Data collection method based on joint source-channel coding, and terminal and network-side device

Abstract

The present application belongs to the technical field of communications. Disclosed are a data collection method based on joint source-channel coding, and a terminal and a network-side device. The data collection method based on joint source-channel coding in the embodiments of the present application comprises: a first device inputting first information into a first artificial intelligence (AI) unit, so as to obtain second information; and transmitting the second information to a second device in a preset manner, wherein the first information comprises channel state information (CSI), the first AI unit is used for coding the first information, and the preset manner is used for adding a channel influence.

Description

Data acquisition method, terminal and network side device of joint source channel coding

[0001] Cross-reference to Related Applications

[0002] The present application claims priority to the Chinese patent application No. 202411250937.4, filed on September 6, 2024, and entitled "Data acquisition method, terminal and network side device of joint source channel coding", the whole content of which is incorporated herein by reference. TECHNICAL FIELD

[0003] The present application belongs to the field of communication technology, and specifically relates to a data acquisition method, terminal and network side device of joint source channel coding. BACKGROUND

[0004] Channel State Information (CSI) feedback enhancement is a kind of bilateral model, and the training methods include three kinds: unilateral training, joint training and separate training. The unilateral training is performed in one entity, and the joint training and the separate training belong to bilateral training, which needs information transmission between the network and the User Equipment (UE).

[0005] At present, the CSI can be source coded by using artificial intelligence (AI) to realize the CSI feedback enhancement. The AI-based CSI source coding adopts the separate source channel coding transmission mode in the actual transmission process, that is, after the source coding is performed by using the AI, the corresponding redundancy is increased by using the channel coding to resist the influence of the wireless channel. The data acquired in this way has high accuracy, which is not conducive to the training of the CSI feedback enhancement. How to acquire the data that is helpful to the training of the CSI feedback enhancement has not been solved. SUMMARY

[0006] The embodiments of the present application provide a data acquisition method, terminal and network side device of joint source channel coding, which can solve the problem of data acquisition for the training of the CSI feedback enhancement.

[0007] In a first aspect, a data acquisition method of joint source channel coding is provided, comprising:

[0008] The first device inputs the first information into a first artificial intelligence (AI) unit to obtain second information;

[0009] The second information is transmitted to a second device by using a preset mode;

[0010] The first information includes channel state information (CSI), the first AI unit is configured to encode the first information, and the preset manner is configured to add channel influence.

[0011] In a second aspect, a data acquisition method for joint source channel coding is provided, including:

[0012] The second device receives second information transmitted by the first device using a preset manner.

[0013] The preset manner is configured to add channel influence, the second information is obtained by the first device inputting first information into a first artificial intelligence (AI) unit, the first information includes channel state information (CSI), and the first AI unit is configured to encode the first information.

[0014] In a third aspect, a data acquisition apparatus for joint source channel coding is provided, including:

[0015] A processing module is configured to obtain second information by inputting first information into a first artificial intelligence (AI) unit.

[0016] A transmission module is configured to transmit the second information to a second device using a preset manner.

[0017] The first information includes channel state information (CSI), the first AI unit is configured to encode the first information, and the preset manner is configured to add channel influence.

[0018] In a fourth aspect, a data acquisition apparatus for joint source channel coding is provided, including:

[0019] A receiving module is configured to receive second information transmitted by a first device using a preset manner.

[0020] The preset manner is configured to add channel influence, the second information is obtained by the first device inputting first information into a first artificial intelligence (AI) unit, the first information includes channel state information (CSI), and the first AI unit is configured to encode the first information.

[0021] In a fifth aspect, a data acquisition apparatus for joint source channel coding is provided, and the apparatus is configured to perform steps of the method according to the first aspect or the second aspect.

[0022] In a sixth aspect, a terminal is provided, including a processor and a memory, the memory stores programs or instructions executable on the processor, and the programs or instructions are executed by the processor to implement steps of the method according to the first aspect or the second aspect.

[0023] In a seventh aspect, a terminal is provided, comprising a processor and a communication interface, wherein the processor is configured to implement the steps of the method according to the first aspect or the second aspect, and the communication interface is configured to be coupled to the processor.

[0024] In an eighth aspect, a network-side device is provided, comprising a processor and a memory, wherein the memory stores programs or instructions executable by the processor, and the programs or instructions, when executed by the processor, implement the steps of the method according to the first aspect or the second aspect.

[0025] In a ninth aspect, a network-side device is provided, comprising a processor and a communication interface, wherein the processor is configured to implement the steps of the method according to the first aspect or the second aspect, and the communication interface is configured to be coupled to the processor.

[0026] In a tenth aspect, a readable storage medium is provided, wherein the readable storage medium stores programs or instructions, and the programs or instructions, when executed by a processor, implement the steps of the method according to the first aspect or the second aspect.

[0027] In an eleventh aspect, a wireless communication system is provided, comprising a first device and a second device, wherein the first device is configured to implement the steps of the method according to the first aspect, and the second device is configured to implement the steps of the method according to the second aspect.

[0028] In a twelfth aspect, a chip is provided, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is configured to run programs or instructions to implement the method according to the first aspect or the second aspect.

[0029] In a thirteenth aspect, a computer program / program product is provided, wherein the computer program / program product is stored in a storage medium, and the computer program / program product is executed by at least one processor to implement the steps of the method according to the first aspect or the second aspect.

[0030] In the embodiments of the present application, the first device inputs first information into the first AI unit to obtain second information, and transmits the second information to the second device in a preset manner, wherein the first information comprises CSI, the first AI unit is configured to encode the first information, and the preset manner is configured to add channel influence, so that the second device can obtain the second information with added channel influence and consider the influence of the channel in the AI-based encoding process, thereby realizing data acquisition based on AI-based joint source channel coding and facilitating CSI feedback enhanced training. BRIEF DESCRIPTION OF DRAWINGS

[0031] FIG. 1 shows a block diagram of a wireless communication system to which the embodiments of the present application can be applied;

[0032] FIG. 2 shows a flow diagram of a data acquisition method of joint source-channel coding according to an embodiment of the present application;

[0033] FIG. 3 shows another flow diagram of a data acquisition method of joint source-channel coding according to an embodiment of the present application;

[0034] FIG. 4 shows still another flow diagram of a data acquisition method of joint source-channel coding according to an embodiment of the present application;

[0035] FIG. 5 shows still another flow diagram of a data acquisition method of joint source-channel coding according to an embodiment of the present application;

[0036] FIG. 6 shows a structure diagram of a data acquisition apparatus of joint source-channel coding according to an embodiment of the present application;

[0037] FIG. 7 shows another structure diagram of a data acquisition apparatus of joint source-channel coding according to an embodiment of the present application;

[0038] FIG. 8 shows a structure diagram of a communication device according to an embodiment of the present application;

[0039] FIG. 9 shows a hardware structure diagram of a terminal according to an embodiment of the present application;

[0040] FIG. 10 shows a hardware structure diagram of a network-side device according to an embodiment of the present application. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be clearly described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those of ordinary skill in the art belong to the scope of the present application.

[0042] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, "or" in this application indicates at least one of the connected objects. For example, the scope of protection for "A or B" covers at least three scenarios: Scenario 1: including A but not B; Scenario 2: including B but not A; Scenario 3: including both A and B. In addition, the terms "A and / or B," "at least one of A and B," and "at least one of A or B" also cover at least the above three scenarios. The character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0043] The term "instruction" in this application can be either a direct instruction (or explicit instruction) or an indirect instruction (or implicit instruction). A direct instruction can be understood as one in which the sender explicitly informs the receiver of specific information, the operation to be performed, or the requested result, etc., in the instruction sent. An indirect instruction can be understood as one in which the receiver determines the corresponding information based on the instruction sent by the sender, or makes a judgment and determines the operation to be performed or the requested result, etc., based on the judgment result.

[0044] It is worth noting that the technology described in the embodiments of the present application is not limited to Long Term Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA) or other systems. The terms "system" and "network" in the embodiments of the present application are often used interchangeably, and the described technology can be used in the above-mentioned systems and radio technologies, as well as in other systems and radio technologies. The following description describes a New Radio (NR) system for example purposes, and NR terminology is used in most of the following description, but these technologies can also be applied to systems other than NR systems, such as 6th Generation (6G) communication systems. th

[0045] ​FIG. 1 shows a block diagram of a wireless communication system to which embodiments of the present application can be applied. The wireless communication system includes a terminal 11 and a network-side device 12. The terminal 11 can be a terminal-side device such as a mobile phone, a Tablet Personal Computer, a Laptop Computer, a notebook computer, a Personal Digital Assistant (PDA), a palmtop computer, a netbook, an Ultra-mobile Personal Computer (UMPC), a Mobile Internet Device (MID), an Augmented Reality (AR) device, a Virtual Reality (VR) device, a robot, a wearable device, a flight vehicle, a Vehicle User Equipment (VUE), a shipboard device, a Pedestrian User Equipment (PUE), a smart home (a home device with a wireless communication function such as a refrigerator, a television, a washing machine, or furniture), a game console, a Personal Computer (PC), a kiosk, or a self-service machine. The wearable device includes a smart watch, a smart bracelet, a smart earphone, smart glasses, smart jewelry (a smart bracelet, a smart necklace, a smart ring, a smart necklace, a smart anklet, a smart necklace, etc.), a smart wristband, smart clothes, etc. The vehicle-mounted device can also be referred to as a vehicle-mounted terminal, a vehicle-mounted controller, a vehicle-mounted module, a vehicle-mounted component, a vehicle-mounted chip, or a vehicle-mounted unit, etc. It should be noted that the specific type of the terminal 11 is not limited in the embodiments of the present application. The network-side device 12 can include an access network device or a core network device. The access network device can also be referred to as a Radio Access Network (RAN) device, a radio access network function, or a radio access network unit. The access network device can include a base station, a Wireless Local Area Network (WLAN) Access Point (AP), or a Wireless Fidelity (WiFi) node, etc.The base station can be referred to as a Node B (NB), an evolved Node B (eNB), a next generation Node B (gNB), a New Radio Node B (NR Node B), an access point, a relay station (RBS), a serving base station (SBS), a base transceiver station (BTS), a radio base station, a radio transceiver, a basic service set (BSS), an extended service set (ESS), a home Node B (HNB), a home evolved Node B, a transmit / receive point (TRP), or some other suitable terminology in the art, and is not limited to a particular technical terminology, provided that the same technical effect is achieved. It should be noted that in the embodiments of the present application, only the base station in the NR system is taken as an example for introduction, and the specific type of the base station is not limited.

[0046] The core network device can also be referred to as a core network node, a core network function, or a core network network element, etc., which includes but is not limited to at least one of the following: a mobility management entity (MME), an access and mobility management function (AMF), a session management function (SMF), a user plane function (UPF), a policy control function (PCF), a policy and charging rules function (PCRF), an edge application server discovery function (EASDF), a unified data management (UDM), a unified data repository (UDR), a home subscriber server (HSS), a centralized network configuration (CNC), a network repository function (NRF), a network exposure function (NEF), a local NEF (L-NEF), a binding support function (BSF), an application function (AF), a location management function (LMF), a gateway mobile location center (GMLC), a network data analytics function (NWDAF), etc. It should be noted that only the core network device in the NR system is taken as an example for introduction in the embodiments of the present application, and the specific type of the core network device is not limited. If the name of the core network device mentioned in the embodiments of the present application changes in the subsequent protocol version (for example, 6G), it is also within the protection scope of the present application.

[0047] Optionally, the core network device can be implemented by one or more function modules in one device, or can be implemented by multiple devices together, and the embodiments of the present application do not make a specific limitation hereon. It can be understood that the above function modules can be network elements in a hardware device, can be software function modules running on a special hardware, or can be virtualized function modules instantiated on a platform (for example, a cloud platform).

[0048] The joint source channel coding data acquisition method, terminal and network side device provided by the embodiments of the present application will be described in detail below in combination with the drawings, some embodiments and application scenarios.

[0049] There are three training cooperation modes in the CSI feedback enhancement in 3GPP Rel-18. The first mode (Type 1): training in one entity, similar to the general AI algorithm training, after training, the trained two-end model is respectively deployed to the UE side and the network (NW) side through model transmission. The second mode (Type 2) and the third mode (Type 3) both need to transmit the encoded information and gradient information from the NW to the UE. However, since the CSI feedback enhancement in 3GPP Rel-18 uses AI for source coding, when designing Type 2 and Type 3, the channel interference is not considered. Based on this, the embodiments of the present application encode the CSI source based on AI and add channel influence for transmission, realize the joint source channel coding based on AI, and solve the problem that the channel interference is not considered in the AI-based coding process in the Type 2 and Type 3 scenarios.

[0050] FIG. 2 shows a flowchart of a joint source channel coding data acquisition method provided by the embodiments of the present application, and the method 200 can be executed by a first device. As shown in FIG. 2, the method can include the following steps.

[0051] S202: The first device inputs first information into a first AI unit to obtain second information.

[0052] The first device can be a terminal or a network side device, and the specific limitation is not limited. The first AI unit can be an encoder or a decoder, and the specific limitation is not limited.

