Method and apparatus for beamforming in multiple input multiple output system

Abstract

Embodiments of the present disclosure provide a communication method, a base station, a user equipment, a storage medium, and a program product, which relate to the field of 5G or 6G communication systems, artificial intelligence or the like, and are used for supporting a higher data rate than that of 4G communication systems such as long term evolution (LTE). The method comprises: acquiring wireless channel information, the wireless channel information comprising first information related to losses of candidate BF algorithms, the first information being determined based on a PMI, singular values and singular vectors, the singular values and the singular vectors being determined based on an SRS measurement related channel matrix; determining a first BF algorithm from at least two candidate BF algorithms based on the wireless channel information by using a first AI model; determining a first BF weight based on the first BF algorithm; and transmitting downlink data based on the first BF weight. The embodiments of the present disclosure can maximize the spectral efficiency in any wireless channel environment. Optionally, the method performed by an electronic device can be performed using an artificial intelligence model.

Description

METHOD AND APPARATUS FOR BEAMFORMING IN MULTIPLE INPUT MULTIPLE OUTPUT SYSTEM

[0001] The present disclosure relates to the technical field of wireless communication, and in particular to a communication method, a base station, a user equipment (UE), a storage medium, and a program product.

[0002]

[0003] Considering the development of wireless communication from generation to generation, the technologies have been developed mainly for services targeting humans, such as voice calls, multimedia services, and data services. Following the commercialization of 5th-generation (5G) communication systems, it is expected that the number of connected devices will exponentially grow. Increasingly, these will be connected to communication networks. Examples of connected things may include vehicles, robots, drones, home appliances, displays, smart sensors connected to various infrastructures, construction machines, and factory equipment. Mobile devices are expected to evolve in various form-factors, such as augmented reality glasses, virtual reality headsets, and hologram devices. In order to provide various services by connecting hundreds of billions of devices and things in the 6th-generation (6G) era, there have been ongoing efforts to develop improved 6G communication systems.

[0004] 6G communication systems, which are expected to be commercialized around 2030, have various significantly improved metrics compared to the current 5G communication systems. The peak data rate will reach at least 50 Gbit / s, and the user experienced data rate will reach at least 300 Mbit / s, the air-interface latency will be less than 1 ms, and the air-interface reliability will reach . In addition to the above basic communication metrics, the 6G communication systems will also have sensing capabilities, AI-related capabilities, better security, better interoperability and better sustainability.

[0005] In order for the 6G communication systems to fulfill the above metrics, more advanced air-interface technologies and network technologies need to be developed. The evolution of extreme Multiple Input Multiple Output (extreme MIMO) has been already under consideration, including the use of ultra-large scale antenna arrays, the development and evolution of distributed antenna systems, and the design of MIMO air-interface algorithms assisted by Artificial Intelligence (AI). This technology enables higher spectral efficiency, greater coverage, and precise localization and sensing capabilities. Additionally, for technologies that contribute to improve high-frequency band coverage, including metamaterial-based lenses and antennas, new antenna architectures, and reconfigurable intelligent surface (RIS), etc., they also need to be better evolved and developed.

[0006] In order to meet some of newly added functions of the 6G communication systems, new technologies need to be developed in the terms of network energy saving, air-interface security, and network security, meanwhile the feasibility of fusion technologies such as Integrated Sensing and Communication, needs to be studied.

[0007] Moreover, in order to improve the spectral efficiency and the overall network performances, the following technologies have been developed for 6G communication systems: a full-duplex technology for enabling an uplink transmission and a downlink transmission to simultaneously use the same frequency resource at the same time; a network technology for utilizing satellites, high-altitude platform stations (HAPS), and the like in an integrated manner; an improved network structure for supporting mobile base stations and the like and enabling network operation optimization and automation and the like; a dynamic spectrum sharing technology via collision avoidance based on a prediction of spectrum usage; an use of artificial intelligence (AI) in wireless communication for improvement of overall network operation by utilizing AI from a designing phase for developing 6G and internalizing end-to-end AI support functions; and a next-generation distributed computing technology for overcoming the limit of user equipment (UE) computing ability through reachable super-high-performance communication and computing resources (such as mobile edge computing (MEC), clouds, and the like) over the network. In addition, through designing new protocols to be used in 6G communication systems, developing mechanisms for implementing a hardware-based security environment and safe use of data, and developing technologies for maintaining privacy, attempts to strengthen the connectivity between devices, optimize the network, promote softwarization of network entities, and increase the openness of wireless communications are continuing.

