Communication method and apparatus

Abstract

The present disclosure relates to a communication method and apparatus. A terminal device obtains M channel estimation results corresponding to M time units, where M is greater than 1. The terminal device performs joint compression on the M channel estimation results by using an encoder network to obtain N pieces of compressed information. The terminal device transmits the N pieces of compressed information to an access network device. Since the terminal device performs joint compression, correlation between downlink channels at different time points is fully utilized. The access network device restores the N pieces of compressed information obtained by joint compression, so that more accurate and effective downlink channel information can be obtained.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to Chinese Patent Application No. 202111590532.1, entitled "Communication Method and Apparatus," filed with the China State Intellectual Property Office on December 23, 2021, the entire contents of which are incorporated herein by reference.

[0002] The present application relates to the field of communication technologies, and in particular to communication methods and devices. [Background technology]

[0003] The 5th generation (5G) mobile communication system has high requirements for system capacity, spectral efficiency, etc. In the 5G communication system, the application of massive multiple-input multiple-output (massive-MIMO) technology plays an important role in improving the spectral efficiency of the system. By using massive-MIMO technology, the base station can provide high-quality services to more user equipment (UE). The key step is for the base station to perform precoding on the downlink data. Precoding can implement spatial multiplexing, reduce interference between different data streams, and increase the signal to interference plus noise ratio (SINR) at the receiving side. This helps to increase the system throughput rate. In order to perform precoding more accurately on the downlink data of the UE, the base station can obtain channel state information (CSI) of the downlink channel, restore the downlink channel information based on the CSI, and determine a precoding matrix by using the restored downlink channel information to perform precoding. How to enable a base station to recover more accurate downlink channel information is a technical problem worthy of consideration. Summary of the Invention

[0004] The present disclosure provides a communication method and apparatus for reducing the overhead of reporting high accuracy channel state information by terminal devices. [Means for solving the problem]

[0005] According to a first aspect, a first communication method is provided. The method may be performed at a terminal device side. The method may be performed by using software, hardware, or a combination of software and hardware. For example, the method may be performed by a terminal device, or by a circuit system, or by a larger device including the terminal device. The circuit system may implement the function of the terminal device. The method includes: obtaining M channel estimation results, where the M channel estimation results correspond to M time units, where M is an integer greater than 1; performing joint compression on the M channel estimation results to obtain N compressed information, where N is a positive integer; and transmitting the N compressed information to an access network device.

[0006] In the present disclosure, the terminal device may perform channel estimation on the downlink channel in M ​​time units, perform joint compression on the obtained M channel estimation results, and report the obtained M channel estimation results to the access network device. Because the joint compression is performed, the correlation between the downlink channels at different time points is fully utilized, and the access network device may restore the N pieces of compressed information obtained by the joint compression, so as to obtain more accurate and effective downlink channel information. Also, because the joint compression is performed, the N pieces of compressed information may be complemented with each other. For example, different pieces of compressed information may correspond to different parameters, so as to reduce redundant information and correspondingly reduce transmission overhead.

[0007] In an optional implementation, performing joint compression on the M channel estimation results to obtain N pieces of compressed information may include performing joint compression on the M channel estimation results by using an encoder network to obtain N pieces of compressed information. The terminal device may process the channel estimation results by using an encoder network, so that the access network device may restore the compressed information based on a corresponding decoder network. Compared with the conventional solution, the same size of feedback may include more channel information by using a neural network to reduce the information loss of the feedback in compression and improve the downlink channel restoration accuracy on the access network device side. Alternatively, compared with the conventional solution, the same channel information may be represented by a smaller amount of feedback to further reduce the feedback overhead.

[0008] In an optional implementation, the multiple channel estimation results include a first channel estimation result, and the first channel estimation result is one of the following: a channel estimation result obtained by performing a measurement based on a received downlink reference signal, a processing result obtained by processing a channel estimation result obtained by performing a measurement based on a received downlink reference signal, or a channel estimation result obtained by prediction. Some time units of the M time units may be past time units or current time units. For a time unit, the terminal device may perform a measurement based on a received downlink reference signal to directly obtain a channel estimation result. However, some time units in the M time units may not arrive. If the terminal device obtains a channel estimation result by performing a measurement after the time unit arrives, the previously obtained channel estimation result may be invalid due to an extremely long latency. Alternatively, in the M time units, the terminal device measures only the downlink reference signal in some time units. Thus, the terminal device may process the obtained channel estimation result to obtain a channel estimation result of a future time unit or an unmeasured time unit, or the terminal device may obtain a channel estimation result of a future time unit or an unmeasured time unit through prediction. In this way, the terminal device can obtain more time-unit channel estimation results in a timely manner to reduce latency and improve channel information feedback efficiency. In addition, the existing channel estimation results can be used as references, regardless of whether the future time-unit channel estimation results are obtained based on the existing channel estimation results or the future time-unit channel estimation results are obtained through prediction. However, because the downlink channel is correlated in the time domain, the future channel estimation results of the downlink channel obtained based on the previous channel estimation results of the downlink channel are more accurate. This helps the access network device to restore the accurate downlink channel.

[0009] In an optional implementation, each of the N pieces of compressed information corresponds to M channel estimation results. Since joint compression is performed on the M channel estimation results, each piece of compressed information can reflect the M channel estimation results, so that a more accurate downlink channel can be restored based on the N pieces of compressed information.

[0010] In an optional implementation, performing joint compression on the M channel estimation results by using an encoder network to obtain N pieces of compressed information includes performing joint compression on the M channel estimation results and history information by using an encoder network to obtain N pieces of compressed information, where the history information includes a channel estimation result corresponding to a time unit before M time units. The downlink channel is correlated in the time domain. Therefore, in addition to the M channel estimation results, the channel estimation result corresponding to a time unit before M time units can be further considered, so that the channel estimation result involved in the compression corresponds to more time units. In this way, the channel information represented by the compressed information obtained by the access network device becomes more diverse, and the access network device restores a more accurate downlink channel based on the channel estimation result in more time units.

[0011] In an optional implementation, each of the N pieces of compressed information corresponds to M channel estimation results and historical information. Since joint compression is performed on the M channel estimation results and historical information, each piece of compressed information can reflect the M channel estimation results and historical information, so that a more accurate downlink channel can be restored based on the N pieces of compressed information.

[0012] In an optional implementation, the method further includes receiving configuration information from the access network device, where the configuration information is used to configure the M time units. The M time units may be configured by the access network device, so that the access network device can configure, based on a request, a channel estimation result in the requested time unit fed back by the terminal device to facilitate a restoration operation of the access network device. Alternatively, the M time units may be determined by the terminal device and do not need to be configured by the access network device. In this way, signaling overhead can be reduced.

[0013] In an optional implementation, the configuration information includes one or more of the following: a start time domain position of the M time units, an end time domain position of the M time units, a duration of the M time units, a number of the M time units, a time domain position of a first sampling point in the M slots, a sampling period of the M time units, a quantity of sampling points of the M time units, or a time domain position of a sampling point in the M time units. In addition, the configuration information may further include other information of the M time units. The terminal device can determine the time domain positions of the sampling points of the M time units based on the configuration information, and can thereby perform channel estimation for the downlink channel of the sampling points.

[0014] In an optional implementation, the method further includes receiving a first indication information from the access network device, the first indication information indicating parameter information of a reference encoder network or an index of the reference encoder network, or sending a second indication information to the access network device, the second indication information indicating parameter information of a reference encoder network or an index of the reference encoder network. The reference encoder network is used to determine an encoder network. The reference encoder network may be indicated to the terminal device by the access network device, and the terminal device does not need to select a reference encoder network. The access network device can select a reference encoder network based on overall requirements, so that the selected reference encoder network is more appropriate. Alternatively, the reference encoder network may be selected by the terminal device instead, and does not need to be selected by the access network device, so that the workload of the access network device can be reduced.

[0015] According to a second aspect, a second communication method is provided. The method can be executed on the access network device side. The method can be executed by using software, hardware, or a combination of software and hardware. For example, the method is executed by an access network device, a large device including an access network device, or a circuit system. The circuit system can implement the function of the access network device. Alternatively, the method is executed by an AI module independent of the access network device with the assistance of the access network device or a network element of the access network device. This is not limited. For example, the access network device is an access network device, such as a base station. The method includes the steps of receiving N pieces of compressed information from a terminal device, where N is a positive integer, and restoring the N pieces of compressed information to obtain K pieces of restored information, where the K pieces of restored information are downlink channel information of M time units, where K is a positive integer, and M is an integer greater than 1.

[0016] In an optional implementation, the step of decompressing the N pieces of compressed information to obtain K pieces of reconstruction information includes the step of decompressing the N pieces of compressed information by using a decoder network to obtain the K pieces of reconstruction information.

[0017] In an optional implementation, each of the K pieces of reconstruction information corresponds to the N pieces of compression information.

[0018] In an optional implementation, the method further includes a step of transmitting configuration information to the terminal device, where the configuration information is used to configure the M time units.

[0019] In an optional implementation, the configuration information includes one or more of the following: a start time domain position of the M time units, an end time domain position of the M time units, a duration of the M time units, a number of the M time units, a time domain position of a first sampling point in the M slots, a sampling period of the M time units, a quantity of sampling points in the M time units, or a time domain position of a sampling point in the M time units.

[0020] In an optional implementation, the quantity of sampling points corresponding to the K pieces of restoration information is the same as or different from the quantity of sampling points corresponding to the N pieces of compression information, and / or the time domain positions of the sampling points corresponding to the K pieces of restoration information are the same as or different from the time domain positions of the sampling points corresponding to the N pieces of compression information. In other words, the time domain positions of the sampling points corresponding to the K pieces of restoration information are located within the M slots, and the quantity of sampling points corresponding to the K pieces of restoration information, the time domain positions within the M slots, etc. are not limited. In this way, for an access network device, the restoration process may be flexible.

[0021] In an optional implementation form, the method further includes: sending first indication information to the terminal device, the first indication information indicating parameter information of a reference encoder network or an index of the reference encoder network, or receiving second indication information from the terminal device, the second indication information indicating parameter information of a reference encoder network selected by the terminal device or an index of a reference encoder network selected by the terminal device. The reference encoder network and the reference decoder network belong to the same reference network, and the reference decoder network is used to determine the decoder network.

[0022] According to a third aspect, a communication device is provided. The communication device can implement the method according to the first aspect. The communication device has the functionality of a terminal device. In an optional implementation, the device can include one-to-one corresponding modules for performing the method / operation / step / action described in the first aspect. The modules may be hardware circuits or software, or may be implemented by hardware circuits in combination with software. In an optional implementation, the communication device includes a baseband device and a radio frequency device. In another optional implementation, the communication device includes a processing unit (sometimes also referred to as a processing module) and a transceiver unit (sometimes also referred to as a transceiver module). The transceiver unit can implement a transmitting function and a receiving function. When the transceiver unit implements a transmitting function, the transceiver unit may be referred to as a transmitting unit (sometimes also referred to as a transmitting module). When the transceiver unit implements a receiving function, the transceiver unit may be referred to as a receiving unit (sometimes also referred to as a receiving module). The transmitting unit and the receiving unit may be the same functional module, and the functional module is referred to as a transceiver unit. A functional module may implement a transmitting function and a receiving function, or the transmitting unit and the receiving unit may be different functional modules, and the transceiver unit is a generic term for the functional modules.

[0023] The processing unit is configured to obtain M channel estimation results and perform joint compression on the M channel estimation results to obtain N pieces of compressed information, where the M channel estimation results correspond to M time units, where M is an integer greater than 1, and N is a positive integer. The transceiver unit is configured to transmit the N pieces of compressed information to the access network device.

[0024] In another example, a communication device includes a processor coupled to a memory and configured to execute instructions in the memory to implement a method according to the first aspect. Optionally, the communication device further includes other components, such as an antenna, an input / output module, and an interface. The components may be hardware, software, or a combination of software and hardware.

[0025] According to a fourth aspect, a communication device is provided. The communication device can implement the method according to the second aspect. The communication device has a function of an access network device. The access network device is, for example, a base station or a baseband device in a base station. In an optional implementation, the device can include a one-to-one corresponding module for performing the method / operation / step / action described in the second aspect. The module may be a hardware circuit or software, or may be implemented by a hardware circuit in combination with software. In an optional implementation, the communication device includes a baseband device and a radio frequency device. In another optional implementation, the communication device includes a processing unit (sometimes also referred to as a processing module) and a transceiver unit (sometimes also referred to as a transceiver module). For the implementation of the transceiver unit, please refer to the related description of the third aspect.

[0026] The transceiver unit is configured to receive N pieces of compressed information from a terminal device, where N is a positive integer. The processing unit is configured to restore the N pieces of compressed information to obtain K pieces of restored information, where each of the K pieces of restored information is downlink channel information of M time units, where K is a positive integer and M is an integer greater than 1.