[0053] In the embodiments of the present application, the first information can include CSI, the first AI unit is used for encoding the first information, and the obtained second information includes the encoded CSI.

[0054] In an implementation, in addition to the CSI, the first information can also include image information or voice information, and the specific limitation is not limited.

[0055] In the embodiments of the present application, the CSI can include at least one of the following parameters:

[0056] 1) Channel Quality Indicator (CQI);

[0057] 2) Precoding Matrix Indicator (PMI);

[0058] 3) CSI-Reference Signal Resource Indicator (CRI);

[0059] 4) SSB resource indicator (SS / PBCH Block Resource Indicator, SSBRI);

[0060] 5) Layer Indicator (LI);

[0061] 6) Rank Indicator (RI);

[0062] 7) Layer 1 Reference Signal Received Power (L1-RSRP);

[0063] 8) Layer 1 Signal to Interference plus Noise Ratio (L1-SINR);

[0064] 9) Capability Index;

[0065] 10) Time-Difference Carrier-Phase (TDCP);

[0066] 11) Raw channel information;

[0067] 12) Transform domain channel information; for example, angle-time delay domain channel information;

[0068] 13) Processed raw channel information; for example, transforming spatial-frequency domain channel information to angle-time delay domain channel and / or performing truncation operation on the channel, etc.

[0069] 14) Source corresponding to the PMI, i.e., V matrix after Singular Value Decomposition (SVD) decomposition of the channel.

[0070] In an implementation, the CSI includes any of the above parameters, which can be replaced by at least one of a mean, a variance, a bias, and a product factor of the parameter, without limitation.

[0071] In an embodiment, the second information can be carried in one of the following:

[0072] 1) Downlink Control Information (DCI);

[0073] 2) Medium Access Control Control Element (MAC CE);

[0074] 3) Radio Resource Control (RRC);

[0075] 4) Physical Uplink Shared Channel (PUSCH);

[0076] 5) Physical Downlink Shared Channel (PDSCH);

[0077] 6) Physical Downlink Control Channel (PDCCH);

[0078] 7) Physical Uplink Control Channel (PUCCH).

[0079] In an embodiment, the second information can be configured by RRC or MAC CE to be periodically transmitted, semi-persistently transmitted, or non-periodically transmitted.

[0080] In the periodic transmission mode, the period of transmission configured by RRC or MAC CE is configured, and periodic transmission of the second information is performed according to the period after configuration.

[0081] In the semi-persistent transmission mode, the second information is periodically transmitted after being activated by MAC CE or DCI, until being deactivated or until a specified period.

[0082] In the non-periodic transmission mode, the second information is transmitted a specified number of times after being activated by MAC CE or DCI. The specified number of times can be configured by the network, such as once or multiple times, without limitation.

[0083] S204: transmitting the second information to the second device in a preset manner.

[0084] The preset manner can include a through-the-channel and lossless transmission manner, a lossless transmission manner, and a lossy transmission manner. The channel influence includes at least signal changes caused by channel transmission, such as at least one of channel interference, channel fading, channel noise, and the like.

[0085] In the embodiment of the present application, the above step S204 can be implemented in various transmission manners, including but not limited to one of the following transmission manners:

[0086] The first transmission manner: after the second information passes through the channel corresponding to the preset channel information, the second information is sent to the second device in a lossless transmission manner. The lossless transmission means that the transmission is protected by channel coding and / or checking, so that the wrong information can be retransmitted, and the transmission is ensured to be error-free. The protection of channel coding can be, for example, increasing the redundancy of channel coding to reduce the code rate of channel coding; the protection of checking manner can be, for example, adding cyclic redundancy check (CRC) and the like. This scenario belongs to the manner of adding channel influence at the sending end, and the first device adds the corresponding channel influence to the second information based on the preset channel information before transmission. The advantage of this manner is that when the actual channel used for transmission and the channel implementing joint source-channel coding do not match, data collection can be performed through the preset channel.

[0087] In addition, the second information can pass through the channel multiple times, and the information after each passing through the channel is sent to the second device in a lossless transmission manner. In this case, the second device can obtain multiple second information with added channel influence, and the effect of data enhancement can be achieved.

[0088] For example, the second information is divided into 5 times to perform the operation: passing through the channel and sending to the second device in a lossless transmission manner. Then, the second device can receive 5 second information with added channel influence, so as to improve the sample size of data collection.

[0089] The second transmission manner: the second information is sent to the second device in a lossless transmission manner. The lossless transmission means that the transmission is protected by channel coding and / or checking, so that the wrong information can be retransmitted, and the transmission is ensured to be error-free. The protection of channel coding can be, for example, increasing the redundancy of channel coding to reduce the code rate of channel coding; the protection of checking manner can be, for example, adding CRC and the like.

[0090] In this scenario, after receiving the second information, the second device can obtain the encoded information added with the influence of the channel through the preset channel information corresponding to the channel, thereby completing data collection. This method is that the second device, i.e., the receiving end, obtains the second information through the corresponding channel, which belongs to a method of adding the influence of the channel at the receiving end. The advantage of this method is that when the actual channel used for transmission and the channel used for implementing joint source-channel coding do not match, data collection can be performed through the preset channel. Further, the second device can also perform training of the second AI unit and the like after completing data collection, which is not limited in particular.

[0091] The third transmission method is to transmit the second information to the second device using a lossy transmission method. The lossy transmission means that the transmission is not protected by channel coding and / or checking and the like, and the transmission process can have errors. This lossy transmission method reduces the bandwidth used for data collection, and at the same time, adds the influence of the channel using the characteristics of the actual channel used for transmission.

[0092] In the embodiments of the present application, the lossless transmission method described above can include at least one of the following:

[0093] 1) Source coding. Correspondingly, the second device performs source decoding.

[0094] 2) Channel coding. Correspondingly, the second device performs channel decoding.

[0095] 3) Modulation. Correspondingly, the second device performs demodulation.

[0096] 4) Layer mapping. Correspondingly, the second device performs layer mapping.

[0097] 5) Interleaving. Correspondingly, the second device performs deinterleaving.

[0098] 6) Padding. Correspondingly, the second device performs depadding.

[0099] 7) Puncturing.

[0100] 8) Rate matching.

[0101] 9) Scrambling. Correspondingly, the second device performs descrambling.

[0102] 10) Precoding. Correspondingly, the second device performs receive combining.

[0103] The lossless transmission method described above can ensure error-free transmission through channel coding, thereby improving the coding quality.

[0104] In the embodiments of the present application, the lossy transmission method described above can include at least one of the following:

[0105] 1) Layer mapping. Wherein the second device correspondingly performs de-mapping.

[0106] 2) Interleaving. Wherein the second device correspondingly performs de-interleaving.

[0107] 3) Padding. Wherein the second device correspondingly performs de-padding.

[0108] 4) Puncturing.

[0109] 5) Rate matching.

[0110] 6) Scrambling. Wherein the second device correspondingly performs de-scrambling.

[0111] 7) Precoding. Wherein the second device correspondingly performs receive combining.

[0112] The above method using lossy transmission can add channel influence in training by using the actual channel environment used for transmission in the scenario where the channel characteristics are the same in the training phase and the deployment phase, reduce the occupation of resources in the training process, and avoid the problem of being unable to perform back propagation.

[0113] In the embodiments of the present application, the first device can be a terminal or a network side device, and the first information can be obtained in at least one of the following ways:

[0114] The first obtaining method is to obtain the first information predefined by a protocol.

[0115] The second obtaining method is to collect the first information. In this case, the first device collects the first information itself, so it does not need to obtain the first information from other devices. For example, the UE collects and obtains the CSI information.

[0116] The third obtaining method is to receive the first information sent by a second device. The second device can be a UE or a NW, which is not limited. For example, the second device collects and obtains the first information and then sends it to the first device, so that the first device can obtain the first information. For example, the second device NW collects the CSI information and sends it to the first device UE, so that the UE obtains the CSI information.

[0117] The fourth obtaining method is to receive the first information sent by a server of the first device. The server of the first device refers to a server that provides services for the first device, and both belong to the same network operator. For example, the first device and the server of the first device are both devices of network operator A.

[0118] The fifth obtaining method is to receive the first information sent by a third party server. The third party server refers to a server of another network operator. For example, the first device is a device of network operator A, and the third party server is a device of network operator B.

[0119] In the second obtaining manner, since the first information is collected by the first device, in order for the second device to train the second AI unit using the label, the first device can further send the sample information collected each time to the second device, and the second device trains the second AI unit using the sample information as the label after receiving the sample information. It is worth mentioning that the first device can send the second information and the sample information together to the second device, or the first device can also send the second information and the sample information separately to the second device, and the specific implementation is not limited.

[0120] In addition to the second obtaining manner, the first device does not need to send the sample information to the second device, and the second device can collect the sample information by itself or obtain the sample information from other devices, and the specific implementation is not limited. It is worth mentioning that in this case, the first device and the second device pre-agree the use mode of the first information to avoid the problem of mispairing.

[0121] In the embodiment of the application, the second information can include at least one of a bit sequence and a symbol sequence. The bit sequence corresponds to bit-level joint source channel coding, and the symbol sequence corresponds to symbol-level joint source channel coding.

[0122] For example, the second information is encoded CSI information including a bit sequence, or is encoded CSI information including a symbol sequence.

[0123] The above method provided by the embodiment can be used to input the first information into the first AI unit to obtain the second information, and transmit the second information to the second device in a preset manner, wherein the first information includes CSI, the first AI unit is used for encoding the first information, and the preset manner is used for adding channel influence, so that the second device can obtain the second information with added channel influence and can consider the influence of the channel in the AI-based encoding process, thereby realizing data collection based on AI-based joint source channel coding, helping to enhance the training of CSI feedback, improving the practicality of the training model, and making the model more widely applicable.

[0124] FIG. 3 shows a flowchart of a joint source channel coding data collection method provided by an embodiment of the application. The method 300 can be performed by a first device. As shown in FIG. 3, the method can include the following steps.

[0125] S302: The first device inputs the first information into the first AI unit to obtain the second information.

[0126] The first device can be a terminal or a network side device, and the specific implementation is not limited. The first information includes CSI, the first AI unit is used for encoding the first information, and the second information includes encoded CSI.

[0127] In the embodiments of the present application, the first device obtaining the second information can be triggered by one of the following:

[0128] 1) network indication;

[0129] 2) terminal reporting;

[0130] 3) predefined event.

[0131] The predefined event triggering the obtaining of the second information can include at least one of the following:

[0132] 1) in a wifi state;

[0133] 2) battery level exceeding a second threshold;

[0134] 3) in a charging state;

[0135] 4) supporting a preset traffic mode; for example, the sending of the second information does not count as traffic overhead;

[0136] 5) in a preset traffic mode;

[0137] 6) no service demand; for example, there is no need to train the first AI unit or the second AI unit;

[0138] 7) performance index below a third threshold.

[0139] S304: transmitting the second information to the second device in a preset manner.

[0140] The preset manner can include a lossless transmission manner, a lossy transmission manner, and a channel-based lossless transmission manner.

[0141] In the embodiments of the present application, the method can further include:

[0142] The first device transmits first related information to the second device in a lossy transmission manner or a lossless transmission manner. The first related information is information associated with the first information and the encoding and decoding. The first related information can also be used for the second device to train the second AI unit.

[0143] It is worth mentioning that the first device can transmit the second information and the first related information to the second device respectively, or the first device can transmit the second information and the first related information to the second device together, and the embodiments of the present application do not make specific limitations.

[0144] In the embodiments of the present application, the first related information can include at least one of the following:

[0145] 1) a sample in a CSI data set;

[0146] 2) a sample ID in a CSI data set;

[0147] 3) encoded dimension information; for example, the encoded dimension information is (16, 32, 32, 16) or the like;

[0148] 4) channel impact information;

[0149] 5) a rule for obtaining the first information.

[0150] In an implementation, the rule for obtaining the first information can be determined between the first device and the second device through a protocol, RRC, DCI, or uplink control information (UCI), or the like.

[0151] In the embodiments of the application, the rule for obtaining the first information can include an iteration number index, which is used to indicate that a preset number of samples are obtained from a current position in the samples arranged in a preset order, the current position being an initial position or a position where the last sample is obtained. The iteration number refers to the number of iterations of training of the first AI unit or the second AI unit. In the first iteration of training, the current position is the initial position, and in each subsequent iteration of training, the current position is the position where the last sample is obtained. In addition, the first device can also send the number of training iterations to the second device.

[0152] For example, the data set of the first information includes 10,000 samples arranged in a preset order, and it is agreed that 100 samples are obtained each time according to the arrangement order, so the rule for obtaining the first information in the first related information can include only the iteration number index. When the iteration number index is 1, the second device takes the first 100 samples from the data set; when the iteration number index is 2, the second device takes the 101st-200th samples from the data set; and so on, until the training is completed.

[0153] In the embodiments of the application, the obtaining of the first related information can be triggered by one of the following:

[0154] 1) network indication;

[0155] 2) terminal reporting;

[0156] 3) predefined event.