[0008] It is expected that research and development of 6G communication systems in hyper-connectivity, including person to machine (P2M) as well as machine to machine (M2M), will allow the next hyper-connected experience. Particularly, it is expected that services such as truly immersive extended reality (XR), high-fidelity mobile hologram, and digital replica could be provided through 6G communication systems. In addition, services such as remote surgery for security and reliability enhancement, industrial automation, and emergency response will be provided through the 6G communication system such that the technologies could be applied in various fields such as industry, medical care, automobiles, and home appliances.

[0009]

[0010] The objective of embodiments of the present disclosure is to solve the technical problem of how to achieve the best beamforming performance in all wireless channel environments.

[0011]

[0012] In accordance with one aspect of the embodiments of the present disclosure, there is provided a method performed by a base station in a wireless communication system, the method comprising: receiving, from a terminal, a sounding reference signal (SRS) for obtaining a channel matrix; generating first information on each performance loss in a channel for a plurality of candidate beamforming (BF) algorithms, based on a precoding matrix indicator (PMI) and the channel matrix for the SRS; determining a BF algorithm with highest gain among the plurality of the candidate BF algorithms based on the first information, the plurality of the candidate BF algorithms including a PMI based BF algorithm and an SRS-based BF algorithm; determining a BF weight based on the BF algorithm with the highest gain; and transmitting, to a terminal, downlink data based on the BF weight.

[0013] In accordance with another aspect of the embodiments of the present disclosure, there is provided a method performed by a terminal in a wireless communication system, the method comprising: transmitting, to a base station, a sounding reference signal (SRS) used for obtaining a channel matrix; and receiving, from a base station, a downlink data based on a beamforming (BF) weight, wherein the BF weight is determined based on a BF algorithm with a highest gain, and wherein the BF algorithm with the highest gain is determined based on a precoding matrix indicator (PMI) transmitted from the terminal and the channel matrix.

[0014] In accordance with still another aspect of the embodiments of the present disclosure, there is provided a base station in a wireless communication system, the base station comprising: a transceiver; and a processor coupled to the transceiver and configured to: receive, from a terminal, a sounding reference signal (SRS) for obtaining a channel matrix; generate first information on each performance loss in a channel for a plurality of candidate beamforming (BF) algorithms, based on a precoding matrix indicator (PMI) and the channel matrix for the SRS; determine a BF algorithm with highest gain among the plurality of the candidate BF algorithms based on the first information, the plurality of the candidate BF algorithms including a PMI based BF algorithm and an SRS-based BF algorithm; determine a BF weight based on the BF algorithm with the highest gain; and transmit, to a terminal, downlink data based on the BF weight.

[0015] In accordance with yet another aspect of the embodiments of the present disclosure, there is provided a terminal in a wireless communication system, the method comprising: a transceiver; and a processor coupled to the transceiver and configured to: transmit, to a base station, a sounding reference signal (SRS) used for obtaining a channel matrix, and receive, from a base station, a downlink data based on a beamforming (BF) weight, wherein the BF weight is determined based on a BF algorithm with a highest gain, and wherein the BF algorithm with the highest gain is determined based on a precoding matrix indicator (PMI) transmitted from the terminal and the channel matrix.

[0016] In the embodiments of the present disclosure, an adaptation of the BF algorithm based on the first AI model can be realized, and the BF algorithm is dynamically selected based on the first information related to the losses of candidate BF algorithms under the current wireless channel, so that the spectral efficiency can be maximized in any wireless channel environment.

[0017]

[0018] According to an example of the present disclosure, by using an AI-based beamforming method, the base station can improve the beamforming gain for a terminal with the allocated SRS resources.

[0019] Additionally, according to an example of the present disclosure, the base station can dynamically select the beamforming algorithm with the highest gain suitable for the current wireless channel by using an AI model.

[0020] Moreover, the base station can perform optimal per-antenna power constraint (PCPA) in a 5G / 6G massive-multi input multi output (MIMO) system by utilizing an AI model, while maintaining a balance between power loss and orthogonality loss.

