the second AI model.

13. The method according to claim 12, characterized in that The second CSI is a first CSI initial value output by the first AI model, and the third CSI is a second CSI initial value output by the second AI model.

14. A method for determining CSI input of an AI model, characterized in that: Executed by a network device, the method includes: Determining that a second AI model of the network device does not have a second CSI input, where the second CSI input is a CSI input of the second AI model at the first time to the first data transmission layer; Determine the third CSI output by the second AI model as the second CSI input.

15. The method according to claim 14, characterized in that The third CSI is the CSI output by the second AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the second AI model outputs the CSI.

16. The method according to claim 15, characterized in that The second moment is the moment corresponding to the first moment minus N times the set period, where N is a positive integer.

17. The method according to claim 14, characterized in that The CSI transmission at the first moment is multi-layer data transmission, the third CSI is the CSI output by the second AI model in the second data transmission layer at the second moment, the second moment is the last output moment before the first moment, the output moment is the moment when the second AI model outputs the CSI, and the index of the second data transmission layer is less than the index of the first data transmission layer.

18. The method according to claim 14, characterized in that The CSI transmission at the first moment is multi-layer data transmission, and the third CSI is the CSI output by the second AI model at the second data transmission layer at the first moment, and the index of the second data transmission layer is smaller than the index of the first data transmission layer.

19. The method according to claim 14, wherein The third CSI includes initial value information of the second AI model.

20. The method according to claim 14, wherein The third CSI is CSI output by the second AI model at a third moment. The third moment is the last output moment within a set period and before the first moment. The output moment is the moment when the second AI model outputs CSI. The set period is a period for the terminal to report CSI.

21. The method according to claim 14, wherein The method further comprises: First indication information is sent to the terminal via signaling, where the first indication information is used to instruct the terminal to determine second CSI of a first AI model of the terminal based on the first indication information, where the second CSI is the CSI output by the first AI model at a second moment, where the second moment is the last output moment before the first moment, and the output moment is the moment when the first AI model outputs the CSI.

22. The method according to claim 21, characterized in that The signaling includes at least one of the following: a radio resource control RRC message, a medium access control layer-control element MAC-CE information and downlink control information DCI.

23. The method according to claim 14, wherein The method further comprises: Send second indication information to the terminal, where the second indication information is used to instruct the terminal to determine a first CSI input of the first AI model based on the second indication information, where the first CSI input is the CSI input of the first AI model at the first moment and the first data transmission layer.

24. The method according to claim 23, wherein The second indication information includes one or more non-periodic channel state reference signal CSI-RS resource information, and there is a time interval between the transmission order of the multiple CSI-RS resource information.

25. The method according to any one of claims 14 to 24, characterized in that The generation time of the third CSI is the same as the generation time of the second CSI, and the second CSI is the CSI input of the terminal corresponding to the first AI model.

26. The method according to claim 25, characterized in that The second CSI is a first CSI initial value output by the first AI model, and the third CSI is a second CSI initial value output by the second AI model.

27. A terminal, characterized in that: include: A first processing module is configured to determine that a first AI model of the terminal does not have a first CSI input, where the first CSI input is a CSI input of the first AI model at a first data transmission layer at a first moment; The second processing module is configured to determine the second CSI output by the first AI model as the first CSI input.

28. A network device, characterized in that: include: a third processing module configured to determine that the second AI model of the network device does not have a second CSI input, where the second CSI input is a CSI input of the second AI model at the first time to the first data transmission layer; The fourth processing module is configured to determine the third CSI output by the second AI model as the second CSI input.

29. A terminal, characterized in that: include: one or more processors; The terminal is used to execute the method for determining the AI model CSI input according to any one of claims 1 to 13.

30. A network device, characterized in that: include: one or more processors; The network device is used to execute the method for determining the AI model CSI input according to any one of claims 14-26.

31. A communication system, characterized in that: The method comprises a terminal and a network device, wherein the terminal is configured to implement the method for determining the AI model CSI input according to any one of claims 1 to 13, and the network device is configured to implement the method for determining the AI model CSI input according to any one of claims 14 to 25.

32. A storage medium storing instructions, characterized in that: When the instruction is executed on the communication device, the communication device executes the method for determining the AI model CSI input according to any one of claims 1 to 13, or the communication device executes the method for determining the AI model CSI input according to any one of claims 14 to 26.

33. A computer program product comprising a computer program and / or instructions, characterized in that When the computer program and / or instructions are executed by a communication device, the method for determining the AI model CSI input as described in any one of claims 1-13 is implemented, or when the computer program and / or instructions are executed by a communication device, the method for determining the AI model CSI input as described in any one of claims 14-26 is implemented.

Images