[0157] The predefined event for triggering the obtaining of the first related information can include at least one of the following:

[0158] 1) in a wifi state;

[0159] 2) power exceeding a second threshold;

[0160] 3) in a charging state;

[0161] 4) Support preset traffic mode; for example, the sending of the second information does not count as traffic overhead;

[0162] 5) In preset traffic mode;

[0163] 6) No service requirement; for example, there is no need to train the first AI unit or the second AI unit;

[0164] 7) Performance index is lower than the third threshold.

[0165] In the embodiments of the present application, the correspondence between the first related information and the second information can also be established. The correspondence between the first related information and the second information can have various forms, including but not limited to one of the following:

[0166] 1) Correspondence between the first information and the encoded first information. For example, (CSI_1, encoded CSI_1), (CSI_2, encoded CSI_2), etc.

[0167] 2) Correspondence between the first information and the encoded first information after lossy transmission. For example, the first device uses the third transmission mode described above, i.e., uses a lossy transmission mode to send the second information to the second device. Where the channel corresponding to the lossy transmission is A, the correspondence can be (CSI_1, encoded CSI_1 after channel A), (CSI_2, encoded CSI_2 after channel A), etc.

[0168] 3) Correspondence between the first information and the encoded first information through the channel corresponding to the preset channel information. For example, the first device uses the first transmission mode described above, i.e., sends the second information through the channel corresponding to the preset channel information, and then uses lossless transmission to send it to the second device. Where the channel corresponding to the preset channel information is B, the correspondence can be (CSI_1, encoded CSI_1 after channel B), (CSI_2, encoded CSI_2 after channel B), etc.

[0169] 4) Correspondence between the first information and the encoded first information and the added channel impact. For example, the channel impact is specifically channel interference, and the correspondence can be (CSI_2, encoded CSI_2, channel interference 2), etc.

[0170] It should be noted that in any of the above four corresponding relationships, the first information such as CSI_1 can be replaced by the ID of the first information such as CSI_ID_1, and correspondingly, (CSI_1, encoded CSI_1) can be replaced by (CSI_ID_1, encoded CSI_1), and (CSI_1, encoded CSI_1, channel interference 1) can be replaced by (CSI_ID_1, encoded CSI_1, channel interference 1), which will not be described in detail here.

[0171] In this way of using the ID of the sample to replace the sample for transmission, the sample does not need to be sent to the second device, and the transmission overhead can be reduced.

[0172] S306: Receive gradient information of the second information sent by the second device.

[0173] The gradient information of the second information is obtained by the second device according to training of the second AI unit on the second information.

[0174] In the embodiments of the present application, the second information and the gradient information of the second information are information propagated forward and information propagated backward between the first device and the second device in the training.

[0175] In the joint training scenario, the first device can be a terminal and the second device can be a network side device, or the first device and the second device can both be terminals. In the separate training scenario, the first device can be a terminal and the second device can be a network side device, or the first device and the second device can both be terminals, or the first device can be a network side device and the second device can be a terminal.

[0176] In the embodiments of the present application, the second AI unit can be a decoder or can also be an encoder, which is not limited in particular. For example, in the joint training or separate training scenario, the second AI unit can be a decoder. In the separate training scenario, the second AI unit can be an encoder.

[0177] In the embodiments of the present application, the gradient information of the second information can be carried in one of the following:

[0178] 1) DCI;

[0179] 2) MAC CE;

[0180] 3) RRC;

[0181] 4) PUSCH;

[0182] 5) PDSCH;

[0183] 6) PDCCH;

[0184] 7) PUCCH.

[0185] In the embodiments of the present application, the gradient information of the second information can be configured by RRC or MAC CE to be periodically transmitted, semi-persistently transmitted or non-periodically transmitted.

[0186] In the periodically transmitted mode, the period of transmission is configured by RRC or MAC CE, and after the configuration, the gradient information of the second information is periodically transmitted according to the period.

[0187] In the semi-persistently transmitted mode, the gradient information of the second information is periodically transmitted after being activated by MAC CE or DCI after the configuration, until being deactivated or after a specified period.

[0188] In the non-periodically transmitted mode, the gradient information of the second information is transmitted a specified number of times after being activated by MAC CE or DCI after the configuration. The specified number of times can be configured by the network, such as once or multiple times, and is not limited in particular.

[0189] In the embodiments of the present application, the gradient information of the second information can be the gradient information of the second information greater than or equal to a preset first threshold. That is, the second device filters the gradient information of the second information greater than or equal to the preset first threshold before transmitting the gradient information of the second information, and then transmits it to the first device. The gradient information of the second information less than the preset first threshold is considered meaningless for training the first AI unit, and therefore does not need to be transmitted to the first device.

[0190] S308: Training the first AI unit according to the gradient information of the second information.

[0191] In one embodiment, the above step S308 can include training the first AI unit according to the training parameter information and the gradient information of the second information. The training parameter information is various parameter information related to training the first AI unit.

[0192] In the embodiments of the present application, the training parameter information can include at least one of the following:

[0193] 1) batch size;

[0194] 2) optimizer type;

[0195] 3) learning rate;

[0196] 4) learning rate decay; for example, the exponential decay rate of the first estimation, the exponential decay rate of the second estimation, the exponential decay rate of the Nth estimation, etc.

[0197] 5) loss function; wherein the loss function can include a combination function of at least one of the following: error of output of the target model and label, mean square error, normalized mean square error, correlation, entropy, mutual information, constant. Wherein the target model is the first AI unit or the second AI unit. The combination function includes at least one of the following: combination of various common mathematical operations such as addition, subtraction, multiplication, division, N power, N root, logarithm, derivation, partial derivation, etc. Wherein N is an arbitrary number, which can be a positive number or a negative number or 0, a real number or a complex number.

[0198] 6) regularization method and parameter;

[0199] 7) initialization method and parameter;

[0200] 8) dropout method and parameter;

[0201] 9) training period;

[0202] 10) number of iterations;

[0203] 11) training termination condition.

[0204] In an embodiment, the loss function described above can be obtained by weighted combination of loss information of multiple parts of the output of the target model. The weighted combination can be linear average, multiplicative average and other common average methods. The multiple parts are obtained by dividing according to at least one of the following: spatial resource, code resource, frequency resource and time resource. Wherein the frequency resource can include resource block RB, subband or precoding resource block group PRG. The time resource can include subcarrier, symbol, time slot or half time slot. The loss information can include loss value and / or loss associated function. Wherein the target model is the first AI unit or the second AI unit.

[0205] In an embodiment of the present application, the training parameter information can be preset by the protocol, or indicated by the first device or the second device using at least one of the following methods:

[0206] 1) RRC indication;

[0207] 2) MAC CE indication;

[0208] 3) control information indication; for example, DCI indication of NW or UCI indication of UE.

[0209] S310: Establish the matching relationship of the trained first AI unit and the second AI unit.

[0210] In an embodiment of the present application, the identification of the matching relationship described above can include at least one of the following:

[0211] 1) model ID (model ID);

[0212] 2) functionality ID;

[0213] 3) dataset ID;

[0214] 4) pairing ID;

[0215] 5) Associate ID.

[0216] In an embodiment, the second information can include a bit sequence, and before the second information is sent to the second device using the lossy transmission, the method further includes modulating the bit sequence in the second information using at least one of the following modulation methods:

[0217] 1) binary phase shift keying (BPSK);

[0218] 2) pi / 2-BPSK;

[0219] 3) quadrature phase shift keying (QPSK);

[0220] 4) 16 quadrature amplitude modulation (QAM);

[0221] 5) 64 QAM;

[0222] 6) 256 QAM;

[0223] 7) 1024 QAM;

[0224] 8) 4096 QAM;

[0225] 9) AI-based modulation.

[0226] In the embodiments of the present application, the above method can further include:

[0227] The first device sends preset channel information to the second device using lossy transmission or lossless transmission. The preset channel information is used to add channel effects to the second information using the corresponding channel.

[0228] In the embodiments of the present application, the preset channel information can include at least one of the following:

[0229] 1) signal to noise ratio (SNR);

[0230] 2) SNR range;

[0231] 3) Bit error rate;

[0232] 4) Bit error rate range;

[0233] 5) CQI;

[0234] 6) CQI range;

[0235] 7) Modulation and Coding Scheme (MCS);

[0236] 8) MCS range;

[0237] 9) Raw channel information of a transmission channel; wherein the raw channel information of the transmission channel can be set by means of a data set and associated using an ID;

[0238] 10) Channel model.

[0239] In an implementation, the preset channel information can be preset by a protocol, for example, it is explicitly stated in the protocol that the preset channel adopted by the first device and the second device is an Additive White Gaussian Noise (AWGN) channel.

[0240] In another implementation, the preset channel information can be indicated by the first device or the second device using at least one of the following ways:

[0241] 1) Radio Resource Control (RRC) indication;

[0242] 2) Medium Access Control Control Element (MAC CE) indication;

[0243] 3) Control information indication; for example, NW DCI indication or UE UCI indication.

[0244] In the embodiments of the application, the training of the first AI unit or the training of the second AI unit can be triggered by one of the following:

[0245] 1) Network indication;

[0246] 2) Terminal reporting;

[0247] 3) Predefined event.

[0248] Among them, the predefined event for triggering the training of the first AI unit or the training of the second AI unit can include at least one of the following:

[0249] 1) in a wifi state;

[0250] 2) power exceeds a second threshold;

[0251] 3) in a charging state;

[0252] 4) supports a preset traffic mode; for example, the sending of the second information does not count as traffic overhead;

[0253] 5) in a preset traffic mode;

[0254] 6) no service demand; for example, there is no need to train the first AI unit or the second AI unit;

[0255] 7) a performance index is lower than a third threshold.

[0256] The above method provided by the embodiments of the present application, by the first device, inputs the first information into the first AI unit to obtain the second information, transmits the second information to the second device using a preset mode, receives gradient information of the second information sent by the second device, trains the first AI unit according to the training parameter information and the gradient information of the second information, and establishes a matching relationship between the trained first AI unit and the second AI unit, wherein the first information is CSI, the first AI unit is used for encoding the first information, and the preset mode is used for adding channel influence, so that the second device can obtain the second information with added channel influence and can consider the influence of the channel in the AI-based encoding process, realizing data acquisition of AI-based joint source channel coding, helping the training of CSI feedback enhancement, and being capable of improving the practicality of the training model and making the model application more extensive.

[0257] FIG. 4 shows a flowchart of a method for data acquisition of joint source channel coding provided by the embodiments of the present application, and the method 400 can be executed by the second device. As shown in FIG. 4, the method can include the following steps.

[0258] S402: The second device receives the second information sent by the first device using a preset mode.

[0259] The second device can be a terminal or a network side device, and is not limited in particular. The preset mode is used for adding channel influence, including: a channel and lossless transmission mode, a lossless transmission mode, and a lossy transmission mode. The channel influence at least includes signal changes caused by channel transmission, such as channel interference, channel fading, channel noise, etc.

[0260] The second information is obtained by the first device inputting the first information into the first AI unit, the first information includes CSI, the first AI unit is used for encoding the first information, and the second information includes encoded CSI.

[0261] In the embodiments of the present application, the second information can include at least one of a bit sequence and a symbol sequence. The bit sequence corresponds to bit-level joint source channel coding, and the symbol sequence corresponds to symbol-level joint source channel coding. The second information can be carried in one of the following: DCI, MAC CE, RRC, PUSCH, PDSCH, PDCCH, and PUCCH. The second information can be configured by RRC or MAC CE to be periodically transmitted, semi-persistently transmitted, or aperiodically transmitted.

[0262] In an implementation, the second information can be used for the second device to train the second AI unit.

[0263] In the embodiments of the present application, the step S402 can include one of the following:

[0264] The second device receives the second information transmitted by the first device using lossless transmission through the corresponding channel according to the preset channel information.

[0265] The second device receives the second information transmitted by the first device using lossless transmission, and transmits the second information through the corresponding channel according to the preset channel information.

[0266] The second device receives the second information transmitted by the first device using lossy transmission.

[0267] In an implementation, after the second information is transmitted through the corresponding channel according to the preset channel information, the second device can further perform an operation of training the second AI unit.

[0268] In the embodiments of the present application, the preset manner is a lossy transmission manner, and correspondingly, after the step S402, the second device further performs at least one of the following operations on the second information:

[0269] 1) De-mapping; wherein the first device performs layer mapping in the process of lossy transmission.

[0270] 2) De-interleaving; wherein the first device performs interleaving in the process of lossy transmission.

[0271] 3) De-padding; wherein the first device performs padding in the process of lossy transmission.

[0272] 4) De-scrambling; wherein the first device performs scrambling in the process of lossy transmission.

[0273] 5) Receive combining; wherein the first device performs precoding in the process of lossy transmission.

[0274] In the embodiments of the present application, the preset manner includes a lossless transmission manner, and correspondingly, after the step S402, the second device further performs at least one of the following operations on the second information:

[0275] 1) source coding; wherein the first device performs source encoding in the course of lossless transmission;

[0276] 2) channel coding; wherein the first device performs channel encoding in the course of lossless transmission;

[0277] 3) demodulation; wherein the first device performs modulation in the course of lossless transmission;

[0278] 4) de-mapping; wherein the first device performs layer mapping in the course of lossless transmission;

[0279] 5) de-interleaving; wherein the first device performs interleaving in the course of lossless transmission;

[0280] 6) de-padding; wherein the first device performs padding in the course of lossless transmission;

[0281] 7) de-scrambling; wherein the first device performs scrambling in the course of lossless transmission;

[0282] 8) receive combining; wherein the first device performs precoding in the course of lossless transmission.