[0027] In another example, a communication device includes a processor coupled to a memory and configured to execute instructions in the memory to implement the method of the second aspect. Optionally, the communication device further includes other components, such as an antenna, an input / output module, and an interface. The components may be hardware, software, or a combination of software and hardware.

[0028] According to a fifth aspect, there is provided a computer readable storage medium configured to store computer programs or instructions which, when executed, implement the method according to the first aspect and / or the method in the second aspect.

[0029] According to a sixth aspect there is provided a computer program product comprising instructions which, when executed on a computer, implement the method according to the first aspect and / or the second aspect.

[0030] According to a seventh aspect, there is provided a chip system. The chip system includes a processor and may further include a memory, the chip system being configured to implement the method according to the first aspect and / or the second aspect. The chip system may include a chip or may include a separate component apart from the chip.

[0031] According to an eighth aspect, there is provided a communication system including a communication device according to the third aspect and a communication device according to the fourth aspect. [Brief description of the drawings]

[0032] [Figure 1] FIG. 1 is a diagram of a communication system. [Diagram 2] 1 is a flowchart of a CSI feedback mechanism. [Diagram 3] FIG. 1 is a diagram of the architecture of a communication network in a communication system. [Figure 4A]A diagram of an application framework for AI in a communications system. [Figure 4B] A diagram of an application framework for AI in a communications system. [Figure 4C] FIG. 1 is a diagram of another application framework for AI in a communications system. [Figure 4D] FIG. 1 is a diagram of yet another application framework for AI in a communications system. [Figure 4E] FIG. 1 is a diagram of yet another application framework for AI in a communications system. [Diagram 5] 1 is a flowchart of a communication method. [Figure 6A] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6B] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6C] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6D] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6E] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6F] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6G] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 6H] 1 is a diagram of several input and output scenarios of an encoder network. [Figure 7] FIG. 1 is a diagram of the CSI feedback process. [Figure 8A] 1A-1D are diagrams of several input and output scenarios of a decoder network. [Figure 8B] 1A-1D are diagrams of several input and output scenarios of a decoder network. [Figure 8C]1A-1D are diagrams of several input and output scenarios of a decoder network. [Figure 8D] 1A-1D are diagrams of several input and output scenarios of a decoder network. [Figure 9] FIG. 1 is a schematic block diagram of a communication device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0033] To make the objectives, technical solutions, and advantages of the present disclosure clearer, the present disclosure will be described in detail hereinafter with reference to the accompanying drawings.

[0034] The technology provided in this disclosure may be applied to a communication system 10 shown in FIG. 1. The communication system 10 includes one or more communication devices 30 (e.g., terminal devices). The one or more communication devices 30 are connected to one or more core network (CN) devices by using one or more access network (RAN) devices 20 to implement communication between the multiple communication devices. For example, the communication system 10 may be a communication system supporting the 4th generation (4G) (including long term evolution (LTE)) access technology, a communication system supporting 5G (sometimes referred to as new radio (NR)) access technology, a wireless fidelity (Wi-Fi) system, a cellular system related to the 3rd generation partnership project (3GPP), a communication system supporting the convergence of multiple wireless technologies, a future-oriented evolution system, etc. This is not limited thereto.

[0035] In the following, the terminal device and the RAN in FIG. 1 are described in detail separately.

[0036] 1. Terminal Device The terminal device may be simply called a terminal. The terminal device may be a device having a wireless transceiver function. The terminal device may be mobile or fixed. The terminal device may be deployed on land, including indoor, outdoor, handheld or vehicle-mounted deployment, may be deployed on water (e.g., on a ship), or may be deployed in the air (e.g., on an aircraft, a balloon, and a satellite). The terminal device may include a mobile phone, a tablet computer (pad), a computer having a wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal device of industrial control, a wireless terminal device of self driving, a wireless terminal device of remote medical surgery, a wireless terminal device of smart grid, a wireless terminal device of transportation safety, a wireless terminal device of smart city, and / or a wireless terminal device of smart home. Alternatively, the terminal device may also be a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device or a computing device with wireless communication capabilities, an in-vehicle device, a wearable device, a terminal device in a future 5th generation (5G) network, a terminal device in a further evolved public land mobile network (PLMN), etc. The terminal device may also be referred to as user equipment (UE). Optionally, the terminal device may communicate with multiple access network devices using different technologies.For example, a terminal device may communicate with an access network device supporting LTE, or may communicate with an access network device supporting 5G, or may implement dual connectivity to an access network device supporting LTE and an access network device supporting 5G, which is not limited in this disclosure.

[0037] In the present disclosure, the apparatus configured to implement the function of the terminal device may be a terminal device, or may be an apparatus capable of supporting the terminal device in implementing the function, such as a chip system, a hardware circuit, a software module, or a hardware circuit combined with a software module. The apparatus may be installed in the terminal device or may be matched with the terminal device for use. In the technical solution provided in the present disclosure, an example in which the apparatus configured to implement the function of the terminal device is a terminal device and the terminal device is a UE is used to explain the technical solution provided in the present disclosure.

[0038] In the present disclosure, a chip system may include a chip, or may include a chip and other separate components.

[0039] 2.RAN The RAN may include one or more RAN devices, such as the RAN device 20. The interface between the RAN device and the terminal device may be a Uu interface (or called an air interface). In future communications, the name of the interface may remain the same or may be replaced by other names. This is not limited in this application.

[0040] A RAN device is a node or device that enables a terminal device to access a wireless network, and may also be referred to as a network device or base station. RAN devices include, but are not limited to, a base station, a 5G generation NodeB (gNB), an evolved node B (eNB), a radio network controller (RNC), a node B (NB), a base station controller (BSC), a base transceiver station (BTS), a home base station (e.g., home evolved nodeB, or home node B, HNB), a base band unit (BBU), a transmission reception point (TRP), a transmitting point (TP), and / or a mobile switching center. Alternatively, the access network device may be at least one of a central unit (CU), a distributed unit (DU), a central unit control plane (CU-CP) node, a central unit user plane (CU-UP) node, an integrated access and backhaul (IAB) node, a radio controller of a cloud radio access network (CRAN), etc. Alternatively, the access network device may be a relay station, an access point, an in-vehicle device, a terminal device, a wearable device, an access network device in a 5G network, an access network device in a future evolved public land mobile network (PLMN), etc.

[0041] In the present disclosure, the apparatus configured to implement the functions of the access network device may be an access network device, or may be an apparatus capable of supporting the access network device in implementing the functions, such as a chip system, a hardware circuit, a software module, or a hardware circuit combined with a software module. The apparatus may be installed in the access network device, or may be matched with the access network device for use. In the technical solution provided in the present disclosure, an example in which the apparatus configured to implement the functions of the access network device is an access network device, and the access network device is a base station, is used to describe the technical solution provided in the present disclosure.

[0042] (1) Protocol Layer Structure The communication between the access network device and the terminal device follows a specific protocol layer structure. The protocol layer structure may include a control plane protocol layer structure and a user plane protocol layer structure. For example, the control plane protocol layer structure may include at least one of the following: a radio resource control (RRC) layer, a packet data convergence protocol (PDCP) layer, a radio link control (RLC) layer, a media access control (MAC) layer, a physical layer (PHY), etc. For example, the user plane protocol layer structure may include at least one of the following: a service data adaptation protocol (SDAP) layer, a PDCP layer, an RLC layer, a MAC layer, a physical layer, etc.

[0043] The protocol layer structure between the access network device and the terminal device can be considered as an access stratum (AS) structure. Optionally, a non-access stratum (NAS) may further exist on the AS, and is used by the access network device to forward information from the core network device to the terminal device, or is used by the access network device to forward information from the terminal device to the core network device. In this case, it can be considered that there is a logical interface between the terminal device and the core network device. Optionally, the access network device may forward information between the terminal device and the core network device via transparent transmission. For example, the NAS message may be mapped to or included in the RRC signaling as an element of the RRC signaling.

[0044] Optionally, the protocol layer structure between the access network device and the terminal device may further include an artificial intelligence (AI) layer for transmitting data related to AI functions.

[0045] (2) Central unit (CU) and distributed unit (DU) The RAN device may include a CU and a DU. This design may be referred to as a CU and DU separated design. Multiple DUs may be centrally controlled by one CU. For example, an interface between a CU and a DU may be referred to as an F1 interface. A control plane (CP) interface may be F1-C, and a user plane (UP) interface may be F1-U. The specific names of the interfaces are not limited in this disclosure. The CU and the DU may be divided based on a protocol layer of a wireless network. For example, the functions of the PDCP layer and the protocol layers above the PDCP layer (e.g., the RRC layer and the SDAP layer) are set in the CU, and the functions of the protocol layers below the PDCP layer (e.g., the RLC layer, the MAC layer, and the PHY layer) are set in the DU. In another example, the functions of the protocol layers above the PDCP layer are set in the CU, and the functions of the PDCP layer and the protocol layers below the PDCP layer are set in the DU.

[0046] The division of the CU and DU into processing functions based on protocol layers is only an example, and may be divided in other ways. For example, the CU or DU may be divided to have functions with more protocol layers. In another example, the CU or DU may be further divided into some processing functions including multiple protocol layers. In a design, some functions of the RLC layer and functions of protocol layers above the RLC layer are set to the CU, and the remaining functions of the RLC layer and functions of protocol layers below the RLC layer are set to the DU. In another design, the division of the functions of the CU or DU may be performed based on service type or another system requirement instead. For example, the division may be performed based on delay. Functions whose processing time needs to meet the delay requirement are set to the DU, and functions whose processing time does not need to meet the delay requirement are set to the CU.

[0047] Optionally, the CU may also have one or more functions of a core network. For example, the CU may be located on the network side to facilitate centralized management.

[0048] Optionally, a radio unit (RU) of the DU is located remotely. The RU has radio frequency functions. For example, the DU and the RU may be divided at the PHY layer. For example, the DU may implement the upper layer functions of the PHY layer, and the RU may implement the lower layer functions of the PHY layer. If transmission is performed, the PHY layer functions may include at least one of the following: adding a cyclic redundancy check (CRC), channel coding, rate matching, scrambling, modulation, layer mapping, precoding, resource mapping, physical antenna mapping, or radio frequency transmission. If reception is performed, the PHY layer functions may include at least one of the following: CRC check, channel decoding, de-rate matching, descrambling, demodulation, layer demapping, channel detection, resource demapping, physical antenna demapping, or radio frequency reception. The upper layer functions of the PHY layer may include some of the functions of the PHY layer. For example, some of the functions are closer to the MAC layer. The lower layer functions of the PHY layer may include another part of the functions of the PHY layer. For example, some of the functions are closer to radio frequency reception functions. For example, the upper layer functions of the PHY layer may include adding CRC check, channel coding, rate matching, scrambling, modulation, and layer mapping, and the lower layer functions of the PHY layer may include precoding, resource mapping, physical antenna mapping, and radio frequency transmission functions. Alternatively, the upper layer functions of the PHY layer may include CRC check, channel coding, rate matching, scrambling, modulation, layer mapping, and precoding, and the lower layer functions of the PHY layer may include resource mapping, physical antenna mapping, and radio frequency transmission functions. For example, the upper layer functions of the PHY layer may include adding CRC check, channel decoding, de-rate matching, decoding, demodulation, and layer demapping, and the lower layer functions of the PHY layer may include channel detection, resource demapping, physical antenna demapping, and radio frequency reception.Alternatively, the upper layer functions of the PHY layer may include CRC checking, channel decoding, de-rate matching, decoding, demodulation, layer demapping, and channel detection, and the lower layer functions of the PHY layer may include resource demapping, physical antenna demapping, and radio frequency reception.

[0049] Optionally, the function of the CU may be further divided, and the control plane and the user plane are separated and implemented by using separate entities, i.e., a control plane CU entity (i.e., a CU-CP entity) and a user plane CU entity (i.e., a CU-UP entity), respectively. The CU-CP entity and the CU-UP entity may be separately coupled or connected to the DU to jointly complete the functions of the RAN device.

[0050] In the above network architecture, the signaling generated by the CU may be transmitted to the terminal device by using the DU, or the signaling generated by the terminal device may be transmitted to the CU by using the DU. For example, the signaling at the RRC layer or the PDCP layer may be finally processed as the signaling at the physical layer and transmitted to the terminal device, or converted from the signaling received from the physical layer. In this architecture, the signaling at the RRC layer or the PDCP layer may be considered to be transmitted by using the DU, or transmitted by using the DU and the RU.

[0051] Optionally, any one of the DU, CU, CU-CP, CU-UP, and RU may be a software module, a hardware structure, or a combination of a software module and a hardware structure. This is not limited. Different entities may exist in different forms. This is not limited. For example, the DU, CU, CU-CP, and CU-UP are software modules, and the RU is a hardware structure. The modules and the methods performed by the modules also fall within the protection scope of the present disclosure. For example, when the method of the present disclosure is performed by an access network device, the method may be specifically performed by at least one of the CU, CU-CP, CU-UP, DU, RU, or near-real-time RIC described below. The methods performed by the modules also fall within the protection scope of the present disclosure.