[0021]

[0022] To more clearly explain the technical schemes in the embodiments of the present disclosure, the drawings to be used in the description of the embodiments of the present disclosure will be briefly introduced below.

[0023] Figure 1a is a schematic diagram of a wireless network according to an embodiment of the present disclosure.

[0024] Figure 1b is a schematic diagram of a base station according to an embodiment of the present disclosure.

[0025] Figure 1c is a schematic diagram of a user equipment according to an embodiment of the present disclosure.

[0026] Figure 1d is a flowchart of a method performed by a base station in a communication system according to an embodiment of the present disclosure.

[0027] Figure 2 is a schematic diagram of a BF effect according to an embodiment of the present disclosure.

[0028] Figure 3 is a schematic diagram of a proportional scaling per-antenna power constraint (PAPC) according to an embodiment of the present disclosure.

[0029] Figure 4 is a schematic diagram of a power loss according to an embodiment of the present disclosure.

[0030] Figure 5 is a schematic diagram of a full-power PAPC according to an embodiment of the present disclosure.

[0031] Figure 6 is a schematic diagram of adaptive BF algorithm selection based on features on one RB according to an embodiment of the present disclosure.

[0032] Figure 7 is a schematic diagram of adaptive BF algorithm selection based on features on multiple RBs according to an embodiment of the present disclosure.

[0033] Figure 8 is a schematic structure diagram of a second AI model according to an embodiment of the present disclosure.

[0034] Figure 9 is a schematic diagram of an AI-based PAPC process according to an embodiment of the present disclosure.

[0035] Figure 10 is a schematic diagram of a training process of the second AI model according to an embodiment of the present disclosure.

[0036] Figure 11 is a schematic diagram of a data collection process of each episode according to an embodiment of the present disclosure.

[0037] Figure 12 is a schematic diagram of an interaction and data collection process in a step according to an embodiment of the present disclosure.

[0038] Figure 13 is a flowchart of an AI-based adaptive beamforming scheme according to an embodiment of the present disclosure.

[0039] Figure 14 is a flowchart of a method performed by a user equipment in a communication system according to an embodiment of the present disclosure.

[0040] Figure 15 is a schematic structure diagram of an electronic device according to an embodiment of the present disclosure.

[0041]

[0042] The following description with reference to the accompanying drawings is provided to assist in a comprehensive understanding of various embodiments of the present disclosure as defined by the claims and their equivalents. It includes various specific details to assist in that understanding but these are to be regarded as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the various embodiments described herein can be made without departing from the scope and spirit of the present disclosure. In addition, descriptions of well-known functions and constructions may be omitted for clarity and conciseness.

[0043] The terms and words used in the following description and claims are not limited to the bibliographical meanings, but, are merely used by the inventor to enable a clear and consistent understanding of the present disclosure. Accordingly, it should be apparent to those skilled in the art that the following description of various embodiments of the present disclosure is provided for illustration purpose only and not for the purpose of limiting the present disclosure as defined by the appended claims and their equivalents.

[0044] It is to be understood that the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to "a component surface" includes reference to one or more of such surfaces. When a component is said to be "connected" or "coupled" to the other component, the component can be directly connected or coupled to the other component, or it can mean that the component and the other component are connected through an intermediate element. In addition, "connected" or "coupled" as used herein may include wireless connection or wireless coupling.

[0045] The term "include" or "may include" refers to the existence of a corresponding disclosed function, operation or component which can be used in various embodiments of the present disclosure and does not limit one or more additional functions, operations, or components. The terms such as "include" and / or "have" may be construed to denote a certain characteristic, number, step, operation, constituent element, component or a combination thereof, but may not be construed to exclude the existence of or a possibility of addition of one or more other characteristics, numbers, steps, operations, constituent elements, components or combinations thereof.

[0046] The term "or" used in various embodiments of the present disclosure includes any or all of combinations of listed words. For example, the expression "A or B" may include A, may include B, or may include both A and B. When describing multiple (two or more) items, if the relationship between multiple items is not explicitly limited, the multiple items can refer to one, many or all of the multiple items. For example, the description of "parameter A includes A1, A2 and A3" can be realized as parameter A includes A1 or A2 or A3, and it can also be realized as parameter A includes at least two of the three parameters A1, A2 and A3.