[0283] The above method provided by the embodiments of the present application, through the second device receiving the second information sent by the first device using the preset manner, wherein the preset manner is used to add channel influence, the second information is obtained by the first device inputting the first information into the first AI unit, the first information is CSI, and the first AI unit is used to encode the first information, so that the second device can obtain the second information with added channel influence, and can consider the influence of the channel in the AI-based encoding process, realizing the data acquisition of the AI-based joint source channel coding, which is helpful for the training of CSI feedback enhancement, can improve the practicality of the training model, and makes the model application more widely.

[0284] FIG. 5 shows a flowchart of a method for data acquisition of joint source channel coding provided by the embodiments of the present application, which method 500 can be performed by the second device. As shown in FIG. 5, the method can include the following steps.

[0285] S502: The second device receives second information sent by the first device using a preset manner.

[0286] The second device can be a terminal or a network side device, which is not limited in particular. The preset manner is used to add channel influence, including: a channel and lossless transmission manner, a lossless transmission manner, and a lossy transmission manner. The channel influence at least includes signal changes caused by channel transmission, such as channel interference, channel fading, channel noise, etc.

[0287] The second information is obtained by inputting, by the first device, the first information into a first AI unit, the first information including CSI, and the first AI unit being configured to encode the first information, and the second information including the encoded CSI.

[0288] S504: inputting the second information into a second AI unit to decode to obtain recovered first information.

[0289] S506: training the second AI unit according to training parameter information, first related information, and the recovered first information to obtain gradient information of the second information.

[0290] The training parameter information refers to parameter information related to the training of the second AI unit. The training parameter information can include at least one of batch size, optimizer type, learning rate, learning rate decay, loss function, regularization method and parameter, initialization method and parameter, dropout method and parameter, training period, iteration number, and training termination condition.

[0291] In an embodiment, the loss function can be obtained by weighted combination of loss information of multiple parts of the output of the target model. The weighted combination can be linear average, multiplicative average, or other common average methods. The multiple parts are obtained by dividing according to at least one of spatial domain resources, code domain resources, frequency domain resources, and time domain resources. The frequency domain resources can include resource blocks (RBs), subbands, or precoding resource block groups (PRGs). The time domain resources can include subcarriers, symbols, slots, or half-slots. The loss information can include loss values and / or loss-related functions. The target model can be the first AI unit or the second AI unit.

[0292] In an embodiment, the training parameter information can be preset by a protocol, or indicated by the first device or the second device using at least one of RRC indication, MAC CE indication, and control information indication. The control information indication includes DCI indication of the NW or UCI indication of the UE.

[0293] In an embodiment, the method can further include receiving, by the second device, the first related information sent by the first device using lossy transmission or lossless transmission.

[0294] The first related information is information related to the first information and the encoding and decoding. The first related information can include at least one of samples in a CSI dataset, sample IDs in the CSI dataset, encoded dimension information, channel impact information, and a rule for obtaining the first information.

[0295] In an embodiment, the rule for obtaining the first information can be determined between the first device and the second device by a protocol, RRC, DCI, or UCI indication.

[0296] In the embodiments of the present application, the acquisition rule of the first information can include an iteration number index, which is used to indicate that a preset number of samples are acquired from a current position in the samples arranged in a preset order, the current position being an initial position or a position where the last sample acquisition ends. The iteration number refers to the iteration number of training of the first AI unit or the second AI unit. In the first iteration of training, the current position is the initial position, and in each subsequent iteration of training, the current position is the position where the last sample acquisition ends. In addition, the first device can also send the training iteration number to the second device.

[0297] In the embodiments of the present application, the acquisition of the first related information can be triggered by one of the following: network indication, terminal reporting, and predefined event. The predefined event can include at least one of the following: being in a wifi state, battery power exceeding a second threshold, being in a charging state, supporting a preset traffic mode, being in a preset traffic mode, having no traffic demand, and a performance index being lower than a third threshold.

[0298] S508: sending gradient information of the second information to the first device.

[0299] The gradient information of the second information is used by the first device to train the first AI unit in combination with the training parameter information.

[0300] In the embodiments of the present application, the gradient information of the second information can be carried in one of the following: DCI, MAC CE, RRC, PUSCH, PDSCH, PDCCH, and PUCCH. The gradient information of the second information can be configured by RRC or MAC CE to be periodically sent, semi-persistently sent, or non-periodically sent.

[0301] In the embodiments of the present application, the gradient information of the second information can be gradient information of the second information greater than or equal to a preset first threshold. That is, the second device filters out the gradient information of the second information greater than or equal to the preset first threshold before sending the gradient information of the second information, and then sends it to the first device. The gradient information of the second information less than the preset first threshold is considered to be meaningless for the training of the first AI unit, and therefore does not need to be sent to the first device.

[0302] In the embodiments of the present application, the above method can further include:

[0303] The second device receives preset channel information sent by the first device using lossy transmission or lossless transmission.

[0304] The preset channel information is used to add channel influence to the second information using the corresponding channel.

[0305] In the embodiments of the present application, the preset channel information can include at least one of the following: SNR, SNR range, bit error rate, bit error rate range, CQI, CQI range, MCS, MCS range, raw channel information of a transmission channel, and a channel model.

[0306] In the embodiments of the present application, the preset channel information can be indicated by the first device or the second device using at least one of the following manners: protocol indication, RRC indication, MAC CE indication, DCI indication, and UCI indication.

[0307] The above method provided by the embodiments of the present application receives, by the second device, the second information sent by the first device using the preset manner, inputs the second information into the second AI unit for decoding to obtain the recovered first information, trains the second AI unit according to the training parameter information, the first related information and the recovered first information to obtain the gradient information of the second information, and sends the gradient information of the second information to the first device, wherein the preset manner is used to add channel influence, the second information is obtained by the first device by inputting the first information into the first AI unit, and the first information is CSI. The first AI unit is used to encode the first information, so that the second device can obtain the second information with added channel influence and can consider the influence of the channel in the AI-based encoding process, thereby realizing data acquisition of AI-based joint source channel coding, helping to enhance the training of CSI feedback, improving the practicality of the training model, and making the model more widely applicable.

[0308] The following describes the AI unit and AI training related to each of the above embodiments.

[0309] In the embodiments of the present application, AI can also refer to machine learning (ML). The AI unit (including the first AI unit and the second AI unit) can also be referred to as an AI model, an AI structure, etc., or the AI unit can refer to a processing unit that can implement specific algorithms, formulas, processing flows, capabilities, etc. related to AI, or the AI unit can be a processing method, algorithm, function, module or unit for a specific data set, or the AI unit can be a processing method, algorithm, function, module or unit running on AI-related hardware such as GPU, NPU, TPU, ASIC, etc. The present application does not make specific limitations on this. Optionally, the specific data set includes the input and / or output of the AI unit.

[0310] Optionally, the identity of the AI unit can be one or more of the following identities: AI unit identity, AI structure identity, AI algorithm identity, function identity, physical identity, logical identity, global identity, local identity, identity of a specific data set associated with the AI unit, identity of a specific scene, environment, channel feature, or device related to the AI, identity of a function, feature, capability, or module related to the AI, which is not specifically limited in the present application.

[0311] In the embodiments of the present application, the first AI unit and the second AI unit can be aligned with the third information before training to ensure the accuracy of training. The third information can include at least one of the following: structural features of the model, load quantization method of the model, estimation accuracy of the model, output accuracy of the model. The model refers to the first AI unit or the second AI unit.

[0312] In the embodiments of the present application, the structural features of the above-mentioned model can include at least one of the following: model structure, model basic structural features, model sub-module structural features, model layer number, model neuron number, model size, model complexity, model parameter quantization parameter.

[0313] In the embodiments of the present application, the model basic structural features can include at least one of the following: whether to contain a full connection structure, whether to contain a convolution structure, whether to contain a long short-term memory model LSTM structure, whether to contain an attention structure, whether to contain a residual structure.

[0314] In the embodiments of the present application, the model neuron number can include at least one of the following: full connection neuron number, convolution neuron number, memory neuron number, attention neuron number, residual neuron number. And / or, the model neuron number can include at least one of the following: all types of neuron number, single type of neuron number, entire model neuron number, single layer or several layers of neuron number.

[0315] In the embodiments of the present application, the model parameter quantization parameter can include at least one of the following: model parameter quantization method, single neuron parameter quantization bit number.

[0316] The model parameter quantization method can include at least one of the following: uniform quantization method, non-uniform quantization method, weight sharing quantization method or grouping quantization method, parameter coding quantization method, transform domain quantization method, product quantization method.

[0317] In the embodiments of the present application, the load quantization method can include at least one of the following: quantization method, dimension of features before and after quantization, quantization method used when quantizing.

[0318] The quantization method used in the quantization can include at least one of the following: synchronization of codebook content and codebook usage method when quantization is performed using a codebook, and synchronization of quantization rules when quantization is performed using specific rules.

[0319] The quantization rules can include at least one of the following: N quantization intervals, quantization methods, and N is a positive integer. The quantization methods can include at least one of the following: uniform quantization method, non-uniform quantization method, weight sharing quantization method or grouping quantization method, parameter coding quantization method, transform domain quantization method, and product quantization method.

[0320] The synchronization method can include any of the following when synchronizing the codebook content and the codebook usage method, and / or the quantization rules: selecting a set number representing the selected method from a predefined method set when synchronizing, and directly sending the codebook content.

[0321] In the embodiments of the present application, the manner in which the first device and the second device align the third information can include at least one of the following:

[0322] The first device or other network element sends the third information to the second device when sending the target information to the second device;

[0323] The second device or other network element sends the third information to the first device when sending the target information to the first device;

[0324] The first device or other network element sends the third information before sending the target information to the second device;

[0325] The second device or other network element sends the third information before sending the target information to the first device;

[0326] The second device sends the third information when requesting the target information;

[0327] The first device sends the third information when requesting the target information;

[0328] The first device or other network element sends the third information and the consent information when the second device requests the target information, and the consent information is used to indicate consent to the request of the second device;

[0329] The second device or other network element sends the third information and the consent information when the first device requests the target information, and the consent information is used to indicate consent to the request of the first device;

[0330] The target information includes at least one first information related to a first action of the model and at least one second information corresponding to the at least one first information.

[0331] In the embodiments of the present application, the second device or other network element sends the target information after the first device sends the confirmation information of the third information; and / or the first device or other network element sends the target information after the second device sends the confirmation information of the third information.

[0332] In the embodiments of the present application, the manner in which the first device and the second device align the third information includes at least one of the following:

[0333] After one of the devices receiving the third information sends the confirmation information of the third information, the first AI unit and / or the second AI unit can use the model associated with the third information;

[0334] After one of the devices receiving the third information sends the confirmation information of the third information and a first time duration elapses, the first AI unit and / or the second AI unit can use the model associated with the third information;

[0335] After the sending time or the receiving time of the third information elapses a first time duration, the first AI unit and / or the second AI unit can use the model associated with the third information.

[0336] The first time duration can be determined by any of the following: carried by the third information, carried by the confirmation information of the third information, carried by other associated information or signaling of the third information, agreed by a protocol, determined by the capability of the first device or the second device.

[0337] In the embodiments of the present application, the joint source-channel coding can be defined or described in the following manners:

[0338] A) input and output of the AI / ML unit;

[0339] B) a signal processing flow;

[0340] C) mapping of input and output information.

[0341] The input and output of the AI / ML unit can include the following:

[0342] A1) AI / ML coding method, specifically including:

[0343] a. AI / ML coding unit;

[0344] b. AI / ML decoding unit.

[0345] A2) the input of the AI / ML coding unit is CSI source information, which can include at least one of the following:

[0346] a. CQI;

[0347] b. PMI;

[0348] c. CRI;

[0349] d.SSBRI;

[0350] e.LI;

[0351] f.RI;

[0352] g.L1-RSRP;

[0353] h.L1-SINR;

[0354] i. Capability Index;

[0355] j.TDCP;

[0356] k. Bare channel information;

[0357] l. Processed raw channel information, such as transforming spatial frequency domain channel information to angular time delay domain channel and / or truncating the channel;

[0358] The information source corresponding to m.PMI is the V matrix after the channel SVD decomposition.

[0359] A3) The output of the AI / ML coding unit is the bit sequence after joint source-channel coding (bit-level joint source-channel coding); or,

[0360] The output of the AI / ML coding unit is the symbol sequence after joint source-channel coding (symbol-level joint source-channel coding).

[0361] A4) The output of the AI / ML coding unit cannot use channel coding.

[0362] The input to the A5) AI / ML decoding unit includes one of the following:

[0363] a. The bit sequence received by NW that is related to the output of the AI / ML encoder;

[0364] b. The symbol sequence received by NW that is associated with the AI / ML encoder output.