[0052] Please note that since the network devices in the present disclosure are mainly access network devices, in the following, unless otherwise specified, a "network device" may also be an "access network device".

[0053] It should be understood that the number of devices in the communication system shown in Figure 1 is used only as an example, and the present disclosure is not limited thereto. In practical application, the communication system may further include more terminal devices and more RAN devices, and may further include other devices, such as core network devices and / or nodes configured to implement artificial intelligence functions.

[0054] The network architecture shown in FIG. 1 may be applicable to communication systems of various radio access technologies (RATs), such as a 4G communication system, a 5G (or called new radio (NR)) communication system, a transition system between an LTE communication system and a 5G communication system, and the transition system may be called a 4.5G communication system, or a future communication system, such as a 6G communication system. The network architecture and service scenarios described in this disclosure are intended to more clearly explain the technical solutions in this disclosure, and do not constitute limitations on the technical solutions provided in this disclosure. Those skilled in the art may know that with the development of network architectures and the emergence of new service scenarios, the technical solutions provided in this disclosure are also applicable to similar technical problems.

[0055] In addition to communication between an access network device and a terminal device, the method provided in the present application can also be used for communication between other communication devices, such as communication between a macro base station and a micro base station in a wireless backhaul link, or communication between a first terminal device and a second terminal device in a sidelink (SL), without being limited thereto. The present disclosure is described by using communication between a network device and a terminal device as an example.

[0056] When transmitting data to the terminal device, the access network device can perform precoding based on channel state information (CSI) fed back by the terminal device. In order to facilitate the understanding of the present disclosure, the following briefly describes some technical terms in the present disclosure.

[0057] 1. Precoding Technology When the channel state information is known, the access network device can process the signal to be transmitted by using a precoding matrix that matches the channel condition. By using this technique, the precoded signal to be transmitted matches the channel, which results in an improved quality of the signal received by the terminal device (e.g., signal to interference plus noise ratio (SINR)), and further improves the system throughput rate. By using the precoding technique, a transmitting device (e.g., an access network device) and multiple receiving devices (e.g., terminal devices) can effectively perform transmission on the same time-frequency resource, i.e., can effectively implement multiple user multiple input multiple output (MU-MIMO). By using the precoding technique, a transmitting device (e.g., an access network device) and a receiving device (e.g., terminal devices) can effectively transmit multiple data streams on the same time-frequency resource, i.e., can effectively implement single user multiple input multiple output (SU-MIMO). It should be noted that the relevant description of the precoding technique is merely an example for ease of understanding, and is not intended to limit the scope of the disclosure of the present disclosure. In a particular implementation process, the transmitting device may alternatively perform precoding in another manner. For example, when channel information (such as, but not limited to, a channel matrix) cannot be obtained, precoding is performed by using a pre-configured precoding matrix or by using a weighting process. For the sake of brevity, the specific contents are not described again in this specification.

[0058] 2. CSI feedback The CSI feedback may also be referred to as a CSI report. According to the CSI feedback, in a wireless communication system, a receiver (e.g., a terminal device) of data (e.g., data carried on a physical downlink shared channel (PDSCH), but not limited thereto) reports information to a transmitter (e.g., an access network device) for describing channel attributes of a communication link. The CSI report includes one or more pieces of information, such as a downlink channel matrix, a precoding matrix indicator (PMI), a rank indicator (RI), a channel quality indicator (CQI), and the like. The contents included in the enumerated CSI are merely examples for illustration and do not constitute any limitation to the present disclosure. The CSI may include one or more of the aforementioned items, or may include other information used to represent the CSI in addition to the aforementioned enumerated information. This is not limited in the present disclosure.

[0059] 3.Neural network (NN) Neural networks are a specific implementation of machine learning technology. According to the universal approximation theorem, neural networks can theoretically approximate any continuous function, and as a result, neural networks can learn any mapping. In traditional communication systems, extensive expertise is required to design communication modules. However, neural network-based deep learning communication systems can automatically find implicit pattern structures from large data sets, establish mapping relationships between data, and obtain better performance than traditional modeling methods.

[0060] For example, a deep neural network (DNN) is a neural network with a large number of layers. Depending on different network structures and / or usage scenarios, the DNN may include a multi-layer perceptron (MLP), a convolutional neural network (CNN), a recurrent neural network (RNN), etc. The specific form of the DNN is not limited in this disclosure.

[0061] 4. Auto-encoder (AE) network, or AE for short. The AE network may include an encoder and a corresponding decoder. For example, the encoder and / or the decoder may be implemented by using a neural network (e.g., DNN). In this case, the encoder may be referred to as an encoder network, and the decoder may be referred to as a decoder network. For example, in the AE network, the encoder and the corresponding decoder may be obtained by joint training. The encoder and the decoder obtained by training may be configured to encode and decode information.

[0062] In this disclosure, unless otherwise stated, the quantity of a noun refers to "singular or plural", i.e., "one or more". "At least one" means one or more, and "multiple" means two or more. "And / or" describes an association relationship between associated objects and indicates that three relationships may exist. For example, A and / or B may indicate three cases: only A is present, both A and B are present, and only B is present, where A and B may be singular or plural. When indicating a feature, the character " / " may indicate an "or" relationship between associated objects. For example, A / B indicates A or B. When indicating an operation, the symbol " / " may further represent the operation of division. Also, in this disclosure, the symbol "x" may also be replaced with the symbol "*".

[0063] In the present disclosure, ordinals such as "first" and "second" are used to distinguish between multiple objects, and are not used to limit the size, content, sequence, time series, application scenario, priority, importance, etc. of the multiple objects. For example, the first instruction information and the second instruction information may be the same instruction information or different instruction information. Furthermore, this type of name does not indicate that the size, transmission mode, instruction content, priority, application scenario, importance, etc. of the two instruction information are different.

[0064] In a possible implementation, the CSI feedback mechanism uses the procedure shown in FIG.

[0065] S21: The base station sends signaling, and correspondingly, the UE receives signaling from the base station.

[0066] The signaling is used to configure channel measurement information, for example, the signaling informs the UE of at least one of the following: time information for performing channel measurement, type of reference signal (RS) for performing channel measurement, time domain resource of the reference signal, frequency domain resource of the reference signal, reporting condition of the measured quantity, etc.

[0067] S22: The base station sends a reference signal to the UE, and in response, the UE receives a reference signal from the base station.

[0068] The UE measures the reference signal to obtain the CSI.

[0069] S23: The UE transmits the CSI to the base station, and in response, the base station receives the CSI from the UE.

[0070] S24: The base station sends data to the UE based on the CSI, and correspondingly, the UE receives data from the base station.

[0071] Data transmitted by the base station to the UE is carried on a downlink channel, for example on the PDSCH.

[0072] The CSI may represent information about the downlink channel matrix. After the UE feeds back the CSI to the base station, the base station may restore the downlink channel information based on the CSI to determine information such as a precoding matrix based on the downlink channel information. A higher accuracy of the CSI fed back by the UE indicates more information about the downlink channel matrix and more accurate downlink channel information that may be restored by the base station based on the CSI. Thus, a more accurate precoding matrix determined by the base station indicates better downlink spatial multiplexing performance, a higher received signal-to-interference-and-noise ratio of the UE, and a higher system throughput rate. Alternatively, after the UE feeds back the CSI to the base station, the base station does not need to restore the downlink channel information, but may determine information such as a precoding matrix based on the CSI. In this case, a higher CSI accuracy indicates a more accurate precoding matrix determined by the base station. However, as the size of the antenna array in the MIMO system increases continuously, the quantity of antenna ports that can be supported also increases. Because the size of the complete downlink channel matrix is ​​directly proportional to the quantity of antenna ports, in a massive MIMO system, a huge feedback overhead is required to ensure a high accuracy of the CSI fed back by the UE. The huge feedback overhead reduces the resources available for data transmission, and thus reduces the capacity of the system. Therefore, it is necessary to consider how to reduce the CSI feedback overhead in order to increase the capacity of the system.

[0073] Deep learning (DL) is a machine learning method based on a deep neural network. The CSI is fed back by deep learning. Compared with the conventional solution, this method can represent almost the same downlink channel information with less feedback amount, so that the CSI feedback overhead can be reduced, or the channel restoration accuracy can be improved by using the same overhead. In the deep learning-based CSI compression feedback technique, the channel compression feedback is performed by using an AE model based on a convolutional neural network (CNN). The AE model includes an encoder model and a decoder model that match each other for use. Specifically, in the UE (i.e., the information transmitting side), the encoder model of the AE maps the downlink channel matrix in a slot to compressed information, and the encoder model is a CNN-type neural network. In the base station (i.e., the information receiving side), the decoder model of the AE can restore the compressed information to the downlink channel information, and the decoder model is a CNN-type neural network. In this way, the UE may need to feed back the CSI of multiple slots to the base station separately. This still incurs a large overhead.

[0074] In consideration of this, the technical solution of the present disclosure is provided. In the present disclosure, the UE can perform channel estimation on the downlink channel in M ​​time units, perform joint compression on the obtained M channel estimation results, and report the obtained M channel estimation results to the access network device. Since the joint compression is performed, the correlation between the downlink channels at different time points is fully utilized, and the access network device can restore the N pieces of compressed information obtained by the joint compression, so as to obtain more accurate and effective downlink channel information. In addition, since the joint compression is performed, the N pieces of compressed information can be complemented with each other. For example, different pieces of compressed information can correspond to different parameters, so as to reduce redundant information and correspondingly reduce transmission overhead.

[0075] 3 illustrates an architecture of a communication network in a communication system 10 according to the present disclosure. Any embodiment provided thereafter is applicable to the architecture. The network device included in FIG. 3 is, for example, an access network device 20 included in the communication system 10, and the terminal device included in FIG. 3 is, for example, a communication device 30 included in the communication system 10. The network device can communicate with the terminal device.

[0076] The machine learning technology of the present disclosure is a specific implementation of artificial intelligence (AI) technology. For ease of understanding, the AI ​​technology will be described below. It can be understood that the description is not intended to limit the present disclosure.

[0077] AI is a technology that performs complex calculations by simulating the human brain. With improvements in data storage and capabilities, AI is being applied more widely.

[0078] In the present disclosure, an independent network element (e.g., referred to as an AI network element, AI node, or AI device) may be introduced into the communication system shown in FIG. 1 to implement the AI ​​function. The AI ​​network element may be directly connected to the access network device, or may be indirectly connected to the access network device by using a third-party network element. Optionally, the third-party network element may be a core network element. Alternatively, the AI ​​entity may be configured or located in another network element in the communication system to implement AI-related operations. The AI ​​entity may also be referred to as an AI module, an AI unit, or another name, and is mainly configured to implement some or all of the AI ​​functions. The specific name of the AI ​​entity is not limited in the present disclosure. Optionally, the other network element may be an access network device, a core network device, an operation, administration and maintenance (OAM), etc. In this case, the network element that performs the AI ​​function is a network element with an AI function built in.

[0079] In this disclosure, AI functions may include at least one of the following: data collection, model training (or model learning), model information publication, model inference (or also referred to as model estimation, inference, prediction, etc.), model monitoring or model validation, inference result publication, etc. AI functions may also be referred to as AI (related) operations or AI-related functions.

[0080] In the present disclosure, the OAM is configured to operate, manage, and / or maintain a core network device (OAM of the core network device) and / or an access network device (OAM of the access network device). For example, the present disclosure includes a first OAM and a second OAM, where the first OAM is an OAM of the core network device and the second OAM is an OAM of the access network device. Optionally, the first OAM and / or the second OAM includes an AI entity. In another example, the present disclosure includes a third OAM, where the third OAM is an OAM of both the core network device and the access network device. Optionally, the third OAM includes an AI entity.

[0081] FIG. 4A is a schematic diagram of a first application framework of AI in a communication system. A data source is used to store training data and inference data. A model training host performs training or update training based on training data provided by a data source to obtain an AI model, and deploys the AI ​​model to a model inference host. The AI ​​model represents a mapping relationship between the input and output of the model. Obtaining an AI model by learning by the model training host is equivalent to obtaining a mapping relationship between the input and output of the model by learning by the model training host by using training data. The model inference host uses the AI ​​model to perform inference based on inference data provided by a data source to obtain an inference result. This method can alternatively be described as follows: the model inference host inputs inference data to the AI ​​model and obtains an output by using the AI ​​model, and the output is an inference result. The inference result can indicate a configuration parameter used (acted upon) by an action subject, and / or an operation performed by an action subject. The inference result may be uniformly planned by an actor and sent to one or more targets of an action object (e.g., a core network element, a base station, or a UE) for action. Optionally, the model inference host may feed back the inference result of the model inference host to the model training host. This process may be referred to as model feedback. The fed back inference result is used by the model training host to update the AI ​​model, and the updated AI model is deployed to the model inference host. Optionally, the action target may feed back the network parameters collected by the action target to the data source. This process may also be referred to as performance feedback, and the fed back network parameters may be used as training data or inference data.