[0047] Unless defined differently, all terms used herein, which include technical terminologies or scientific terminologies, have the same meaning as that understood by a person skilled in the art to which the present disclosure belongs. Such terms as those defined in a generally used dictionary are to be interpreted to have the meanings equal to the contextual meanings in the relevant field of art, and are not to be interpreted to have ideal or excessively formal meanings unless clearly defined in the present disclosure.

[0048] The figures included herein, and the various embodiments used to describe the principles of the present disclosure are by way of illustration only and should not be construed in any way to limit the scope of the disclosure. Further, those skilled in the art will understand that the principles of the present disclosure may be implemented in any suitably arranged wireless communication system.

[0049] Figures 1a-1c below describe various embodiments of the present disclosure implemented in wireless communications systems. The descriptions of FIGS. 1a-1c are not meant to imply physical or architectural limitations to the manner in which different embodiments may be implemented. Different embodiments of the present disclosure may be implemented in any suitably-arranged communications system.

[0050] Figure 1a illustrates an example wireless network according to embodiments of the present disclosure.

[0051] The embodiment of the wireless network shown in Figure 1a is for illustration only. Other embodiments of the wireless network 100 could be used without departing from the scope of the present disclosure.

[0052] As shown in Figure 1a, the wireless network includes a base station (next generation nodeB, gNB or gNodeB) 101, a gNB 102, and a gNB 103. The gNB 101 communicates with the gNB 102 and the gNB 103. The gNB 101 also communicates with at least one network 130, such as the Internet, a proprietary Internet Protocol (IP) network, or other data network.

[0053] The gNB 102 provides wireless broadband access to the network 130 for a first plurality of user equipments (UEs) within a coverage area 120 of the gNB 102. The first plurality of UEs includes a UE 111, which may be located in a small business (SB); a UE 112, which may be located in an enterprise (E); a UE 113, which may be located in a WiFi hotspot (HS); a UE 114, which may be located in a first residence (R1); a UE 115, which may be located in a second residence (R2); and a UE 116, which may be a mobile device (M), such as a cell phone, a wireless laptop, a wireless personal digital assistant (PDA), or the like. The gNB 103 provides wireless broadband access to the network 130 for a second plurality of UEs within a coverage area 125 of the gNB 103. The second plurality of UEs includes the UE 115 and the UE 116, as well as subscriber stations (SS, for example, UEs) 117, 118 and 119. In some embodiments, one or more of the gNBs 101-103 may communicate with each other and with the UEs 111-116 using existing wireless communication techniques, and one or more of the UE 111-119 may communicate directly with each other (e.g., UEs 117-119) using other existing or proposed wireless communication techniques.

[0054] Depending on the network type, the term "base station" or "BS" can refer to any component (or collection of components) configured to provide wireless access to a network, such as transmit point (TP), transmit-receive point (TRP), an enhanced (or "evolved") base station (eNodeB or eNB), a 5G base station (gNB), a macrocell, a femtocell, a wireless fidelity (WiFi) access point (AP), or other wirelessly enabled devices. Base stations may provide wireless access in accordance with one or more wireless communication protocols, e.g., 3GPP 5G New Radio (NR), Long Term Evolution (LTE), LTE Advanced (LTE-A), high speed packet access (HSPA), Wi-Fi 802.11a / b / g / n / ac, etc. For the sake of convenience, the various names for a base station-type apparatus and functionality are used interchangeably in this patent document to refer to network infrastructure components that provide wireless access to remote terminals. Also, depending on the network type, the term "user equipment" (UE) can refer to any component such as a mobile station (MS), subscriber station (SS), remote terminal, wireless terminal, receive point, or user device. For the sake of convenience, the various names for a user equipment-type device and functionality are used interchangeably in this patent document to refer to remote wireless equipment that wirelessly accesses a BS, whether the UE is a mobile device (such as a mobile telephone or smartphone) or is normally considered a stationary device (such as a desktop computer or vending machine).

[0055] Dotted lines show the approximate extents of the coverage areas 120 and 125, which are shown as approximately circular for the purposes of illustration and explanation only. It should be clearly understood that the coverage areas associated with gNBs, such as the coverage areas 120 and 125, may have other shapes, including irregular shapes, depending upon the configuration of the gNBs and variations in the radio environment associated with natural and man-made obstructions.