[0365] A6) The output of the AI / ML decoding unit is the recovered CSI source information, which may include at least one of the following:

[0366] a.CQI;

[0367] b. PMI;

[0368] c.CRI;

[0369] d.SSBRI;

[0370] e.LI;

[0371] f.RI;

[0372] g. L1-RSRP;

[0373] h. L1-SINR;

[0374] i. Capability Index;

[0375] j. TDCP;

[0376] k. Raw channel information;

[0377] l. Processed raw channel information, such as transforming spatial-frequency domain channel information to angular-delay domain channel information and / or truncating the channel;

[0378] m. Source corresponding to PMI, i.e. V matrix after SVD decomposition of the channel.

[0379] The above-mentioned signal processing procedure can include the following four contents:

[0380] B1) A joint source-channel encoding processing procedure for CSI information feedback, including the following modules:

[0381] a. Joint source-channel encoding processing module, located at the UE side;

[0382] b. Joint source-channel decoding processing module, located at the NW side, such as at the gNB side.

[0383] B2) Channel estimation on CSI-RS to obtain CSI information, which can include at least one of the following:

[0384] a. CQI;

[0385] b. PMI;

[0386] c. CRI;

[0387] d. SSBRI;

[0388] e. LI;

[0389] f. RI;

[0390] g. L1-RSRP;

[0391] h. L1-SINR;

[0392] i. Capability Index;

[0393] j. TDCP;

[0394] k. Raw channel information;

[0395] l. Processed bare channel information, such as converting the space-frequency domain channel information to the angle-delay domain channel and / or performing a truncation operation on the channel.

[0396] m. The source corresponding to the PMI, that is, the V matrix after SVD decomposition of the channel.

[0397] B3) The UE side uses a joint source channel coding processing module to process the CSI information to obtain encoded information, which can include one of the following:

[0398] a. Bit sequence (bit-level joint source coding) ;

[0399] b. Symbol sequence (symbol-level joint source coding).

[0400] B4) The NW side uses a joint source channel decoding processing module to process the encoded bit sequence or symbol sequence containing channel interference and noise, and can obtain recovered CSI information.

[0401] Wherein, the above input-output information mapping can include one of the following:

[0402] C1) A joint source channel coding mapping on the UE side, which maps the CSI information to the joint source channel coded information bit sequence (bit-level joint source channel coding) or symbol sequence (symbol-level joint source channel coding).

[0403] C2) A joint source channel decoding mapping on the NW side, which maps the encoded output bit sequence (bit-level joint source channel coding) or symbol sequence (symbol-level joint source channel coding) containing channel interference and noise to the recovered CSI information.

[0404] The application of each embodiment will be described in detail below in combination with six scenarios.

[0405] The embodiments of the present application relate to the following two encoding methods: joint source channel coding (JSCC) and separate source channel coding (SSCC).

[0406] In the embodiments of the present application, the first device transmits the second information to the second device using a preset method, including but not limited to one of the following three transmission methods:

[0407] The first transmission method (channel impact adding mode at the sending end) : After the second information passes through the channel corresponding to the preset channel information, it is sent to the second device using lossless transmission.

[0408] The second transmission mode (the mode of adding channel influence at the receiving end): the second information is transmitted to the second device using lossless transmission. In this mode, after receiving the second information, the second device can obtain the encoded information added with channel influence through the channel corresponding to the preset channel information, thereby completing data acquisition.

[0409] The third transmission mode (the mode of lossy transmission): the second information is transmitted to the second device using lossy transmission.

[0410] In combination with the above two encoding modes and three transmission modes, a plurality of scenarios can be obtained, and six scenarios are exemplified as follows.

[0411] Scenario one: symbol-level JSCC joint training (Joint training), the mode of lossy transmission;

[0412] Scenario two: bit-level JSCC joint training (Joint training), the mode of adding channel influence at the sending end UE;

[0413] Scenario three: symbol-level JSCC joint training (Joint training), the mode of adding channel influence at the receiving end NW;

[0414] Scenario four: symbol-level JSCC separate training (Separate training), the receiving end NW trains the decoder, and the mode of adding channel influence at the NW side;

[0415] Scenario five: symbol-level JSCC separate training (Separate training), the receiving end NW trains the decoder, and the mode of adding channel influence at the sending end UE;

[0416] Scenario six: symbol-level JSCC separate training (Separate training), the UE trains the encoder, and the mode of not adding channel influence.

[0417] In scenario one, the first device is a UE, the second device is a NW, the channel of the UE and the NW is approximately or consistent with the actual deployment when Joint training, and the UE and the NW know the selected training data set, and the symbol-level JSCC Joint training mode is used. The data acquisition method of joint source channel coding in this scenario can include the following processes:

[0418] Step 1: the UE selects samples with a batch size of 128 from the CSI data set, records the sample Id (such as 3, 6, …), and obtains the first information, i.e., the CSI information to be encoded.

[0419] Step 2: The UE inputs the CSI information to be encoded into a first AI unit, i.e., a symbol-level joint source-channel encoder, to obtain second information, i.e., encoded CSI symbol information.

[0420] Step 3: The UE performs layer / resource mapping on the encoded CSI symbol information and transmits the encoded CSI symbol information to the NW through lossy transmission.

[0421] Step 4: The UE transmits first related information to the NW through lossless transmission, wherein the first related information includes the shape of the encoded CSI symbol information and a sample ID.

[0422] Step 5: The NW receives the encoded CSI symbol information and the first related information, performs demapping on the encoded CSI symbol information, performs a reshape operation on the encoded CSI symbol information using the first related information to obtain the original shape, inputs the original shape into a second AI unit, i.e., a symbol-level joint source-channel decoder, and obtains recovered CSI information.

[0423] Step 6: The NW obtains CSI information to be encoded according to the sample ID, calculates the loss of the recovered CSI information and the CSI information to be encoded using a loss function according to training parameter information, and trains the symbol-level joint source-channel decoder using a Backpropagation algorithm to obtain gradient information of the encoded CSI symbol information.

[0424] Step 7: The NW transmits the gradient information of the encoded CSI symbol information to the UE through lossless transmission.

[0425] Step 8: The UE receives the gradient information of the encoded CSI symbol information, and trains the symbol-level joint source-channel encoder according to the training parameter information and the gradient information of the encoded CSI symbol information.

[0426] Step 9: Repeat the above steps 1-8 until the training termination condition is met or the number of training iterations is completed.

[0427] Step 10: Establish a matching relationship between the trained UE-side symbol-level joint source-channel encoder and the NW-side symbol-level joint source-channel decoder.

[0428] In scenario two: the first device is the UE, the second device is the NW, and the bit-level JSCC joint training method is used. The channel impact is added at the sending end UE side, and the NW does not know the training data set selected by the UE. The data collection method of the joint source-channel encoding in this scenario can include the following processes:

[0429] Step 1: The UE selects a sample of batch_size of 128 from the CSI dataset to obtain first information, i.e., CSI information to be encoded.

[0430] Step 2: The UE inputs the CSI information to be encoded into a first AI unit, i.e., a bit-level joint source-channel encoder, to obtain second information, i.e., encoded CSI bit information.

[0431] Step 3: The UE transmits the encoded CSI bit information through the channel according to the first channel information, adds channel effects, and obtains the encoded CSI bit information through the channel, and then transmits it to the NW through lossless transmission.

[0432] Step 4: The UE transmits the first related information to the NW through lossless transmission, wherein the first related information includes: the shape of the encoded CSI bit information and the CSI sample information.

[0433] Step 5: The NW receives the encoded CSI bit information through the channel and the first related information, uses the first related information to perform a reshape operation on the encoded CSI bit information through the channel, obtains the original shape, and then inputs it into a second AI unit, i.e., a bit-level joint source-channel decoder, to obtain recovered CSI information.

[0434] Step 6: The NW uses a loss function to calculate the loss of the recovered CSI information and the CSI sample information according to the training parameter information, and uses a Backpropagation algorithm to train the bit-level joint source-channel decoder to obtain gradient information of the encoded CSI bit information.

[0435] Step 7: The NW transmits the gradient information of the encoded CSI bit information to the UE through lossless transmission.

[0436] Step 8: The UE receives the gradient information of the encoded CSI bit information, and trains the bit-level joint source-channel encoder according to the training parameter information and the gradient information of the encoded CSI bit information.

[0437] Step 9: Repeat steps 1-8 above until the training termination condition is met or the number of training iterations is completed.

[0438] Step 10: Establish a matching relationship between the trained UE-side bit-level joint source-channel encoder and the NW-side bit-level joint source-channel decoder.

[0439] In scenario three, the first device is a UE, the second device is a NW, and a joint training method of symbol-level JSCC is used. A channel impact is added at the receiving end NW side. The NW does not know the training data set selected by the UE. The data acquisition method of joint source channel coding in this scenario can include the following processes:

[0440] Step 1: The UE selects a sample with a batch size of 128 from a CSI data set to obtain first information, i.e., CSI information to be encoded.

[0441] Step 2: The UE inputs the CSI information to be encoded into a first AI unit, i.e., a symbol-level joint source channel encoder, to obtain second information, i.e., encoded CSI symbol information.

[0442] Step 3: The UE sends the encoded CSI symbol information to the NW through lossless transmission. After receiving the information, the NW transmits the encoded CSI symbol information through a corresponding channel according to first channel information.

[0443] Step 4: The UE sends first related information to the NW through lossless transmission. The first related information includes the shape of the encoded CSI symbol information and CSI sample information.

[0444] Step 5: The NW uses the first related information to perform a reshape operation on the encoded CSI symbol information through the channel, obtains the original shape, and inputs the information into a second AI unit, i.e., a symbol-level joint source channel decoder, to obtain recovered CSI information.

[0445] Step 6: The NW uses a loss function to calculate the loss of the recovered CSI information and the CSI information to be encoded according to training parameter information, and uses a backpropagation algorithm to train the symbol-level joint source channel decoder to obtain gradient information of the encoded CSI symbol information.

[0446] Step 7: The NW sends the gradient information of the encoded CSI symbol information to the UE through lossless transmission.

[0447] Step 8: The UE receives the gradient information of the encoded CSI symbol information, and trains the symbol-level joint source channel encoder according to the training parameter information and the gradient information of the encoded CSI symbol information.

[0448] Step 9: Repeat the above steps 1 to 8 until the training termination condition is met or the number of training iterations is completed.

[0449] Step 10: Establish a matching relationship between the trained UE-side symbol-level joint source channel encoder and the NW-side symbol-level joint source channel decoder.

[0450] In scenario four: the first device is a UE, the second device is a NW, and a symbol-level SSCC Separate training method is used. Before Separate training, the UE has trained the symbol-level joint source-channel encoder on the UE side through a CSI data set and preset channel information, and there is no need to train the encoder again. Moreover, a channel impact is added on the receiving end NW side, and the decoder is trained by the NW. In this scenario, the data acquisition method of joint source-channel coding can include the following processes:

[0451] Step 1: The UE encodes all samples in the first information, i.e., a CSI data set, using the first AI unit, i.e., a symbol-level joint source-channel encoder, to obtain the second information, i.e., an encoded CSI data set.

[0452] Step 2: The UE sends the encoded CSI data set to the NW through lossless transmission, and the NW obtains the encoded CSI data set through the channel according to the first channel information.

[0453] Step 3: The UE sends the first related information, i.e., the CSI data set, to the NW through lossless transmission.

[0454] Step 4: The NW receives the CSI data set, and trains the second AI unit, i.e., a symbol-level joint source-channel decoder, on the NW side using the encoded CSI data set through the channel and the received CSI data set according to the training parameter information until the training termination condition is met or the training iteration number is completed.

[0455] Step 5: The matching relationship between the trained UE-side symbol-level joint source-channel encoder and the NW-side symbol-level joint source-channel decoder is established.

[0456] In scenario five: the first device is a UE, the second device is a NW, and a symbol-level SSCC Separate training method is used. Before Separate training, the UE has trained the joint source-channel encoder on the UE side through a CSI data set and preset channel information, and there is no need to train the encoder again. Moreover, a channel impact is added on the sending end UE side, and the decoder is trained by the NW side. In this scenario, the data acquisition method of joint source-channel coding can include the following processes:

[0457] Step 1: The UE encodes all samples in the first information, i.e., a CSI data set, using the first AI unit, i.e., a symbol-level joint source-channel encoder, to obtain the second information, i.e., an encoded CSI data set.

[0458] Step 2: The UE transmits the encoded CSI data set through the corresponding channel according to the first channel information to obtain an encoded CSI data set through the channel, and then transmits the encoded CSI data set to the NW through lossless transmission.

[0459] Step 3: The UE transmits the first related information, i.e., the CSI data set, to the NW through lossless transmission.

[0460] Step 4: The NW receives the CSI data set, and uses the encoded CSI data set through the channel and the received CSI data set to train the second AI unit, i.e., the symbol-level joint source channel decoder, on the NW side according to the training parameter information until a training termination condition is met or a training iteration number is completed.

[0461] Step 5: The matching relationship between the trained UE-side symbol-level joint source channel encoder and the NW-side symbol-level joint source channel decoder is established.