[0082] For example, the AI ​​model includes a decoder network in an AE network. The decoder network is deployed on an access network device side. The inference result of the decoder network is, for example, used to reconstruct a downlink channel matrix. The AI ​​model includes an encoder network in an AE network. The encoder network is deployed on a UE side. The inference result of the encoder network is, for example, used to encode a downlink channel matrix.

[0083] The application framework shown in FIG. 4A may be deployed on the network elements shown in FIG. 1. For example, the application framework in FIG. 4A may be deployed on at least one of the terminal device, the access network device, the core network device (not shown), or the independently deployed AI network element (not shown) in FIG. 1. For example, the AI ​​network element (which may be considered as a model training host) may analyze or train training data provided by the terminal device and / or the access network device to obtain a model. At least one of the terminal device, the access network device, or the core network device (which may be considered as a model inference host) may perform inference by using the model and the inference data to obtain an output of the model. The inference data may be provided by the terminal device and / or the access network device. The input of the model includes the inference data, and the output of the model is the inference result corresponding to the model. At least one of the terminal device, the access network device, or the core network device (which may be considered as a target of the action) may perform a corresponding operation based on the inference result. The model inference host and the target of the action may be the same or different. This is not limited.

[0084] With reference to FIGS. 4B to 4E, the following describes, by using an example, a network architecture to which the method provided in the present disclosure can be applied.

[0085] As shown in FIG. 4B, in a first possible implementation, the access network device includes a near-real-time radio access network intelligent controller (RAN intelligent controller, RIC) module for model training and inference. For example, the near-real-time RIC may be configured to train an AI model and use the AI ​​model for inference. For example, the near-real-time RIC may obtain network-side and / or terminal-side information from at least one of the CU, DU, and RU, and the information may be used as training data or inference data. Optionally, the near-real-time RIC may submit inference results to at least one of the CU, DU, RU, or terminal device. Optionally, the CU and the DU may exchange inference results. Optionally, the DU and the RU may exchange inference results. For example, the near-real-time RIC submits the inference results to the DU, and the DU forwards the inference results to the RU.

[0086] As shown in FIG. 4B, in a second possible implementation, a non-real-time RIC (optionally, the non-real-time RIC may be located in an OAM or core network device) is located outside the access network and is used for model training and inference. For example, the non-real-time RIC is configured to train an AI model and use the model for inference. For example, the non-real-time RIC may obtain network-side and / or terminal-side information from at least one of a CU, a DU, or a RU, which may be used as training data or inference data. The inference result may be submitted to at least one of a CU, a DU, a RU, or a terminal device. Optionally, the CU and the DU may exchange the inference result. Optionally, the DU and the RU may exchange the inference result. For example, the non-real-time RIC submits the inference result to the DU, and the DU forwards the inference result to the RU.

[0087] As shown in FIG. 4B, in a third possible implementation, the access network device includes a near-real-time RIC, and the non-real-time RIC is located outside the access network device (optionally, the non-real-time RIC may be located in an OAM or core network device). As in the second possible implementation described above, the non-real-time RIC can be used for model training and inference, and / or as in the first possible implementation, the near-real-time RIC can be used for model training and inference, and / or the non-real-time RIC can perform model training, and the near-real-time RIC can obtain AI model information from the non-real-time RIC, obtain network-side and / or terminal-side information from at least one of the CU, DU, or RU, and obtain an inference result by using the information and the AI ​​model information. Optionally, the near-real-time RIC can submit an inference result to at least one of the CU, DU, RU, or terminal device. Optionally, the CU and the DU can exchange inference results. Optionally, the DU and the RU can exchange inference results. For example, the near-real-time RIC submits an inference result to the DU, and the DU forwards the inference result to the RU. For example, the near real-time RIC is configured to train model A and use model A for inference. For example, the non-real-time RIC is configured to train model B and use model B for inference. For example, the non-real-time RIC is configured to train model C and send information about model C to the near real-time RIC, which uses model C for inference.

[0088] 4C is an exemplary diagram of a network architecture in which the method according to the present disclosure can be applied. Compared with FIG. 4B, the CU in FIG. 4C is separated into a CU-CP and a CU-UP.

[0089] 4D is an example diagram of a network architecture to which the method according to the present disclosure may be applied. As shown in FIG. 4D, optionally, the access network device includes one or more AI entities, and the functions of the AI ​​entities are similar to those of the near-real-time RIC. Optionally, the OAM includes one or more AI entities, and the functions of the AI ​​entities are similar to those of the non-real-time RIC. Optionally, the core network device includes one or more AI entities, and the functions of the AI ​​entities are similar to those of the non-real-time RIC. When the OAM and the core network device each include an AI entity, the models trained and obtained by the AI ​​entities of the OAM and the core network device are different, and / or the models used for inference are different.

[0090] In this disclosure, the models differ in at least one of the following: structural parameters of the model (e.g., at least one of the number of neural network layers, the neural network width, the connectivity between layers, the weights of the neurons, the activation functions of the neurons, or the offsets in the activation functions), input parameters of the model (e.g., the types of the input parameters and / or the dimensions of the input parameters), or output parameters of the model (e.g., the types of the output parameters and / or the dimensions of the output parameters).

[0091] FIG. 4E is an exemplary diagram of a network architecture to which the method according to the present disclosure may be applied. Compared with FIG. 4D, the access network device in FIG. 4E is separated into a CU and a DU. Optionally, the CU may include an AI entity, and the function of the AI ​​entity is similar to that of the near-real-time RIC. Optionally, the DU may include an AI entity, and the function of the AI ​​entity is similar to that of the near-real-time RIC. When the CU and the DU each include an AI entity, the models obtained by training by the AI ​​entities of the CU and the DU are different, and / or the models used for inference are different. Optionally, the CU in FIG. 4E may be further divided into a CU-CP and a CU-UP. Optionally, one or more AI models may be deployed in the CU-CP. Optionally, one or more AI models may be deployed in the CU-UP.

[0092] In Fig. 4D or Fig. 4E, the OAM of the access network device and the OAM of the core network device are deployed uniformly. Alternatively, as described above, in Fig. 4D or Fig. 4E, the OAM of the access network device and the OAM of the core network device may be deployed separately.

[0093] In the present disclosure, a model can obtain an output through inference, and the output includes one or more parameters. The learning or training processes of different models may be deployed on different devices or nodes, or may be deployed on the same device or node. The inference processes of different models may be deployed on different devices or nodes, or may be deployed on the same device or node.

[0094] Optionally, the AI ​​model includes a decoder network in the AE network. At the base station side, the inference result of the decoder network is used, for example, to reconstruct the downlink channel matrix. Optionally, the AI ​​model includes an encoder network in the AE network. Model information of the encoder network can be sent to the UE for inference.

[0095] In the framework of FIG. 4A-FIG. 4D, it should be noted that the AI ​​model may be referred to as a model or a network model for short, and may be considered as a mapping from input parameters (e.g., input matrix) to output parameters (e.g., output matrix). For example, in the case of a decoder network on the base station side, the input matrix may be a matrix determined based on the received CSI. The training data may include a known input matrix, or may include a known input matrix and a corresponding output matrix, and is used to train the AI ​​model. The training data may be data from the access network device, the CU, the CU-CP, the CU-UP, the DU, the RU, the UE, and / or another entity, and / or data obtained by inference using AI techniques. This is not limited. The inference data includes the input matrix and is used to infer the output matrix by using the model. The inference data may be data from the access network device, the CU, the CU-CP, the CU-UP, the DU, the RU, the UE, and / or another entity. The inferred matrix may be considered as policy information and is transmitted to the target of the action. The matrix obtained by inference may be transmitted to an access network device, a CU, a CU-CP, a CU-UP, a DU, a RU, or a UE for further processing, e.g., reconstruction of a downlink channel matrix.

[0096] In the present disclosure, on the network side, the decoder network in the AE network may be deployed in an access network device (such as a base station) or outside the access network device, for example, in an OAM, an AI network element, a core network device, an RU, a DU, or a near-real-time RIC. This is not limited. The inference result of the decoder network may be obtained by inference performed by the access network device, or may be sent to the access network device after a non-real-time RIC performs inference. For simplicity of explanation, the present disclosure is described by using an example in which the decoder network is deployed in the access network device.

[0097] In this disclosure, an encoder network in an AE network is deployed in a UE, and the UE can perform inference by using the encoder network.

[0098] The method provided in the present disclosure will be described below with reference to the accompanying drawings. The steps or operations included in the method are merely examples, and other operations or variations of various operations may be further performed in the present disclosure. Furthermore, the steps may be performed in an order different from the order presented in the present disclosure, and all operations of the present disclosure may not be performed.

[0099] FIG. 5 is a flow chart of a communication method according to the present disclosure.

[0100] Optionally, S501: an access network device sends configuration information to a UE. In response, the UE receives configuration information from the access network device.

[0101] The configuration information may have another name, for example, may be called CSI feedback configuration information or feedback configuration information. The name is not limited in this disclosure. The configuration information may be used to configure M time units, where M is an integer greater than 1. The M time units are time units in which the downlink channel information that the UE needs to feed back is located. In other words, the UE needs to feed back downlink channel information in M ​​time units. M is greater than 1, i.e., the UE can feed back downlink channel information in multiple time units to the access network device. The M time units may be consecutive, i.e., the M time units are of consecutive duration. In this case, the M time units may also be considered as one time frame, for example, referred to as a first time frame. In other words, the configuration information is used to configure the first time frame. Alternatively, the M time units may be distinct. For example, at least two adjacent time units in the M time units are discontinuous. For example, the time unit is a subframe, a slot, a mini-slot, or an orthogonal frequency division multiplexing (OFDM) symbol. In this disclosure, an example in which the time unit is a slot is used. Therefore, "slot" is used in the following description. That is, "slot" in the following can be read as "time unit".

[0102] The start time domain positions of the M slots may be located before the current time domain position, or after the current time domain position, or may be the current time domain position. The current time domain position is, for example, the time domain position where the UE receives the configuration information. In other words, the downlink channels in the M slots may include one or more of a past downlink channel, a current downlink channel, and a future downlink channel.

[0103] For example, the configuration information includes one or more of the following: start time domain position of the M slots, end time domain position of the M slots, duration of the M slots, number of the M slots, sampling period in the M slots, quantity of sampling points in the M slots, time domain position of the first sampling point in the M slots, or time domain positions of sampling points (e.g., all sampling points) in the M slots. The number of slots is, for example, the number of slots in a subframe or the number of slots in a radio frame. The sampling points are sampling instants at which the UE needs to estimate the downlink channel at the sampling points. The first sampling point in the M slots is the first sampling point in the M slots in the time domain. For example, the sampling points are represented by using slots, and one sampling point corresponds to one slot. The UE performs channel estimation for the downlink channel at the sampling points, i.e., performs information estimation for the downlink channel in the slot. Alternatively, the sampling points may be represented by using OFDM symbols.

[0104] For example, if the configuration information includes a start time domain position of the M slots, an end time domain position of the M slots, and a sampling period within the M slots, the UE may determine a position of a sampling point within the M slots based on the configuration information to perform channel estimation for the downlink channel at the sampling point. As another example, if the configuration information includes a start time domain position (or an end time domain position) of the M slots, a duration of the M slots, and a sampling period within the M slots, the UE may determine a position of a sampling point within the M slots based on the configuration information to perform channel estimation for the downlink channel at the sampling point. As another example, if the configuration information includes a start time domain position (or an end time domain position) of the M slots, a sampling period within the M slots, and a quantity of sampling points, the UE may determine a position of a sampling point within the configuration information to perform channel estimation for the downlink channel at the sampling point. As another example, if the configuration information includes a start time domain position of the M slots, a duration of the M slots, and a quantity of sampling points within the M slots, the UE may determine a position of a sampling point within the configuration information to perform channel estimation for the downlink channel at the sampling point. For example, the M slots are consecutive, the configuration information includes a quantity M of sampling points in the M slots, the starting time-domain position of the M slots is n1, and the duration of the M slots is T. In this case, the UE may determine that the sampling period in the M slots is T / M, so that the time-domain positions of the sampling points in the M slots are [n1, n1+T / M, n1+2T / M,..., n1+(N-1)T / M]. Some of the configuration schemes described above are applicable when the M slots are consecutive.

[0105] For example, when the configuration information includes the number of M slots, the time-domain positions of the first sampling points in the M slots, and the sampling period in the M slots, the UE may determine the time-domain positions of the M slots based on the number of M slots, and then may determine the time-domain positions of each sampling point in the M slots based on the time-domain positions of the first sampling points in the M slots and the sampling period, so as to perform channel estimation for the downlink channel at the sampling points. The configuration scheme is applicable when the M slots are separate.

[0106] For example, if the configuration information includes time-domain locations of sampling points within the M slots, the UE may perform channel estimation for the downlink channel at the sampling points based on the configuration information. The configuration scheme is applicable to both the case where the M slots are distinct and the case where the M slots are consecutive.

[0107] Optionally, one or more of the following information, namely, the start time domain position of the M slots, the end time domain position of the M slots, the duration of the M slots, the number of the M slots, the sampling period of the M slots, the quantity of the sampling points of the M slots, or the time domain positions of the sampling points of the M slots, may be predefined in the protocol, preconfigured in the UE, or may use default values. For example, if the start time domain position of the M slots is predefined in the protocol, the first offset is predefined in the protocol. For example, the start time domain position of the M slots is a time domain position obtained by adding the first offset at the moment the UE receives the configuration information, and the first offset is a real number. In another example, the sampling period in the M slots is predefined in the protocol as one transmission time interval (TTI). In another example, the duration of the M slots is the duration of the shortest time window that includes the required sampling points by default. In another example, the quantity of the sampling points in the M slots is P by default, and P is a real number. One TTI may be, but is not limited to, one or more subframes, one or more slots, one or more symbols, or another possible configuration.

[0108] The UE may determine the M slots based on configuration information and information predefined in the protocol (or information preconfigured in the UE, or default information). For example, the configuration information includes that the start time-domain position of the M slots is n1, the duration is T, and the sampling period in the M slots predefined in the protocol is 1 TTI. In this case, the UE may determine the time-domain positions of the sampling points in the M slots based on n1, T, and the sampling period to perform channel estimation for the downlink channel at the sampling points.

[0109] Alternatively, the UE can determine the M slots based on information predefined in the protocol (or preconfigured in the UE, or default information), and the access network device does not need to send configuration information. Therefore, S501 is an optional step.

[0110] Optionally, the access network device may further transmit information used to configure the reference signal to the UE. For example, the information may be referred to as reference signal configuration information. This information may be used to configure a time domain resource, a frequency domain resource, and / or the like of the reference signal, and the UE may detect the reference signal from the access network device based on this information. The access network device may convey the information used to configure the M slots and the configuration information in one message for transmission, or may transmit the information used to configure the M slots and the configuration information separately. When the access network device transmits the information used to configure the M slots and the configuration information separately, the two pieces of information may be transmitted simultaneously, or the information may be transmitted first, or the configuration information used to configure the M slots may be transmitted first.

[0111] Optionally, the access network device further transmits a downlink reference signal to the UE, and the UE can receive the downlink reference signal based on the information used to configure the downlink reference signal, and perform channel estimation for the downlink channel based on the downlink reference signal. The access network device can transmit the downlink reference signal to the UE in the M slots, or the access network device can transmit the downlink reference signal to the UE before the start time-domain position of the M slots arrives, or the access network device can transmit the downlink reference signal to the UE in the M slots before the start time-domain position of the M slots arrives. The downlink reference signal is, for example, a synchronization signal and physical broadcast channel (PBCH) block (SSB) and a channel state information reference signal (CSI-RS).

[0112] S502: The UE obtains M channel estimation results. The M channel estimation results correspond to the M time units. For example, the M channel estimation results may reflect characteristics of a downlink channel in the M time units.

[0113] The UE may determine time-domain locations of the sampling points within the M slots based on the configuration information and / or the information used to configure the reference signal, such that the UE performs channel estimation on the downlink channel at the sampling points. For example, the sampling points are represented by using slots. In this disclosure, one slot corresponds to one sampling point is used as an example. In this case, the M slots may correspond to the M sampling points. The UE may perform channel estimation on the downlink channel at one sampling point and obtain a channel estimation result. In this case, the UE may perform channel estimation on the downlink channel at the M sampling points and obtain M channel estimation results.

[0114] For example, when one of the M channel estimation results is referred to as a first channel estimation result, the first channel estimation result may be a channel estimation result obtained by the UE by performing a measurement based on a received downlink reference signal. For example, the access network device may transmit a downlink reference signal to the UE in the M slots, or the access network device may transmit a downlink reference signal to the UE before the start time-domain position of the M slot arrives, or the access network device may transmit a downlink reference signal to the UE in the M slots before the start time-domain position of the M slot arrives. The UE may receive a downlink reference signal in the M slots, or may receive a downlink reference signal before the start time-domain position of the M slot arrives, or may receive a downlink reference signal before the start time-domain position of the M slot arrives and within the M slots. The time-domain position at which the UE receives the downlink reference signal is not limited in this disclosure. If the time-domain location of the sampling point within the M slots is the current time-domain location or is before the current time-domain location, the UE may measure the downlink reference signal, regardless of the particular time-domain location or the particular time-domain location at which the UE receives the downlink reference signal, to obtain a channel estimation result at the sampling point.

[0115] Alternatively, the first channel estimation result may be a processing result obtained by the UE by processing the second channel estimation result, and the second channel estimation result is obtained by the UE by measuring the received downlink reference signal. In other words, the first channel estimation result is obtained after the UE processes the obtained channel estimation result. For example, the time domain position of the sampling point in the M slots is a time domain position after the current time domain position, i.e., the time domain position of the sampling point has not been reached. If the UE performs channel estimation when the time domain position of the sampling point arrives, the previously obtained channel estimation result may be invalid due to excessively long latency. Alternatively, in the M slots, the UE only measures the downlink reference signal in some slots. Thus, in the processing manner, the UE can process the obtained channel estimation result to obtain the channel estimation result of the future slot or the unmeasured slot, or the UE can obtain the channel estimation result of the future time unit or the unmeasured time unit through prediction. In this way, the UE can obtain the channel estimation result of more slots in a timely manner to reduce latency and improve channel information feedback efficiency. For the process of the UE measuring the received downlink reference signal to obtain the second channel estimation result, please refer to the description in the previous paragraph. After obtaining the second channel estimation result, the UE can process the second channel estimation result, and the processing result can be used as the first channel estimation result.

[0116] For example, the first channel estimation result may be obtained by using a convolution of the second channel estimation result with a shaped waveform.

number

number

number

[0117] In formula 1, j is an integer ranging from 1 to M, and g j (m) represents the coefficient of the shaped waveform. In particular, if the M′ second channel estimation results are channel estimation results obtained by equally spaced measurements and the measurement period is T, then:

number

number

[0118] Alternatively, the first channel estimation result may be a channel estimation result obtained through prediction by the UE. As mentioned above, when the time domain position of the sampling point in the M slots is a time domain position after the current time domain position, the UE can obtain the channel estimation result of the future sampling point through corresponding processing, and prediction is another processing method.

[0119] For example, the UE may predict the first channel estimation result by using a prony algorithm. For example, the UE may perform a weighted summation on the channel estimation results of b slots from the (nb)th slot to the (n-1)th slot to obtain the first channel estimation result of the nth slot. Alternatively, the UE may perform channel prediction by using a long short-term memory (LSTM) artificial neural network to obtain the first channel estimation result. The input of the LSTM is, for example, the channel estimation results of b slots from the (nb)th slot to the (n-1)th slot, and the output of the LSTM is, for example, the first channel estimation result of the nth slot. Optionally, the parameters of the LSTM may be obtained through training by using a neural network optimizer such as an Adam optimizer. In this case, b is a positive integer. For example, b may be equal to 2 or 4, or may be another value.

[0120] The first channel estimation result can be obtained in the three possible implementation forms mentioned above. Thus, the M channel estimation results may include one or more of the following: a channel estimation result obtained by performing a measurement based on a received downlink reference signal, a processing result obtained by processing an existing channel estimation result, or a channel estimation result obtained by prediction. The methods of determining different channel estimation results may be the same or different. This is not limited. For example, M1 channel estimation results among the M channel estimation results are obtained by performing a measurement based on a received downlink reference signal, and the other M-M1 channel estimation results are processing results obtained by processing the M1 channel estimation results. M1 is an integer greater than 0 and less than M. For simplicity of explanation, examples are not provided one by one.

[0121] S503: The UE performs joint compression on the M channel estimation results to obtain N compressed information, where N is a positive integer.

[0122] In this disclosure, the UE can compress the M channel estimation results together by using an encoder network. In consideration of this, the UE first needs to determine the encoder network to be used.

[0123] In the technology in which the CSI feedback is performed with reference to a neural network, a typical neural network architecture is a dual architecture. An autoencoder is used as an example. The compressed transmission may be implemented by joint optimization of the encoder and the decoder. For example, one or more groups of reference networks (or referred to as one or more reference networks) may be obtained by training, and the reference networks may be an encoder network (or referred to as a reference encoder network)-decoder network (or referred to as a reference decoder network) that appear in pairs. That is, the group of reference networks may include a reference encoder network and a corresponding reference decoder network. These reference networks may be trained offline or online. If the training is performed offline, the training may be specified in a protocol. For example, the protocol may provide parameters such as network structures and / or weights of one or more reference networks (including a reference encoder network and a reference decoder network). The UE or the access network device may implement the reference network with reference to the protocol. Optionally, the protocol may provide evaluation performance corresponding to a specific reference network in the agreed dataset.

[0124] When there are multiple groups of reference networks, the access network device may indicate a specific group of reference networks or a reference encoder network in a specific group of reference networks to the UE for use. For example, the access network device transmits first indication information to the UE, and the first indication information may indicate parameter information of the encoder network. The first indication information indicates the parameter information of the encoder network. In the indication manner, the first indication information includes the parameter information of the encoder network, and the parameter information of the encoder network includes, for example, parameters such as the structure and / or weight of the encoder network. The encoder network indicated by the first indication information is, for example, a reference encoder network, and the UE may determine the corresponding reference encoder network based on the first indication information. Alternatively, the first indication information may indicate an index of the encoder network (e.g., an index of the reference encoder network) or an index of the reference network, and the UE may also determine the corresponding reference encoder network based on the first indication information. After determining the reference encoder network, the UE may directly use the determined reference encoder network. In other words, the reference encoder network is the encoder network that is finally used by the UE. Alternatively, after determining the reference encoder network, the UE does not directly use the reference encoder network, but the characteristics of the encoder network that is ultimately used by the UE may be determined based on the characteristics of the reference encoder network. For example, the input dimension of the encoder network currently used by the UE is determined based on the input dimension of the reference encoder network. For example, the input dimension of the encoder network used by the UE is equal to the input dimension of the reference encoder network. Similarly, for example, the output dimension of the encoder network used by the UE is also determined based on the output dimension of the reference encoder network.For example, the output dimension of the encoder network used by the UE is the same as the output dimension of the reference encoder network. For example, if the input of the encoder network used by the UE is the same as the input of the reference encoder network, the difference between the output of the encoder network used by the UE and the output of the reference encoder network is less than a threshold.

[0125] Alternatively, the UE may select a specific reference network group or a reference encoder network in a specific reference network group by the UE. For example, the UE may select one of a plurality of groups of reference networks based on factors such as evaluation performance of the reference network, and the reference network includes a reference encoder network. For how the UE determines the encoder network to be used after the reference encoder network is determined, please refer to the description in the previous few paragraphs. When the UE determines the reference encoder network, the UE may send second indication information to the access network device, and the second indication information may indicate parameter information of the encoder network selected by the UE. If the encoder network finally used by the UE is not the same as the reference encoder network, the second indication information may indicate parameter information of the reference encoder network. The access network device may determine the reference network to which the reference encoder network belongs based on the parameter information of the reference encoder network to determine the reference decoder network included in the reference network. Alternatively, the second indication information may indicate parameter information of the reference network selected by the UE, and the access network device may determine the reference decoder network included in the reference network based on the parameter information of the reference network. Alternatively, the second indication information may indicate an index of the encoder network selected by the UE. If the encoder network finally used by the UE is not the same as the reference encoder network, the second indication information can indicate the index of the reference encoder network or the index of the reference network, so that the access network device can determine the corresponding reference decoder network. The second indication information needs to indicate the parameter information of the network. In the indication manner, the second indication information includes the parameter information of the network.For example, when the second instruction information indicates parameter information of a reference encoder network, the second instruction information may specifically include parameter information of the reference encoder network, for example, may include parameters such as the structure and / or weight of the encoder network. The second instruction information needs to indicate the index of the network. In the instruction manner, the second instruction information includes the index of the network. For example, when the second instruction information indicates the index of the reference encoder network, the second instruction information may specifically include the index of the reference encoder network.