[0462] In scenario six, the first device is the NW, and the second device is the UE. The symbol-level SSCC Separate training method is used. Before Separate training, the NW has trained the joint source channel codec on the NW side using the CSI data set and preset channel information, and does not need to train the decoder again. The encoder is trained on the UE side, and no channel influence needs to be added. The data collection method for joint source channel coding in this scenario can include the following processes:

[0463] Step 1: The NW uses the symbol-level joint source channel encoder to encode all samples in the first information, i.e., the CSI data set, to obtain the second information, i.e., the encoded CSI data set.

[0464] Step 2: The NW transmits the encoded CSI data set and the first related information, i.e., the CSI data set, to the UE through lossless transmission respectively or together.

[0465] Step 3: The UE receives the encoded CSI data set and the first related information, and uses the first CSI data set and the encoded CSI data set to train the second AI unit, i.e., the symbol-level joint source channel encoder, on the UE side according to the training parameter information until a training termination condition is met or a training iteration number is completed.

[0466] Step 4: The matching relationship between the trained UE-side symbol-level joint source channel encoder and the NW-side symbol-level joint source channel decoder is established.

[0467] The application of the above method is specifically described below by taking a terminal and a network-side device as an example.

[0468] The first example: the first device is a terminal, and the second device is a network side device. Based on the data collection method of the joint source channel coding, the training and application based on the joint training method can specifically include the following steps:

[0469] Step 1: The terminal inputs first information into a first AI unit to obtain second information; wherein the first information includes CSI, the first AI unit is used for encoding the first information, and the second information includes the encoded CSI.

[0470] Step 2: The terminal transmits the second information to the network side device using a preset method, wherein the preset method is used to add channel influence.

[0471] Step 3: The terminal receives gradient information of the second information sent by the network side device, wherein the gradient information of the second information is obtained by the network side device according to the training of the second AI unit based on the second information, and the second AI unit is used for decoding the second information.

[0472] Step 4: The terminal trains the first AI unit according to the gradient information of the second information.

[0473] Step 5: The terminal establishes a matching relationship between the trained first AI unit and the trained second AI unit of the network side device.

[0474] Correspondingly, the network side device establishes a matching relationship between the trained second AI unit and the trained first AI unit of the terminal.

[0475] The above process is the training process, and after the training is completed, the training result can be used for actual CSI feedback.

[0476] Step 6: In the CSI feedback scenario, the terminal encodes the CSI using the trained first AI unit and sends it to the network side device. Correspondingly, the network side device can decode the received encoded CSI using the trained second AI unit to obtain the recovered CSI.

[0477] Wherein, the CSI feedback scenario refers to the scenario in which the terminal feeds back the CSI to the network side device, which can be various, including but not limited to at least one of the following: a scenario in which the cell edge needs to feed back the CSI, a scenario in which the bandwidth allocated for CSI feedback is small, a scenario in which the feedback channel is a fast-varying channel and cannot be accurately estimated, and the like.

[0478] The above process can consider the influence of the channel in the AI-based encoding process, and realize data collection of AI-based joint source channel coding. The bilateral joint training of CSI feedback enhancement based on the results of data collection can improve the practicality of the trained model, making the model more widely applicable. The application of the trained model to CSI feedback can bring gain to CSI, improve the accuracy of CSI feedback, and improve transmission efficiency.

[0479] The second example: the first device is a network side device, and the second device is a terminal. Based on the above-mentioned joint source channel coding data collection method, the training and application process based on the separate training mode can specifically include the following steps:

[0480] Step 1: the network side device initially trains a decoder, inputs first information into a first AI unit to obtain second information; wherein the first information includes CSI, and the first AI unit is used for encoding the first information; and the second information includes the encoded CSI.

[0481] Step 2: the network side device transmits the second information to the terminal using a preset mode, wherein the preset mode is used to add channel influence.

[0482] Step 3: the terminal receives the second information sent by the network side, and trains a second AI unit according to the training parameter information, the first related information and the second information.

[0483] The training parameter information is the parameter information related to the training of the second AI unit. The first related information is the information related to the first information and the coding and decoding, which can be sent by the network side device to the terminal. The second AI unit is used for encoding the CSI.

[0484] In an embodiment, the first related information can be a CSI data set, including a plurality of CSI samples. The network side device can send the CSI data set to the terminal for terminal side training. In the process of terminal training, the first related information, i.e., the CSI data set, is input, and the second information, i.e., the encoded CSI, is used as a label to train the second AI unit, i.e., the encoder.

[0485] Step 4: the network side device establishes a matching relationship between the decoder and the trained second AI unit of the terminal.

[0486] Correspondingly, the terminal establishes a matching relationship between the trained second AI unit and the decoder of the network side device.

[0487] The above process is the training process, and after the training is completed, the training results can be used for actual CSI feedback.

[0488] Step 5: In the CSI feedback scenario, the terminal uses the trained second AI unit to encode the CSI and then sends it to the network side device. Correspondingly, the network side device can use the decoder to decode the received encoded CSI to obtain the recovered CSI.

[0489] The CSI feedback scenario refers to the scenario in which the terminal feeds back the CSI to the network side device. There can be multiple scenarios, including but not limited to at least one of the following: a scenario in which the cell edge needs to feed back the CSI, a scenario in which the bandwidth allocated for CSI feedback is small, a scenario in which the feedback channel is a fast-varying channel and accurate channel estimation cannot be performed, and the like.

[0490] The above process can consider the influence of the channel in the AI-based encoding process, and realize data collection based on AI-based joint source channel coding. The enhanced two-sided separation training based on data collection can improve the practicality of the training model, making the model more widely applicable. The CSI feedback application based on the training model can bring gain to the CSI, improve the accuracy of the CSI feedback, and improve the transmission efficiency.

[0491] FIG. 6 shows a structure of a data collection device for joint source channel coding according to an embodiment of the present application. As shown in FIG. 6, the device 600 is applied to a first device and can include a processing module 601 and a transmission module 602.

[0492] The processing module 601 is configured to input first information into a first AI unit to obtain second information.

[0493] The transmission module 602 is configured to transmit the second information to a second device using a preset manner.

[0494] The first information includes channel state information (CSI), the first AI unit is configured to encode the first information, and the preset manner is configured to add channel influence.

[0495] In an embodiment of the present application, the preset manner includes a lossless transmission manner through a channel, a lossless transmission manner, and a lossy transmission manner. The transmission module 602 transmits the second information to the second device using the preset manner, which can include one of the following:

[0496] After the second information is transmitted to the second device using the lossless transmission manner through the channel corresponding to the preset channel information;

[0497] The second information is transmitted to the second device using the lossless transmission manner;

[0498] The second information is transmitted to the second device using the lossy transmission manner.

[0499] In the embodiments of the present application, the device is further configured to receive gradient information of second information sent by the second device, and train the first AI unit according to the gradient information of the second information.

[0500] The gradient information of the second information is obtained by the second device according to training of a second AI unit on the second information, and the second AI unit is configured to decode the second information.

[0501] In an implementation, the training of the first AI unit according to the gradient information of the second information can include training the first AI unit according to training parameter information and the gradient information of the second information.

[0502] The training parameter information includes parameter information related to the training of the first AI unit.

[0503] In the embodiments of the present application, the CSI can include at least one of the following parameters:

[0504] Channel Quality Indication (CQI);

[0505] Precoding Matrix Indication (PMI);

[0506] Channel State Information Reference Signal Resource Indicator (CRI);

[0507] Synchronization Signal Physical Broadcast Channel Block Resource Indicator (SSBRI);

[0508] Layer Indicator (LI);

[0509] Rank Indicator (RI);

[0510] Layer 1 Reference Signal Received Power (L1-RSRP);

[0511] Layer 1 Signal to Interference plus Noise Ratio (L1-SINR);

[0512] Capacity Index;

[0513] Time Difference Carrier Phase (TDCP);

[0514] Raw Channel Information;

[0515] Transform Domain Channel Information;

[0516] Processed Raw Channel Information;

[0517] Source corresponding to the PMI.

[0518] In the embodiments of the present application, any one of the parameters included in the CSI can be replaced by at least one of the mean, variance, deviation, and product factor of the parameter, without limitation.

[0519] In an embodiment of the present application, the second information can include at least one of a bit sequence and a symbol sequence. The bit sequence corresponds to a bit-level joint source-channel coding, and the symbol sequence corresponds to a symbol-level joint source-channel coding.

[0520] In an embodiment, the second information can include a bit sequence, and the transmission module 602 can be further configured to modulate the bit sequence in the second information using at least one of the following modulation modes before transmitting the second information to the second device using the lossy transmission:

[0521] binary phase shift keying (BPSK);

[0522] pi / 2-BPSK;

[0523] quadrature phase shift keying (QPSK);

[0524] 16 quadrature amplitude modulation (QAM);

[0525] 64 QAM;

[0526] 256 QAM;

[0527] 1024 QAM;

[0528] 4096 QAM;

[0529] AI-based modulation.

[0530] In an embodiment of the present application, the apparatus is further configured to transmit the first related information to the second device using the lossy transmission or the lossless transmission.

[0531] The first related information is information related to the first information and the encoding and decoding.

[0532] In an embodiment of the present application, the first related information can include at least one of the following:

[0533] a sample in the CSI dataset;

[0534] a sample ID in the CSI dataset;

[0535] dimension information after encoding;

[0536] channel impact information;

[0537] an acquisition rule of the first information.

[0538] In an embodiment of the present application, the acquisition rule of the first information can include an iteration number index, and the iteration number index is used to indicate that a preset number of samples are acquired from a current position in the samples arranged in a preset order. The current position is an initial position or a position where the last sample acquisition ends.

[0539] In the embodiments of the present application, the first information can be obtained in at least one of the following ways:

[0540] obtaining the first information predefined by a protocol;

[0541] collecting the first information;

[0542] receiving the first information sent by a second device;

[0543] receiving the first information sent by a server of the first device;

[0544] receiving the first information sent by a third-party server.

[0545] In the embodiments of the present application, the device is further configured to send preset channel information to the second device by using a lossy transmission or a lossless transmission.

[0546] The preset channel information is used to add channel influence to the second information by using a corresponding channel.

[0547] In the embodiments of the present application, the preset channel information can include at least one of the following:

[0548] a signal-to-noise ratio (SNR);

[0549] an SNR range;

[0550] a bit error rate;

[0551] a bit error rate range;

[0552] a channel quality indicator (CQI);

[0553] a CQI range;

[0554] a modulation and coding strategy (MCS);

[0555] an MCS range;

[0556] raw channel information of a transmission channel;

[0557] a channel model.

[0558] In the embodiments of the present application, the preset channel information can be predefined by a protocol, or indicated by the first device or the second device in at least one of the following ways:

[0559] wireless resource control (RRC) indication;

[0560] media access control control element (MAC CE) indication;

[0561] control information indication.

[0562] In the embodiments of the present application, the lossy transmission can include at least one of the following:

[0563] Layer mapping;

[0564] Interleaving;

[0565] Padding;

[0566] Puncturing;

[0567] Rate matching;

[0568] Scrambling;

[0569] Precoding.

[0570] In the embodiments of the present application, the lossless transmission mode can include at least one of the following:

[0571] Source coding;

[0572] Channel coding;

[0573] Modulation;

[0574] Layer mapping;

[0575] Interleaving;

[0576] Padding;

[0577] Puncturing;

[0578] Rate matching;

[0579] Scrambling;

[0580] Precoding.

[0581] In the embodiments of the present application, the training parameter information can include at least one of the following:

[0582] Batch size;

[0583] Optimizer type;

[0584] Learning rate;

[0585] Learning rate decay;

[0586] Loss function;

[0587] Regularization method and parameter;

[0588] Initialization method and parameter;

[0589] Dropout method and parameter;

[0590] Training period;

[0591] Iteration number;

[0592] Training termination condition.

[0593] In the embodiments of the present application, the loss function can include:

[0594] a combination function of at least one of an error of an output of the target model and a label, a mean square error, a normalized mean square error, a correlation, an entropy, mutual information, a constant.

[0595] The target model is the first AI unit or the second AI unit.

[0596] In an implementation, the loss function is obtained by weighted combination of loss information of multiple parts of the output of the target model. The multiple parts are obtained by division according to at least one of a spatial resource, a code resource, a frequency resource, and a time resource. The frequency resource includes a resource block (RB), a sub-band, or a precoding resource block group (PRG). The time resource includes a sub-carrier, a symbol, a time slot, or a half time slot. The loss information includes a loss value and / or a loss-related function.

[0597] In the embodiments of the present application, the training parameter information can be preset by a protocol, or indicated by the first device or the second device using at least one of the following ways:

[0598] RRC indication;

[0599] MAC CE indication;

[0600] Control information indication.

[0601] In the embodiments of the present application, the second information can be carried in one of the following:

[0602] DCI;

[0603] MAC CE;

[0604] RRC;

[0605] Physical uplink shared channel (PUSCH);

[0606] Physical downlink shared channel (PDSCH);

[0607] Physical downlink control channel (PDCCH);

[0608] Physical uplink control channel (PUCCH).

[0609] In the embodiments of the present application, the second information can be configured by RRC or MAC CE to be periodically transmitted, semi-persistently transmitted, or non-periodically transmitted. The semi-persistent transmission is activated by MAC CE or DCI, and is periodically transmitted after activation until deactivation or after a specified period. The non-periodic transmission is activated by MAC CE or DCI, and is transmitted for a specified number of times.