[0126] Regardless of whether the access network device sends the first instruction information to the UE or the UE sends the second instruction information to the access network device, when the parameter information of the encoder network (or the reference encoder network or the reference network) is sent, in order to reduce transmission overhead, optionally, the sender (the access network device or the UE) can compress the parameter information of the encoder network (or the reference encoder network or the reference network), and the first instruction information or the second instruction information can include the compressed parameter information. The compression includes, but is not limited to, one or more of model pruning, model distillation, model quantization, and the like. Model pruning indicates that some parameter information of the encoder network (or the reference encoder network or the reference network) is sent to the UE, but the remaining parameter information is not sent. The UE obtains only some parameter information of the encoder network (or the reference encoder network or the reference network), and not all parameter information. This corresponds to pruning some parameters to reduce transmission overhead. Model distillation indicates that a first model is obtained based on the parameter information of the encoder network (or the reference encoder network or the reference network). The scale of the first model is smaller than the scale of the encoder network (or the reference encoder network or the reference network). It can be understood that the encoder network (or the reference encoder network or the reference network) is a large-scale model, and the encoder network (or the reference encoder network or the reference network) can have a large amount of parameters. However, the first model is a small-scale model, and the number of parameters corresponding to the first model may be small. In this case, the first instruction information or the second instruction information may include the first model (or may include the parameter information of the first model). In this way, the transmission overhead can also be reduced.After receiving the first model, the receiving side (UE or access network device) directly uses the first model. It can be understood that the receiving side uses the first model as the encoder network (or the reference encoder network or the reference network). Model quantization means that the parameter information of the encoder network (or the reference encoder network or the reference network) is, for example, a floating-point number, and the transmitting side can quantize the parameter information of the encoder network (or the reference encoder network or the reference network). The parameter information of the encoder network (or the reference encoder network or the reference network) included in the first instruction information or the second instruction information may be a quantization parameter to reduce transmission overhead.

[0127] When the UE determines the encoder network, an implementation form in which the UE performs joint compression on M channel estimation results is as follows: the UE inputs the M channel estimation results into the encoder network, and the encoder network can perform joint compression on the M channel estimation results. After the joint compression, the encoder network can output N pieces of compressed information. Since the N pieces of compressed information are obtained by performing joint compression on the M channel estimation results, each of the N pieces of compressed information corresponds to some or all of the M channel estimation results, and the channel estimation results corresponding to different pieces of compressed information may be the same, or different, or not completely the same. The different pieces of compressed information may correspond to the same parameters of the same channel estimation result, or the different pieces of compressed information may correspond to different parameters of the same channel estimation result, or the different pieces of compressed information may correspond to parameters that are not completely the same in the same channel estimation result. One piece of compressed information corresponding to one channel estimation result may be understood as the compressed information reflecting the characteristics of the channel estimation result, or the process of generating the compressed information being related to the channel estimation result, or the compressed information corresponding to all or some parameters of the channel estimation result.

[0128] For example, when M=3, M channel estimation results are channel estimation result 1, channel estimation result 2, and channel estimation result 3, respectively, and all three channel estimation results correspond to parameters A to E, and when N=3, N compressed information are compressed information 1, compressed information 2, and compressed information 3, respectively. Compressed information 1 can reflect parameters A and B corresponding to channel estimation result 1, parameters C and D corresponding to channel estimation result 2, and parameters A, D, and E corresponding to channel estimation result 3. Compressed information 2 can reflect parameters C and E corresponding to channel estimation result 1, parameters C and D corresponding to channel estimation result 2, and parameters B and C corresponding to channel estimation result 3. Compressed information 3 can reflect parameter D corresponding to channel estimation result 1, parameters A, B, and E corresponding to channel estimation result 2, and parameters A, B corresponding to channel estimation result 3. In this case, different pieces of compressed information correspond to the same channel estimation result, and parameters corresponding to different pieces of compressed information include several cases below. Different pieces of compressed information may correspond to the same parameters of the same channel estimation result (e.g., compressed information 1 corresponds to parameters C and D of channel estimation result 2, and compressed information 2 also corresponds to parameters C and D of channel estimation result 2). Different pieces of compressed information may correspond to different parameters of the same channel estimation result (e.g., compressed information 1 corresponds to parameters A and B of channel estimation result 1, and compressed information 2 corresponds to parameters C and E of channel estimation result 1). Different pieces of compressed information may also correspond to parameters that are not completely the same in the same channel estimation result (e.g., compressed information 1 corresponds to parameters A, D, E of channel estimation result 3, and compressed information 3 corresponds to parameters A, B of channel estimation result 3). Also, in this example, for example, the parameters corresponding to M channel estimation results are all the same (i.e., all three channel estimation results correspond to parameters A to E). In practice, the parameters corresponding to different channel estimation results may be different or may not be completely the same. This is not limited here.Also, in this example, for example, the parameters corresponding to N pieces of compressed information are all the same.In practice, the parameters corresponding to different pieces of compressed information may also be different or may not be completely the same.For example, N=2, compressed information 1 may correspond to M parameters A of M channel estimation results, and compressed information 2 no longer corresponds to parameter A.

[0129] Optionally, in order to ensure that the downlink channel restored by the access network device is more accurate, when performing compression, the UE can not only perform joint compression on the M channel estimation results, but also perform processing on the history information. In other words, the UE can perform joint compression on the M channel estimation results and the history information to obtain N compressed information. The history information includes, for example, channel estimation results corresponding to L slots M slots before. For example, the history information includes L channel estimation results. L is a positive integer, and the L slots may be consecutive or separate. For example, the L slots are configured by using configuration information. For example, the number of the L slots can be configured by using the configuration information. Alternatively, the L slots may be determined by the UE. For a method in which the UE obtains the history information, please refer to the aforementioned method in which the UE obtains the M channel estimation results. For example, the history information includes the L channel estimation results. For example, the UE can input both the M channel estimation results and the L channel estimation results to an encoder network, and the encoder network can perform joint compression on the M channel estimation results and the L channel estimation results. After the joint compression, the encoder network can output N pieces of compressed information. The N pieces of compressed information are obtained by performing joint compression on the M channel estimation results and the L channel estimation results, so that each of the N pieces of compressed information corresponds to some or all of the M channel estimation results and the L channel estimation results, and the channel estimation results corresponding to different pieces of compressed information may be the same, or different, or not completely the same. The different pieces of compressed information may correspond to the same parameters of the same channel estimation result, or the different pieces of compressed information may correspond to different parameters of the same channel estimation result, or the different pieces of compressed information may correspond to parameters that are not completely the same in the same channel estimation result. Please refer to the above example for this example.For example, any one or two channel estimation results in the above examples are considered as channel estimation results included in the historical information.

[0130] For example, the value of N may be N = 1. Specifically, the UE may obtain one piece of compressed information based on the M channel estimation results (or based on the M channel estimation results and the L channel estimation results).

[0131] In an implementation of the channel estimation result, the channel estimation result is channel information, and the UE can obtain one compressed information based on the M pieces of channel information (or based on the M pieces of channel information and the L pieces of channel information). The channel information may also be called a channel response, a channel matrix, etc. The dimension of the channel information is, for example, [N tx ,N rx ,N RB ], where N tx represents the number of antennas or ports on the transmitting side (e.g., an access network device), and N rx represents the number of antennas or ports at the receiving end (e.g., UE), and N RB represents the number of frequency domain units, e.g., the number of resource blocks (RB), and N tx , N rx , and N RB is a positive integer. The dimension reduction is performed by using singular value decomposition (SVD) to obtain the eigensubspace matrices of the downlink channel [N tx ,N rx ,N RB ], and the eigensubspace matrices have dimensions [N tx ,N sb ]. N sb is the number of frequency domain subbands, e.g., N sb =N RB / a, where a represents the frequency domain subband granularity or subband bandwidth, i.e., the number of RBs included in each subband. Common granularities of frequency domain subbands are 1RB, 2RB, or 4RB. When the granularity is 4RB, N sb =N RB / 4. One RB includes a positive integer number of subcarriers, for example, 12 or 16 subcarriers. When performing dimensionality reduction of channel information by using SVD, the UE can process different ranks, and different ranks can also be understood as different streams or different layers. One piece of channel information (or one channel estimation result) can correspond to one or more layers. The following describes the processing process of the UE of the L'th layer, where L' is a positive integer.

[0132] If each subband of the L'th layer includes an RB, the UE can calculate an equivalent downlink channel based on the downlink channel of the RB. i Assuming that , the equivalent downlink channel in a subband can be expressed as:

number

[0133] UE is

number

number

[0134] That is,

number

[0135] In formulas 2 and 3, Hi The dimension of is [N tx ,N rx ], where H i H H i represents the conjugate transpose of

number

number

number

[0136] The UE combines the eigenvectors of all subbands in the L'th layer to obtain the eigensubspace matrix V = [V1V2...V Nsb ], where V is [N tx ,N sb ]. The two groups of DFT bases are the discrete fourier transform (DFT) or spatial bases

number

number

number

[0137] S H is the hermitian matrix of S, also called the self-conjugate matrix, which can be obtained by performing a conjugate transpose on the matrix S, where S represents the spatial basis and the specific form of S is related to the antenna panel. Assuming that the antenna panel is dual polarized, with a horizontal component of Nh and a vertical component of Nv, the representation form of S is obtained as follows:

number

[0138] F represents the frequency basis, and the expression format of F is subband N sb For example, F can satisfy the following equation: F = DFT(N sb ) (Equation 6)

[0139] Optionally, an oversampling factor may be further added in the DFT process. For example, a group of multiple orthogonal spatial bases {S1, S2, S3...} and a group of multiple orthogonal frequency bases {F1, F2, F3...} may be generated in an oversampling manner, and S i and F j A set of the above is selected from the group as the spatial and frequency bases of the present disclosure. For example, a group with a more accurate projection direction may be selected. For example, the oversampling factor of both the spatial and frequency domains is 4.

[0140] The complex matrix C obtained by performing the above operations complexis a sparse representation of the eigensubspace of the original channel, and the dimensions of the complex matrix match those of the eigensubspace matrix before projection, [N tx ,N sb ]. A channel estimation result can correspond to one or more layers, and processing one layer produces a complex matrix C complex Then, multiple complex matrices C can be obtained based on one channel estimation result. complex The UE can obtain all the complex matrices C corresponding to the M channel estimation results (or the M channel estimation results and the L channel estimation results). complex to the encoder network, or the UE may input each complex matrix C corresponding to the M channel estimation results (or the M channel estimation results and the L channel estimation results) complex One may perform corresponding processing on C (e.g., converting the complex matrix to a 3D real-valued tensor) and then input the complex matrix to the encoder network. complex The process of converting to a 3D real-valued tensor is described below.

[0141] In another implementation of the channel estimation result, the channel estimation result is an eigenvector, and the UE may obtain one compressed information based on the M eigenvectors (or based on the M eigenvectors and the L eigenvectors). When the channel estimation result is an eigenvector, each layer is an eigensubspace matrix V=[V1V2...V Nsb ] and V has [N tx ,N sb ] is a complex matrix having a dimension of . The UE can perform data pre-processing on the eigensubspace matrix in each layer. For example, the calculation process for performing data pre-processing on the eigensubspace matrix of a layer is as follows:

[0142] UE is the DFT formula: spatial basis

number

number

[0143] Optionally, an oversampling factor may be further added in the DFT process, see above for details.

[0144] The UE calculates the complex matrix C based on one channel estimation result. complex Based on the M channel estimation results, M complex matrices C complex If there are still L channel estimation results, the UE may obtain L complex matrices C complex We can further obtain the complex matrix C complex After obtaining the complex matrix C, the UE complex may be directly input to the encoder network, or the UE may perform corresponding processing on the obtained complex matrix and then input the obtained complex matrix to the encoder network. For example, in the processing method, the UE complex To convert the complex matrix C into a 3D real-valued tensor, we first input the resulting 3D real-valued tensor into the encoder network. complex In this case, the UE tx ,N sb To obtain a real matrix with dimensions complex Similarly, UE extracts the real part of each element of [N tx ,N sbTo obtain a real matrix with dimensions complex In addition, the UE can extract the imaginary part of each element in . tx ,N sb ], i.e., H[0]=real(C complex ), H[1]=imag(C complex ), where the tensor H is a 3D real-valued tensor.

[0145] The above process of determining the channel estimation result is only used as an example. In the present disclosure, the channel estimation may be performed in other possible ways, which is not limited.

[0146] For example, FIG. 6A shows an example of an input / output scenario of an encoder network. FIG. 6A uses an example in which the input information of the encoder network does not include history information. In FIG. 6A,

number

[0147] Optionally,

number

number

[0148] The finally output compressed information c is a [C, 1]-dimensional real vector, where C represents the length of the compressed information, and can be obtained based on the requirement of feedback overhead.

[0149] Further, FIG. 6C shows an example of an input and output scenario of the encoder network. FIG. 6C uses an example where the input information of the encoder network includes history information. In FIG. 6C,

number

number

[0150] Optionally,

number

number

[0151] The compressed information c that is finally output is a [C, 1]-dimensional real vector.

[0152] In this case, the UE can obtain one compressed information by using an encoder network. Specifically, for a downlink channel in M ​​slots, the UE can represent the downlink channel by using one compressed information. The UE only needs to transmit one compressed information to the access network device. In addition, the compressed information can reflect the characteristics of the downlink channel in M ​​slots, so that the UE does not need to transmit the compressed information in each slot. In this way, the transmission overhead can be greatly reduced. In addition, the compressed information is obtained by performing joint compression on the M channel estimation results (or the M channel estimation results and the L channel estimation results), and can reflect the state of the downlink channel in the M slots. In addition, the downlink channel is correlated in the time domain. The compressed information can better reflect the related situation of the downlink channel in the time domain, so that the access network device can restore a more accurate downlink channel based on the compressed information.

[0153] In another example, the value of N may be N>1. Specifically, the UE can obtain a number of pieces of compressed information based on M channel estimation results (or based on M channel estimation results and L channel estimation results). When N is greater than 1, N may be equal to M, or may be smaller than M, or may be greater than M. It can be seen that the quantity of channel estimation results input by the encoder network may be equal to or not equal to the quantity of a number of pieces of compressed information output by the encoder network, so that the channel compression process becomes more flexible. Even when N is greater than 1, each of the N pieces of compressed information corresponds to M channel estimation results, or corresponds to M channel estimation results and L channel estimation results. The value of N is related to the change speed of the downlink channel. In a low-speed scenario (e.g., in the 3GPP channel model, the indoor speed is 3km / h (80%) and the outdoor speed is 30km / h (20%)), N=M=10 is a suitable value. In this case, the M consecutive channel estimation results change slightly, and the change rule of the M consecutive channel estimation results has been learned by the network for channel compression. When N=M, the network structure is better designed.

[0154] For example, FIG. 6E shows an example of an input / output scenario of an encoder network. FIG. 6E uses an example where the input information of the encoder network does not include history information. In FIG. 6E,

number

number

[0155] Optionally,

number

number

number

[0156]

number

number

[0157] Further, FIG. 6G shows an example of an input and output scenario of the encoder network. FIG. 6G uses an example where the input information of the encoder network includes history information. In FIG. 6G,

number

number

number

[0158] Optionally,

number

number

number

[0159]

number

number

[0160] In this case, the N pieces of compressed information are obtained by performing joint compression on the M channel estimation results (or the M channel estimation results and the L channel estimation results), and can reflect the state of the downlink channel in the M slots. In addition, the downlink channel is correlated in the time domain. The compressed information can better reflect the related situation of the downlink channel in the time domain, so that the access network device can restore a more accurate downlink channel based on the compressed information. In addition, since the joint compression is performed on the M channel estimation results (or the M channel estimation results and the L channel estimation results), each piece of compressed information does not need to represent all parameters of the channel estimation result involved in the compression. For example, the N pieces of compressed information can respectively correspond to different parameters of the M channel estimation results (or the M channel estimation results and the L channel estimation results), so that the amount of information of the N pieces of compressed information can be reduced to reduce the transmission overhead. Also, in this case, the input information of the encoder network includes multiple channel estimation results, and the output information includes multiple pieces of compressed information. Compared with the solution in which a single piece of compressed information is output, the multiple-input multiple-output solution can simplify the structure of the encoder network.

[0161] S504: The UE sends the N pieces of compressed information to the access network device. In response, the access network device receives the N pieces of compressed information from the UE. The N pieces of compressed information may be regarded as CSI or PMI.

[0162] FIG. 7 is a diagram of a CSI feedback process. In FIG. 7, the access network device transmits configuration information to the UE and then transmits a downlink reference signal to the UE. For example, the downlink reference signal is CSI-RS. The UE can transmit compressed information to the access network device multiple times. For example, after performing channel estimation on the downlink channel of M slots, the UE can transmit N pieces of compressed information to the access network device. Then, the UE continues to perform channel estimation on the downlink channel in M ​​slots, and then transmits the obtained compressed information to the access network device. In FIG. 7, the UE transmits compressed information to the access network device twice or feeds back CSI twice. The CSI fed back by the UE to the access network device for the first time may include one or more of the following channel estimation results: a channel estimation result obtained by the UE based on the first received CSI-RS (the first CSI-RS from left to right in FIG. 7), a channel estimation result corresponding to a downlink channel before the downlink channel corresponding to the first CSI-RS, or a channel estimation result corresponding to a downlink channel after the downlink channel corresponding to the first CSI-RS. The slot in which the UE feeds back CSI to the access network device for the first time is later than the slot in which the UE receives CSI-RS for the first time (or later than the slot in which the access network device delivers CSI-RS for the first time). The slot in which the UE feeds back CSI to the access network device for the first time may be earlier or later than the slot in which the UE receives CSI-RS for the second time (or earlier or later than the slot in which the access network device delivers CSI-RS for the second time). FIG. 7 uses an earlier slot than the slot in which the UE receives CSI-RS for the second time as an example.Similarly, the CSI fed back by the UE to the access network device at the kth time may include one or more of the following channel estimation results: a channel estimation result obtained by the UE based on the kth received CSI-RS, a channel estimation result corresponding to a downlink channel before the downlink channel corresponding to the kth received CSI-RS, or a channel estimation result corresponding to a downlink channel after the downlink channel corresponding to the kth received CSI-RS. The slot in which the UE feeds back the CSI to the access network device at the kth time is later than the slot in which the UE receives the kth CSI-RS (or later than the slot in which the access network device delivers the kth CSI-RS). The slot in which the UE feeds back the CSI to the access network device at the kth time may be earlier or later than the slot in which the UE receives the (k+1)th CSI-RS (or earlier or later than the slot in which the access network device delivers the (k+1)th CSI-RS). In addition, it can be seen from FIG. 7 that the M slots indicated by the configuration information may include the moment the configuration information is received, and may further include a moment before the configuration information is received and a moment after the configuration information is received.

[0163] S505: The access network device restores the N pieces of compressed information to obtain K pieces of restored information. The K pieces of restored information are information about the downlink channel in M ​​slots, that is, the K pieces of restored information can represent the downlink channel in M ​​slots. K is a positive integer. For example, K may be equal to 1 or may be greater than 1. If K is greater than 1, K may be equal to N, or may be greater than N, or may be less than N. K may be equal to M, or may be greater than M, or may be less than M. K may be equal to M+L, or may be greater than M+L, or may be less than M+L. That is, the quantity of the channel estimation result input by the UE to the encoder network may be equal to or may not be equal to the quantity of the plurality of response information output by the decoder network. This is more flexible.

[0164] In the present disclosure, the UE can perform joint compression on multiple channel estimation results by using an encoder network, and the access network device can restore N pieces of compressed information by using a corresponding decoder network. Therefore, the access network device needs to first determine the decoder network to be used. If the access network device sends first indication information to the UE to indicate a reference encoder network, the access network device can determine a reference decoder network corresponding to the reference encoder network. Alternatively, if the UE sends second indication information to the access network device to indicate a reference encoder network, the access network device can determine a reference decoder network corresponding to the reference encoder network. Alternatively, if the UE sends second indication information to the access network device to indicate a reference network, the access network device can determine a reference decoder network included in the reference network.

[0165] After determining the reference decoder network, the access network device can directly use the determined reference decoder network. In other words, the reference decoder network is the decoder network that is ultimately used by the access network device. Alternatively, after determining the reference decoder network, the access network device does not directly use the reference decoder network, but the characteristics of the decoder network that is ultimately used by the access network device may be determined based on the characteristics of the reference decoder network. For example, the input dimension of the decoder network used by the access network device may be determined based on the input dimension of the reference decoder network. For example, the input dimension of the decoder network used by the access network device is equal to the input dimension of the reference decoder network of the reference decoder network. Similarly, for example, the output dimension of the decoder network used by the access network device may also be determined based on the output dimension of the reference decoder network. For example, the output dimension of the decoder network used by the access network device is equal to the output dimension of the reference decoder network. For example, if the input of the decoder network used by the access network device is the same as the input of the reference decoder network, the difference between the output of the decoder network used by the access network device and the output of the reference decoder network is less than a threshold value.

[0166] After determining the decoder network, the access network device can restore the N pieces of compressed information based on the decoder network. For example, the access network device inputs the N pieces of compressed information into the decoder network, and the decoder network outputs K pieces of restored information. Each of the K pieces of restored information corresponds to a part or all of the N pieces of compressed information. The compressed information corresponding to different pieces of restored information may be the same, or different, or not completely the same. The different pieces of restored information may correspond to the same parameters of the same compressed information, or the different pieces of restored information may correspond to different parameters of the same compressed information, or the different pieces of restored information may correspond to parameters that are not completely the same of the same compressed information. One piece of restored information corresponds to one piece of compressed information may be understood as the restored information reflecting the characteristics of the compressed information, or the process of generating the restored information being related to the compressed information, or the restored information corresponding to all or some parameters of the compressed information.

[0167] For example, when N=3, the N pieces of compression information are compression information 1, compression information 2, and compression information 3, and all three pieces of compression information correspond to parameters A to E; when K=3, the K pieces of restoration information are restoration information 1, restoration information 2, and restoration information 3. The restoration information 1 may reflect parameters A and B corresponding to the compression information 1, parameters A and C corresponding to the compression information 2, and parameters D and E corresponding to the compression information 3. The restoration information 2 may reflect parameters A, B, and C corresponding to the compression information 1, parameters A and C corresponding to the compression information 2, and parameters C and E corresponding to the compression information 3. The restoration information 3 may reflect parameters D and E corresponding to the compression information 1, parameters B, D, and E corresponding to the compression information 2, and parameters A and B corresponding to the compression information 3. In this case, different pieces of restoration information correspond to the same pieces of compression information, and the parameters corresponding to different pieces of restoration information include several cases below. Different pieces of restoration information may correspond to the same parameters of the same compressed information (for example, restoration information 1 corresponds to parameters A and C of compressed information 2, and restoration information 2 also corresponds to parameters A and C of compressed information 1). Different pieces of restoration information may correspond to different parameters of the same compressed information (for example, restoration information 1 corresponds to parameters A and B of compressed information 1, and restoration information 3 corresponds to parameters D and E of compressed information 1). Different pieces of restoration information may also correspond to parameters that are not completely the same of the same compressed information (for example, restoration information 1 corresponds to parameters A and B of compressed information 1, and restoration information 2 corresponds to parameters A, B, and C of compressed information 1). In this example, for example, the parameters corresponding to N pieces of compression information are all the same (i.e., all three pieces of compression information correspond to parameters A to E). In practice, the parameters corresponding to different pieces of compression information may be different or may not be completely the same. This is not limited here. Also, in this example, for example, the parameters corresponding to K pieces of restoration information are all the same. In practice, the parameters corresponding to different pieces of reconstruction information may also be different or may not be exactly the same.For example, with K=2, reconstruction information 1 may correspond to N compressed information of M parameters A, and reconstruction information 2 no longer corresponds to parameter A.

[0168] FIG. 8A shows an example of input / output information of the decoder network when N=1. In FIG. 8A, c represents the compressed information input by the decoder network;

number

[0169] The compressed information input by the network is represented as above. Optionally, c is

number

number

[0170] Also, FIG. 8C shows an example of input / output information of the decoder network for N>1. In FIG. 8C,

number

number

[0171] The compressed information input by the network is represented as above. Optionally,

number

number

number

[0172] The K pieces of reconstruction information obtained by the decoder network can be considered as CSI, or PMI, or information having a function similar to CSI or PMI. Optionally, the decoder network may further process the K pieces of reconstruction information to obtain reconstructed downlink channel information. In other words, the access network device can obtain reconstructed downlink channel information based on the K pieces of reconstruction information. The downlink channel information is, for example, a downlink channel matrix, or a parameter of the downlink channel matrix (e.g., a parameter such as a weight of the downlink channel matrix).

[0173] Sampling points corresponding to K pieces of reconstruction information are arranged in the M slots, and sampling points corresponding to N pieces of compressed information (or sampling points corresponding to M pieces of channel estimation results) are also located in the M slots. However, the number of sampling points corresponding to the K pieces of reconstruction information may or may not be equal to the number of sampling points corresponding to the N pieces of compressed information.