[0610] In the embodiments of the present application, the gradient information of the second information can be carried in one of the following:

[0611] DCI;

[0612] MAC CE;

[0613] RRC;

[0614] PUSCH;

[0615] PDSCH;

[0616] PDCCH;

[0617] PUCCH.

[0618] In the embodiments of the present application, the gradient information of the second information can be configured by RRC or MAC CE to be periodically transmitted, semi-persistently transmitted or non-periodically transmitted. The semi-persistent transmission is activated by MAC CE or DCI, and is periodically transmitted after being activated until being deactivated or after a specified period. The non-periodic transmission is activated by MAC CE or DCI, and is transmitted for a specified number of times.

[0619] In the embodiments of the present application, the gradient information of the second information can be gradient information of the second information greater than or equal to a preset first threshold.

[0620] In the embodiments of the present application, the device is further configured to: establish a matching relationship between the trained first AI unit and the second AI unit.

[0621] In the embodiments of the present application, the identifier of the matching relationship can include at least one of:

[0622] model ID;

[0623] function ID;

[0624] data set ID;

[0625] pairing ID;

[0626] association ID.

[0627] In the embodiments of the present application, the training of the first AI unit or the training of the second AI unit can be triggered by one of:

[0628] network indication;

[0629] terminal reporting;

[0630] predefined event.

[0631] In the embodiments of the present application, obtaining the second information can be triggered by one of:

[0632] network indication;

[0633] terminal reporting;

[0634] predefined event.

[0635] In the embodiments of the present application, the acquisition of the first related information can be triggered by one of the following:

[0636] Network indication;

[0637] Terminal reporting;

[0638] Predefined event.

[0639] In the embodiments of the present application, the predefined event can include at least one of the following:

[0640] In a wifi state;

[0641] The power exceeds a second threshold;

[0642] In a charging state;

[0643] Supporting a preset traffic mode;

[0644] In a preset traffic mode;

[0645] No service demand;

[0646] The performance index is lower than a third threshold.

[0647] In the embodiments of the present application, the second information can be used for the second device to train the second AI unit.

[0648] In the embodiments of the present application, the first AI unit and the second AI unit can be aligned with the third information before training, and the third information includes at least one of the following: structural features of the model, load quantization method of the model, estimation accuracy of the model, and output accuracy of the model.

[0649] The above device provided by the embodiments of the present application can execute the method in any of the above method embodiments with the first device as the execution subject, and the detailed process is described in the method embodiments, which will not be repeated here.

[0650] The above device provided by the embodiments of the present application, by the first device, inputs the first information into the first AI unit to obtain the second information, and uses a preset manner to transmit the second information to the second device, wherein the first information is CSI, the first AI unit is used for encoding the first information, and the preset manner is used to add channel influence, so that the second device can obtain the second information with added channel influence, and can consider the influence of the channel in the AI-based encoding process, realizing the data acquisition of the AI-based joint source channel coding, helping the training of CSI feedback enhancement, and being able to improve the practicality of the training model, making the model more widely applied.

[0651] FIG. 7 shows a structure schematic diagram of a data acquisition device for joint source channel coding provided by the embodiments of the present application, as shown in FIG. 7, the device 700 applied to the second device can include a receiving module 701.

[0652] The receiving module 701 is configured to receive second information sent by the first device using a preset manner.

[0653] The preset manner is used to add channel influence, the second information is obtained by inputting the first information into the first AI unit by the first device, the first information includes CSI, the first AI unit is used to encode the first information, and the second information includes the encoded CSI.

[0654] In the embodiments of the present application, the apparatus is further configured to train the second AI unit based on the second information.

[0655] The apparatus trains the second AI unit based on the second information, and can include the following:

[0656] The second information is input into the second AI unit for decoding to obtain recovered first information;

[0657] The second AI unit is trained based on the training parameter information, the first related information and the recovered first information to obtain gradient information of the second information.

[0658] In the embodiments of the present application, the receiving module 701 receives the second information sent by the first device using a preset manner, which can include one of the following:

[0659] The second information sent by the first device through a corresponding channel after the preset channel information is received using a lossless transmission manner;

[0660] The second information sent by the first device using a lossless transmission manner is received;

[0661] The second information sent by the first device using a lossy transmission manner is received.

[0662] In the embodiments of the present application, the apparatus can also be configured to train the second AI unit after receiving the second information sent by the first device using a lossless transmission manner, and then transmitting the second information through a corresponding channel according to preset channel information.

[0663] In the embodiments of the present application, the apparatus can also be configured to send the gradient information of the second information to the first device.

[0664] The gradient information of the second information is used by the first device to train the first AI unit in combination with the training parameter information.

[0665] In the embodiments of the present application, the apparatus can also be configured to receive the first related information sent by the first device using a lossy transmission or a lossless transmission manner.

[0666] The first related information is information associated with the first information and the encoding and decoding.

[0667] In the embodiments of the present application, the apparatus can also be configured to receive preset channel information sent by the first device using a lossy transmission or a lossless transmission.

[0668] The preset channel information is used to add channel influence to the second information using a corresponding channel.

[0669] In the embodiments of the present application, the preset manner is a lossy transmission, and after the receiving module 701 receives the second information sent by the first device using the preset manner, the receiving module 701 further performs at least one of the following operations on the second information:

[0670] demapping;

[0671] deinterleaving;

[0672] depadding;

[0673] descrambling;

[0674] receiving and combining.

[0675] In the embodiments of the present application, the preset manner includes a lossless transmission, and after the receiving module 701 receives the second information sent by the first device using the preset manner, the receiving module 701 further performs at least one of the following operations on the second information:

[0676] source decoding;

[0677] channel decoding;

[0678] demodulation;

[0679] demapping;

[0680] deinterleaving;

[0681] depadding;

[0682] descrambling;

[0683] receiving and combining.

[0684] The apparatus provided in the embodiments of the present application can execute the method in any method embodiment described above with the second device as the execution subject, and the detailed process is described in the method embodiments, which will not be repeated here.

[0685] The device provided by the embodiment of the present application can receive the second information sent by the first device in a preset manner through the second device, wherein the preset manner is used to add channel influence, the second information is obtained by inputting the first information into the first AI unit by the first device, the first information is CSI, and the first AI unit is used to encode the first information, so that the second device can obtain the second information with added channel influence and consider the influence of the channel in the AI-based encoding process, thereby realizing data acquisition of AI-based joint source channel coding, helping to enhance the training of CSI feedback, improving the practicality of the training model, and making the model more widely applied.

[0686] The embodiment of the present application provides a data acquisition device for joint source channel coding. As an example, the device can be a communication device or a component in a communication device, such as a chip. The communication device can be a terminal, a network side device, a server, or the like. For example, the terminal can include but is not limited to the types of terminals listed above, the network side device can include but is not limited to the types of network side devices listed above, and the embodiment of the present application does not make specific limitations.

[0687] The data acquisition device for joint source channel coding includes a receiving module, a sending module, and a processing module. The receiving module, the sending module, and the processing module can be implemented by software or by hardware. When implemented by hardware, the processing module can be implemented by a processor, which can include a general-purpose processor, a special-purpose processor, or the like, such as a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), an artificial intelligent (AI) processor, a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a network processor (NP), a field programmable gate array (FPGA), or other programmable logic devices, gate circuits, transistors, discrete hardware components, or the like. The receiving module and the sending module can be implemented by a communication interface, which can include one or more of a transceiver, a pin, a circuit, a bus, a radio frequency unit, or the like.

[0688] The data acquisition device for joint source channel coding provided by the embodiment of the present application can implement the processes of the method embodiments of FIGS. 2 to 5 and achieve the same technical effects. To avoid repetition, details are not described here.

[0689] As shown in FIG. 8, the embodiment of the present application further provides a communication device 800, comprising a processor 801 and a memory 802, wherein the memory 802 stores programs or instructions executable by the processor 801. For example, when the communication device 800 is a terminal, the programs or instructions are executed by the processor 801 to implement the steps of the above method embodiments and achieve the same technical effects. When the communication device 800 is a network side device, the programs or instructions are executed by the processor 801 to implement the steps of the above method embodiments and achieve the same technical effects. To avoid repetition, details are not described herein.

[0690] The embodiment of the present application further provides a terminal, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is configured to run programs or instructions to implement the steps in the above method embodiments. The terminal embodiment corresponds to the above terminal side method embodiments, and each implementation process and implementation manner of the above method embodiments can be applied to the terminal embodiment and achieve the same technical effects. The terminal can be the apparatus shown in FIG. 6 or 7. Specifically, FIG. 9 is a schematic diagram of a hardware structure of a terminal for implementing the embodiment of the present application.

[0691] The terminal 900 includes, but is not limited to, at least part of the following components: a radio frequency unit 901, a network module 902, an audio output unit 903, an input unit 904, a sensor 905, a display unit 906, a user input unit 907, an interface unit 908, a memory 909, and a processor 910.

[0692] Those skilled in the art can understand that the terminal 900 can further include a power supply (such as a battery) for supplying power to each component, and the power supply can be logically connected to the processor 910 through a power management system, so as to realize the functions of power management, such as charging, discharging and power consumption management, through the power management system. The terminal structure shown in FIG. 9 does not constitute a limitation on the terminal, and the terminal can include more or fewer components than those shown, or combine certain components, or different component arrangements, which are not described herein.

[0693] It should be understood that in the embodiments of the present application, the input unit 904 can include a graphics processor 9041 and a microphone 9042, and the graphics processor 9041 processes image data of a still picture or a video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The display unit 906 can include a display panel 9061, which can be configured in the form of a liquid crystal display, an organic light-emitting diode, etc. The user input unit 907 includes at least one of a touch panel 9071 and other input devices 9072. The touch panel 9071 is also called a touch screen. The touch panel 9071 can include two parts of a touch detection device and a touch controller. The other input devices 9072 can include, but are not limited to, a physical keyboard, function keys (such as volume control keys, on-off keys, etc.), a trackball, a mouse, a joystick, and the like, which will not be described here.

[0694] In the embodiments of the present application, after the radio frequency unit 901 receives the downlink data from the network side device, it can be transmitted to the processor 910 for processing. In addition, the radio frequency unit 901 can send uplink data to the network side device. Generally, the radio frequency unit 901 includes, but is not limited to, an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.

[0695] The memory 909 can be used to store software programs or instructions and various data. The memory 909 can mainly include a first storage area storing programs or instructions and a second storage area storing data, wherein the first storage area can store an operating system, application programs or instructions required by at least one function (such as a sound playing function, an image playing function, etc.), etc. In addition, the memory 909 can include a volatile memory or a non-volatile memory. The non-volatile memory can be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically EPROM (EEPROM), or a flash memory. The volatile memory can be a Random Access Memory (RAM), a Static RAM (SRAM), a Dynamic RAM (DRAM), a Synchronous DRAM (SDRAM), a Double Data Rate SDRAM (DDR SDRAM), an Enhanced SDRAM (ESDRAM), a Synch link DRAM (SLDRAM), and a Direct Rambus RAM (DRRAM). The memory 909 in the embodiments of the present application includes but is not limited to these and any other suitable types of memory.

[0696] The processor 910 can include one or more processing units; optionally, the processor 910 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and an application program, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 910.

[0697] The processor 910 is configured to input first information into a first artificial intelligence (AI) unit to obtain second information. The radio frequency unit 901 is configured to transmit the second information to a second device using a preset manner. The preset manner is used to add channel influence. The second information is obtained by the first device inputting the first information into the first AI unit. The first information includes CSI. The first AI unit is configured to encode the first information.

[0698] Or, the radio frequency unit 901 is configured to receive second information sent by the first device using a preset manner. The preset manner is used to add channel influence. The second information is obtained by inputting the first information into the first AI unit. The first information includes CSI. The first AI unit is configured to encode the first information.

[0699] The terminal provided in the embodiments of the present application obtains the second information by inputting the first information into the first AI unit and transmits the second information to the second device using the preset manner. The first information is CSI. The first AI unit is configured to encode the first information. The preset manner is used to add channel influence. The second device can obtain the second information with added channel influence and can consider the influence of the channel in the AI-based encoding process. The data acquisition based on the AI-based joint source channel encoding is achieved. The training of the CSI feedback enhancement is facilitated. The practicability of the training model is improved. The model application is more extensive.

[0700] It can be understood that the implementation processes of the implementation manners mentioned in the embodiments can refer to the related descriptions of the method embodiments and achieve the same or corresponding technical effects. To avoid repetition, details are not described herein again.

[0701] The embodiments of the present application further provide a network side device including a processor and a communication interface. The communication interface and the processor are coupled. The processor is configured to run programs or instructions to implement the steps of the above method embodiments. The network side device embodiments correspond to the above network side device method embodiments. The implementation processes and implementation manners of the above method embodiments can be applied to the network side device embodiments and achieve the same technical effects.

[0702] Specifically, the embodiments of the present application further provide a network side device, which can be the apparatus shown in FIG. 6 or 7. As shown in FIG. 10, the network side device 1000 includes an antenna 101, a radio frequency device 102, a baseband device 103, a processor 104 and a memory 105. The antenna 101 is connected with the radio frequency device 102. In the uplink direction, the radio frequency device 102 receives information through the antenna 101 and sends the received information to the baseband device 103 for processing. In the downlink direction, the baseband device 103 processes the information to be sent and sends it to the radio frequency device 102. The radio frequency device 102 processes the received information and sends it out through the antenna 101.

[0703] The method performed by the network side device in the above embodiments can be implemented in the baseband device 103. The baseband device 103 includes a baseband processor.

[0704] The baseband device 103 can include at least one baseband board on which a plurality of chips are disposed, as shown in FIG. 10, one of the chips being, for example, a baseband processor, connected with the memory 105 through a bus interface to invoke programs in the memory 105 to perform the network device operations shown in the above method embodiments.

[0705] The network side device can further include a network interface 106, which is, for example, a Common Public Radio Interface (CPRI).

[0706] Specifically, the network side device 1000 of the embodiments of the present application further includes instructions or programs stored in the memory 105 and executable on the processor 104, the processor 104 invoking the instructions or programs in the memory 105 to perform the methods performed by the modules shown in FIG. 6 or 7 and achieve the same technical effects, and thus the details are not repeated here.

[0707] The embodiments of the present application further provide a readable storage medium having programs or instructions stored thereon, the programs or instructions being executed by a processor to implement the various processes of the above method embodiments and achieve the same technical effects, and thus the details are not repeated here.

[0708] The processor is the processor in the terminal described in the above embodiments. The readable storage medium includes a computer readable storage medium, such as a computer readable only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. In some examples, the readable storage medium can be a non-transitory readable storage medium.

[0709] The embodiments of the present application further provide a chip including a processor and a communication interface, the communication interface being coupled with the processor, the processor being configured to run programs or instructions to implement the various processes of the above method embodiments and achieve the same technical effects, and thus the details are not repeated here.

[0710] It should be understood that the chip mentioned in the embodiments of the present application can also be referred to as a system level chip, a system chip, a chip system or a system on chip, etc.

[0711] The embodiments of the present application further provide a computer program / program product stored in a storage medium, the computer program / program product being executed by at least one processor to implement the various processes of the above method embodiments and achieve the same technical effects, and thus the details are not repeated here.

[0712] This application also provides a data acquisition system using joint source-channel coding, comprising: a terminal and a network-side device. The terminal can be used to execute the steps of the first device-side joint source-channel coding data acquisition method described above, and the network-side device can be used to execute the steps of the second device-side joint source-channel coding data acquisition method described above. Alternatively, the network-side device can be used to execute the steps of the first device-side joint source-channel coding data acquisition method described above, and the terminal can be used to execute the steps of the second device-side joint source-channel coding data acquisition method described above.

[0713] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0714] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.) and includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[0715] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

1. A data acquisition method of joint source channel coding, comprising: a first device inputs first information into a first artificial intelligence (AI) unit to obtain second information; transmitting the second information to a second device using a preset mode; wherein the first information comprises channel state information (CSI), the first AI unit is configured to encode the first information, and the preset mode is configured to add channel influence.

2. The method of claim 1, wherein, The preset mode comprises: a lossless transmission mode through a channel, a lossless transmission mode, and a lossy transmission mode.

3. The method of claim 1, wherein, Further comprising: receiving gradient information of the second information sent by the second device, wherein the gradient information of the second information is obtained by the second device according to training of a second AI unit on the second information, and the second AI unit is configured to decode the second information; training the first AI unit according to the gradient information of the second information.

4. The method of any one of claims 1-3, wherein, The CSI comprises at least one of the following parameters: channel quality indicator (CQI); precoding matrix indicator (PMI); channel state information reference signal resource indicator (CRI); synchronization signal physical broadcast channel block resource indicator (SSBRI); layer indicator (LI); rank indicator (RI); layer 1 reference signal received power (L1-RSRP); layer 1 signal to interference plus noise ratio (L1-SINR); capacity index; time difference carrier phase (TDCP); raw channel information; transform domain channel information; processed raw channel information; a source corresponding to the PMI.

5. The method of claim 1, wherein, The second information comprises at least one of a bit sequence and a symbol sequence; wherein the bit sequence corresponds to bit-level joint source channel coding, and the symbol sequence corresponds to symbol-level joint source channel coding.

6. The method of claim 2, wherein, The second information comprises a bit sequence, and before the second information is transmitted to the second device using the lossy transmission mode, the method further comprises modulating the bit sequence in the second information using at least one of the following modulation modes: binary phase shift keying (BPSK); pi / 2-BPSK; quadrature phase shift keying (QPSK); 16 quadrature amplitude modulation (QAM); 64 QAM; 256 QAM; 1024 QAM; 4096 QAM; AI-based modulation.

7. The method of any one of claims 1-6, wherein, Further comprising: transmitting first related information to the second device using a lossy transmission or lossless transmission mode; wherein the first related information is information related to the first information and encoding and decoding; The first related information comprises at least one of the following: a sample in a CSI data set; a sample ID in a CSI data set; dimension information after encoding; channel influence information; acquisition rule of the first information.

8. The method of claim 7, wherein, The acquisition rule comprises an iteration number index, and the iteration number index is used to indicate that a preset number of samples are obtained from a current position in samples arranged in a preset order, and the current position is an initial position or a position where the last sample acquisition ends.

9. The method of any one of claims 1-8, wherein, The first device is a terminal or a network side device, and the first information is obtained in at least one of the following ways: acquiring first information predefined by a protocol; acquiring first information; receiving first information sent by the second device; receiving first information sent by a server of the first device, wherein the server of the first device is a server of a same network operator as the first device; receiving first information sent by a third-party server.

10. The method of claim 1, wherein, Further comprising: sending preset channel information to the second device in a lossy transmission manner or a lossless transmission manner, wherein the preset channel information is used to add channel influence on the second information using a corresponding channel; the preset channel information comprises at least one of: a signal-to-noise ratio (SNR); an SNR range; a bit error rate (BER); a BER range; a channel quality indicator (CQI); a CQI range; a modulation and coding strategy (MCS); an MCS range; raw channel information of a transmission channel; and a channel model.

11. The method of claim 2 or 10, wherein, The preset channel information is preset by a protocol or indicated by the first device or the second device in at least one of the following manners: a radio resource control (RRC) indication; a medium access control control element (MAC CE) indication; and control information indication.

12. The method of claim 2, 7, or 10, wherein, The lossy transmission manner comprises at least one of: layer mapping; interleaving; padding; puncturing; rate matching; scrambling; and precoding.

13. The method of claim 2, 7, or 10, wherein, The lossless transmission manner comprises at least one of: source coding; channel coding; modulation; layer mapping; interleaving; padding; puncturing; rate matching; scrambling; and precoding.

14. The method of claim 3, wherein, The training parameter information of the trained first AI unit is preset by a protocol or indicated by the first device or the second device in at least one of the following manners: an RRC indication; a MAC CE indication; and control information indication.

15. The method of claim 1, wherein, The second information is carried in one of the following: a DCI; a MAC CE; an RRC; a physical uplink shared channel (PUSCH); a physical downlink shared channel (PDSCH); a physical downlink control channel (PDCCH); and a physical uplink control channel (PUCCH).

16. The method of claim 1, wherein, The second information is configured by an RRC or a MAC CE to be periodically transmitted, semi-persistently transmitted, or aperiodically transmitted; the semi-persistent transmission is activated by a MAC CE or a DCI and periodically transmitted after being activated until being deactivated or after a specified period; the aperiodic transmission is activated by a MAC CE or a DCI and transmitted a specified number of times.

17. The method of claim 3, wherein, Gradient information of the second information is carried in one of the following: a DCI; a MAC CE; an RRC; a PUSCH; a PDSCH; a PDCCH; and a PUCCH.

18. The method of claim 3, wherein, Gradient information of the second information is configured by an RRC or a MAC CE to be periodically transmitted, semi-persistently transmitted, or aperiodically transmitted; the semi-persistent transmission is activated by a MAC CE or a DCI and periodically transmitted after being activated until being deactivated or after a specified period; the aperiodic transmission is activated by a MAC CE or a DCI and transmitted a specified number of times.

19. The method of claim 3, wherein, Gradient information of the second information is gradient information of the second information greater than or equal to a preset first threshold.

20. The method of claim 3, wherein, Further comprising: establishing a matching relationship between the trained first AI unit and a second AI unit.

21. The method of claim 20, wherein, An identifier of the matching relationship comprises at least one of: a model ID; a function ID; a data set ID; a pairing ID; and an association ID.

22. The method of any one of claims 3, 17-21, wherein, Training of the first AI unit or training of the second AI unit is triggered by one of the following: a network indication; terminal reporting; and Predefined event.

23. The method of claim 1, wherein, The obtaining of the second information is triggered by one of the following: Network indication; Terminal reporting; Predefined event.

24. The method of claim 7, wherein, The obtaining of the first related information is triggered by one of the following: Network indication; Terminal reporting; Predefined event.

25. The method of claim 22, 23, or 24, wherein, The predefined event includes at least one of the following: In a wifi state; Battery power exceeding a second threshold; In a charging state; Supporting a preset traffic mode; In a preset traffic mode; No service demand; Performance index below a third threshold.

26. The method of any one of claims 1-25, wherein, The second information is used for training the second AI unit by the second device.

27. The method of claim 3, wherein, The first AI unit and the second AI unit are aligned with third information before training, and the third information includes at least one of the following: structural features of a model, load quantization method of a model, estimation accuracy of a model, and output accuracy of a model.

28. A data acquisition method combining source channel coding, comprising: The second device receives second information sent by the first device using a preset mode; The preset mode is used to add channel influence, the second information is obtained by inputting first information into a first AI unit by the first device, the first information includes CSI, and the first AI unit is used to code the first information.

29. The method of claim 28, wherein, Further comprising: Inputting the second information into a second AI unit for decoding to obtain recovered first information; According to the training parameter information, the first related information, and the recovered first information, training the second AI unit to obtain gradient information of the second information; The gradient information of the second information is sent to the first device, and the gradient information of the second information is used for the first device to train the first AI unit in combination with the training parameter information.

30. The method of claim 28, wherein, The preset mode includes: a lossless transmission mode through a channel, a lossless transmission mode, and a lossy transmission mode, and the second device receives second information sent by the first device using a preset mode, including one of the following: The second device receives second information sent by the first device using a lossless transmission mode through a corresponding channel according to preset channel information; The second device receives second information sent by the first device using a lossless transmission mode, and transmits the second information through a corresponding channel according to preset channel information; The second device receives second information sent by the first device using a lossy transmission mode.

31. The method of any one of claims 28-30, wherein, Further comprising: Receiving first related information sent by the first device using a lossy transmission or a lossless transmission mode; wherein the first related information is information related to the first information and coding and decoding; The first related information includes at least one of the following: Sample in CSI data set; Sample ID in CSI data set; Encoded dimension information; Channel influence information; First information acquisition rule.

32. The method of any one of claims 28-30, wherein, Further comprising: Receiving preset channel information sent by the first device using a lossy transmission or a lossless transmission mode; wherein the preset channel information is used to add channel influence to the second information using a corresponding channel; The preset channel information includes at least one of the following: Signal-to-noise ratio (SNR); SNR range; Bit error rate; Bit error rate range; CQI; CQI range; Modulation and coding strategy (MCS); MCS range; Raw channel information of a transmission channel; Channel model.

33. The method of any one of claims 30-32, wherein, The preset mode is a lossy transmission mode, and after receiving the second information sent by the first device using the preset mode, the second device further performs at least one of the following operations on the second information: demapping; deinterleaving; depadding; descrambling; reception combining.

34. The method of any one of claims 30-32, wherein, The preset mode includes a lossless transmission mode, and after receiving the second information sent by the first device using the preset mode, the second device further performs at least one of the following operations on the second information: source decoding; channel decoding; demodulation; demapping; deinterleaving; depadding; descrambling; reception combining.

35. A data acquisition device for joint source-channel coding, comprising: a processing module configured to input first information into a first artificial intelligence (AI) unit to obtain second information; a transmission module configured to transmit the second information to a second device using a preset mode; wherein the first information includes channel state information (CSI), the first AI unit is configured to encode the first information, and the preset mode is configured to add channel influence.

36. A data acquisition device for joint source-channel coding, comprising: a receiving module configured to receive second information sent by a first device using a preset mode; wherein the preset mode is configured to add channel influence, the second information is obtained by inputting first information into a first AI unit by the first device, the first information includes CSI, and the first AI unit is configured to encode the first information.

37. A terminal comprising a processor and a memory, wherein the memory stores programs or instructions executable on the processor, and the programs or instructions are executed by the processor to implement the steps of the data acquisition method for joint source-channel coding according to any one of claims 1-34.

38. A network-side device comprising a processor and a memory, wherein the memory stores programs or instructions executable on the processor, and the programs or instructions are executed by the processor to implement the steps of the data acquisition method for joint source-channel coding according to any one of claims 1-34.

39. A readable storage medium, wherein the readable storage medium stores programs or instructions, and the programs or instructions are executed by a processor to implement the steps of the data acquisition method for joint source-channel coding according to any one of claims 1-34.

Images