[0174] In addition, the time domain positions of the sampling points corresponding to the K pieces of reconstruction information may be the same as, different from, or not completely the same as the time domain positions of the sampling points corresponding to the N pieces of compressed information. For example, the K pieces of reconstruction information correspond to three sampling points, and the time domain positions are T1, T2, and T3, respectively. The N pieces of compressed information correspond to three sampling points, and the time domain positions are T1, T4, and T3, respectively. It can be seen that the time domain positions of the sampling points corresponding to the K pieces of reconstruction information are not completely the same as the time domain positions of the sampling points corresponding to the N pieces of compressed information. In another example, the K pieces of reconstruction information correspond to two sampling points, and the time domain positions are T1 and T2, respectively. The N pieces of compressed information correspond to three sampling points, and the time domain positions are T4, T5, and T6, respectively. It can be seen that the time domain positions of the sampling points corresponding to the K pieces of reconstruction information are different from the time domain positions of the sampling points corresponding to the N pieces of compressed information. In another example, the K pieces of reconstruction information correspond to three sampling points, and the time domain positions are T1, T2, and T3, respectively. The N pieces of compressed information correspond to three sampling points, and the time domain positions are T1, T2, and T3, respectively. It can be seen that the time domain positions of the sampling points corresponding to the K pieces of restored information are the same as the time domain positions of the sampling points corresponding to the N pieces of compressed information.

[0175] In other words, the time domain positions of the sampling points corresponding to the K pieces of restoration information are located within the M slots, and the quantity of the sampling points corresponding to the K pieces of restoration information, the time domain positions within the M slots, etc. are not limited. In this way, for an access network device, the restoration process can be flexible.

[0176] Optionally, after obtaining the downlink channel information based on the K pieces of restoration information, the access network device may determine, based on the downlink channel information, one or more pieces of information, such as a quantity of streams used when data is transmitted to the UE, a modulation order used when data is transmitted to the UE, or a code rate of a channel carrying data (e.g., a physical downlink control channel (PDCCH)). In addition, the access network device may further determine, based on the downlink channel information, a precoding matrix used when data is transmitted to the UE, etc. Alternatively, the access network device does not restore the downlink channel information based on the K pieces of restoration information, but directly uses the K pieces of restoration information. In this case, the access network device may determine, based on the K pieces of restoration information, one or more pieces of information, such as a quantity of streams used when data is transmitted to the UE, a modulation order used when data is transmitted to the UE, or a coding rate of a channel carrying data. In addition, the access network device may further determine, based on the K pieces of restoration information, a precoding matrix used when data is transmitted to the UE, etc.

[0177] In the present disclosure, the UE may perform channel estimation on the downlink channel in M ​​slots, perform joint compression on the obtained channel estimation results, and report the channel estimation results to the access network device. Since the joint compression is performed, the correlation between the downlink channels in different slots is fully utilized, and the access network device may restore the N pieces of compressed information obtained by the joint compression, so as to obtain more accurate and effective downlink channel information. Also, since the joint compression is performed, the N pieces of compressed information may be complemented with each other. For example, different pieces of compressed information may correspond to different parameters to reduce redundant information and correspondingly reduce transmission overhead. In addition, during the joint compression, in addition to the M channel estimation results, history information may be further taken into account, so that the compressed information may reflect the characteristics of the downlink channel in more slots, and the access network device may restore a more accurate downlink channel matrix.

[0178] The communication device provided in the present disclosure is described according to the embodiments of the aforementioned method.

[0179] The present disclosure provides a communication device. The communication device includes, for example, a processing unit and a transceiver unit (also referred to as a communication unit). The processing unit may be configured to implement the processing functions of the UE in the embodiment shown in Figure 5, and the transceiver unit may be configured to implement all or some of the transmission and reception functions of the UE in the embodiment shown in Figure 5. Alternatively, the processing unit may be configured to implement the processing functions performed by the access network device in the embodiment shown in Figure 5, and the transceiver unit may be configured to implement all or some of the transmission and reception functions of the access network device in the embodiment shown in Figure 5.

[0180] Optionally, the processing unit and / or the transceiver unit may be implemented by using a virtual module. For example, the processing unit may be implemented by using a software functional unit or a virtual device, and the transceiver unit may be implemented by using a software functional unit or a virtual device. Alternatively, the processing unit and / or the transceiver unit may be implemented by using a physical device (e.g., a circuit system and / or a processor). The following describes the case where the processing unit and the transceiver unit are implemented by using a physical device.

[0181] FIG. 9 is a diagram of the structure of a communication device according to the present disclosure. The communication device 900 may be a UE in the embodiment shown in FIG. 5, a circuit system of the UE, a circuit system that can be used in the UE, etc., configured to implement a method corresponding to the UE in the above-mentioned method embodiment. Alternatively, the communication device 900 may be an access network device in the embodiment shown in FIG. 5, a circuit system of the access network device, a circuit system that can be used in the access network device, etc., configured to implement a method corresponding to the access network device in the above-mentioned method embodiment. For specific functions, please refer to the description of the above-mentioned method embodiment. For example, the circuit system is a chip system.

[0182] The communication device 900 includes one or more processors 901. The processor 901 may implement certain control functions. The processor 901 may be a general-purpose processor or a dedicated processor, etc. For example, the processor 901 includes a baseband processor, a central processing unit, etc. The baseband processor may be configured to process communication protocols and communication data. The central processing unit may be configured to control the communication device 900, execute software programs, and / or process data. The different processors may be independent components or may be located in one or more processing circuits, for example integrated in one or more application specific integrated circuits.

[0183] Optionally, the communication device 900 includes one or more memories 902 configured to store instructions 904. The instructions 904 may be executed by a processor, thereby enabling the communication device 900 to perform the methods described in the method embodiments above. Optionally, the memory 902 may further store data. The processor and the memory may be located separately or may be integrated.

[0184] Optionally, the communication device 900 may store instructions 903 (which may sometimes be referred to as code or programs) that may be executed on the processor to enable the communication device 900 to perform the methods described in the previous embodiments. The processor 901 may store data.

[0185] For example, the processing unit may be implemented using one or more processors 901, or the processing unit may be implemented using one or more processors 901 and one or more memories 902, or the processing unit may be implemented using one or more processors 901, one or more memories 902, and instructions 903.

[0186] Optionally, the communications device 900 may further include a transceiver 905 and an antenna 906. The transceiver 905 may also be referred to as a transceiver unit, a transceiver machine, a transceiver circuit, a transceiver machine, an input / output interface, etc., and is configured to implement a transceiver function of the communications device 900 by using the antenna 906. For example, the transceiver unit is implemented by using the transceiver 905, or the transceiver unit is implemented by using the transceiver 905 and the antenna 906.

[0187] Optionally, the communication device 900 may further include any one or more of the following components: a wireless communication module, an audio module, an external memory interface, an internal memory, a universal serial bus (USB) interface, a power management module, an antenna, a speaker, a microphone, an input / output module, a sensor module, a motor, a camera, a display, etc. It will be understood that in some embodiments, the communication device 900 may include more or fewer components, or some components are integrated, or some components are split. The components may be implemented by hardware, software, or a combination of software and hardware.

[0188] The processor 901 and transceiver 905 described in this disclosure may be implemented in an integrated circuit (IC), an analog IC, a radio frequency integrated circuit (RFID), a mixed signal IC, an application specific integrated circuit (ASIC), a printed circuit board (PCB), an electronic device, etc. The communication device described herein may be an independent device (e.g., an independent integrated circuit or a mobile phone) or may be part of a larger device (e.g., a module that can be incorporated into another device). For details, please refer to the description of the UE and the access network device in the preceding embodiment. Details will not be repeated here.

[0189] The present disclosure provides a terminal device, which may be used in the aforementioned embodiments. The terminal device includes corresponding means, units, and / or circuits for implementing the functions of the UE in the embodiment shown in Figure 5. For example, the terminal device includes a transceiver module (or referred to as a transceiver unit) configured to support the terminal device in implementing transceiver functions, and a processing module (or referred to as a processing unit) configured to support the terminal device in processing signals.

[0190] The present disclosure further provides an access network device, which may be used in the aforementioned embodiments. The access network device includes corresponding means, units, and / or circuits for implementing the functions of the access network device in the embodiment shown in Figure 5. For example, the access network device includes a transceiver module (or referred to as a transceiver unit) configured to support the access network device in implementing a transmission function, and a processing module (or referred to as a processing unit) configured to support the access network device in processing signals.

[0191] All or part of the technical solutions provided in the present disclosure may be implemented by using software, hardware, firmware, or any combination thereof. When implementing the embodiments using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded into a computer and executed, the procedures or functions according to the present disclosure are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, an access network device, a terminal device, an AI node, or another programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from a computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired (e.g., coaxial cable, optical fiber, or digital subscriber line (DSL)) or wireless (e.g., infrared, radio wave, or microwave) manner. A computer-readable storage medium may be any available medium accessible by a computer, or may be a data storage device incorporating one or more available media, such as a server or a data center. The available media may be magnetic media (e.g., floppy disks, hard disk drives, magnetic tapes), optical media (e.g., digital video discs (DVDs)), semiconductor media, etc.

[0192] The above description is merely a specific implementation of the present disclosure, and is not intended to limit the scope of protection of the present disclosure. Any variations or replacements that can be easily imagined by those skilled in the art within the technical scope disclosed in the present disclosure shall fall within the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure shall be subject to the scope of protection of the claims. [Explanation of symbols]

[0193] 10. Communication Systems 20 RAN Devices, Access Network Devices 30 Communication Equipment 900 Communication Equipment 901 Processor 902 Memory 903 Instructions 904 Command 905 Transceiver 906 Antenna

Claims

1. obtaining M channel estimation results corresponding to M time units, where M is an integer greater than 1; performing joint compression on the M channel estimation results by using an encoder network to obtain N compressed information, where N is a positive integer; transmitting the N pieces of compressed information to an access network device; A communication method comprising:

2. The plurality of channel estimation results includes a first channel estimation result, the first channel estimation result being: a channel estimation result obtained by performing measurements based on a received downlink reference signal; a processed result obtained by processing a channel estimation result obtained by performing measurements based on a received downlink reference signal; or Channel estimation results obtained by prediction The method of claim 1, wherein the

3. The method according to claim 1 or 2, wherein each of the N compressed information corresponds to the M channel estimation results.

4. performing joint compression on the M channel estimation results by using an encoder network to obtain N compressed information; performing joint compression on the M channel estimation results and history information by using the encoder network to obtain the N compressed information, the history information including channel estimation results corresponding to a time unit prior to the M time units; 3. The method of claim 1 or 2, comprising:

5. The method of claim 4 , wherein each of the N pieces of compressed information corresponds to the M channel estimation results and the history information.

6. The method comprises: receiving configuration information from the access network device, the configuration information being used to configure the M time units; 6. The method of claim 1, further comprising:

7. The configuration information includes the following: a start time-domain position of said M time units; an end time domain position of said M time units; the duration of said M time units, the number of said M time units; the time domain location of the first sampling point within the M slots; a sampling period of the M time units; the number of sampling points for the M time units; or Time domain positions of the sampling points in the M time units The method of claim 6, comprising one or more of:

8. The method comprises: receiving first indication information from the access network device, the first indication information indicating parameter information of a reference encoder network or an index of a reference encoder network; or sending second indication information to the access network device, the second indication information indicating parameter information of a reference encoder network or an index of a reference encoder network; Further comprising: the reference encoder network is used to determine the encoder network; 8. The method according to any one of claims 1 to 7.

9. receiving N pieces of compressed information from a terminal device, where N is a positive integer; using a decoder network to recover the N pieces of compressed information to obtain K pieces of recovered information, where the K pieces of recovered information are M pieces of downlink channel information of time units, where K is a positive integer and M is an integer greater than 1; A communication method comprising:

10. The method of claim 9 , wherein each of the K pieces of reconstruction information corresponds to the N pieces of compression information.

11. The method comprises: transmitting configuration information to the terminal device, the configuration information being used to configure the M time units; 11. The method of claim 9 or 10, further comprising:

12. The configuration information includes the following: a start time-domain position of said M time units; an end time domain position of said M time units; the duration of said M time units, the number of said M time units; the time domain location of the first sampling point within the M slots; a sampling period of the M time units; the number of sampling points for the M time units; or The method of claim 11 , comprising one or more of the time domain locations of sampling points in the M time units.

13. The method comprises: sending first indication information to the terminal device, the first indication information indicating parameter information of a reference encoder network or an index of a reference encoder network; or receiving second indication information from the terminal device, the second indication information indicating parameter information of a reference encoder network selected by the terminal device or an index of a reference encoder network selected by the terminal device; Further comprising: The reference encoder network and the reference decoder network belong to the same reference network, and the reference decoder network is used to determine the decoder network.

13. The method according to any one of claims 9 to 12.

14. A communication device configured to implement the method according to any one of claims 1 to 8 or the method according to any one of claims 9 to 13.

15. A communications device comprising a processor and a memory, the memory being connected to the processor, the processor being configured to execute a method according to any one of claims 1 to 8 or any one of claims 9 to 13.

16. A computer readable storage medium, the computer readable storage medium configured to store a computer program, the computer program being enabled, when run on a computer, to perform the method according to any one of claims 1 to 8, or the computer being enabled to perform the method according to any one of claims 9 to 13.

17. A communication system comprising an apparatus for carrying out the method according to any one of claims 1 to 8 and an apparatus for carrying out the method according to any one of claims 9 to 13.

Images