[0056] As described in more detail below, one or more of the UEs 111-119 include circuitry, programing, or a combination thereof. In certain embodiments, and one or more of the gNBs 101-103 includes circuitry, programing, or a combination thereof.

[0057] Although Figure 1a illustrates one example of a wireless network, various changes may be made to Figure 1a. For example, the wireless network could include any number of gNBs and any number of UEs in any suitable arrangement. Also, the gNB 101 could communicate directly with any number of UEs and provide those UEs with wireless broadband access to the network 130. Similarly, each gNB 102-103 could communicate directly with the network 130 and provide UEs with direct wireless broadband access to the network 130. Further, the gNBs 101, 102, and / or 103 could provide access to other or additional external networks, such as external telephone networks or other types of data networks.

[0058] Figure 1b illustrates an example base station according to embodiments of the present disclosure.

[0059] The embodiment of the gNB 102 illustrated in Figure 1b is for illustration only, and the gNBs 101 and 103 of Figure 1a could have the same or similar configuration. However, gNBs come in a wide variety of configurations, and Figure 1b does not limit the scope of the present disclosure to any particular implementation of a gNB.

[0060] As shown in figure 1b, the gNB 102 includes multiple antennas 200a-200n, multiple radio frequency (RF) transceivers 201a-201n, transmit (TX) processing circuitry 203, and receive (RX) processing circuitry 204. The gNB 102 also includes a controller / processor 205, a memory 206, and a backhaul or network interface 207.

[0061] The RF transceivers 201a-201n receive, from the antennas 200a-200n, incoming RF signals, such as signals transmitted by UEs in the network 100. The RF transceivers 201a-201n down-convert the incoming RF signals to generate intermediate frequency (IF) or baseband signals. The IF or baseband signals are sent to the RX processing circuitry 204, which generates processed baseband signals by filtering, decoding, and / or digitizing the baseband or IF signals. The RX processing circuitry 204 transmits the processed baseband signals to the controller / processor 205 for further processing.

[0062] The TX processing circuitry 203 receives analog or digital data (such as voice data, web data, electronic mail, or interactive video game data) from the controller / processor 205. The TX processing circuitry 203 encodes, multiplexes, and / or digitizes the outgoing baseband data to generate processed baseband or IF signals. The RF transceivers 201a-201n receive the outgoing processed baseband or IF signals from the TX processing circuitry 203 and up-converts the baseband or IF signals to RF signals that are transmitted via the antennas 201a-201n.

[0063] The controller / processor 205 can include one or more processors or other processing devices that control the overall operation of the gNB 102. For example, the controller / processor 205 could control the reception of forward channel signals and the transmission of reverse channel signals by the RF transceivers 201a-201n, the RX processing circuitry 204, and the TX processing circuitry 203 in accordance with well-known principles. The controller / processor 205 could support additional functions as well, such as more advanced wireless communication functions.

[0064] For instance, the controller / processor 205 could support beam forming or directional routing operations in which outgoing signals from multiple antennas 200a-200n are weighted differently to effectively steer the outgoing signals in a desired direction. Any of a wide variety of other functions could be supported in the gNB 102 by the controller / processor 205.

[0065] The controller / processor 205 is also capable of executing programs and other processes resident in the memory 206, such as an operating system (OS). The controller / processor 205 can move data into or out of the memory 206 as required by an executing process.

[0066] The controller / processor 205 is also coupled to the backhaul or network interface 207. The backhaul or network interface 207 allows the gNB 102 to communicate with other devices or systems over a backhaul connection or over a network. The interface 207 could support communications over any suitable wired or wireless connection(s). For example, when the gNB 102 is implemented as part of a cellular communication system (such as one supporting 5G, LTE, or LTE-A), the interface 207 could allow the gNB 102 to communicate with other gNBs over a wired or wireless backhaul connection. When the gNB 102 is implemented as an access point, the interface 207 could allow the gNB 102 to communicate over a wired or wireless local area network or over a wired or wireless connection to a larger network (such as the Internet). The interface 207 includes any suitable structure supporting communications over a wired or wireless connection, such as an Ethernet or RF transceiver.

[0067] The memory 206 is coupled to the controller / processor 205. Part of the memory 206 could include a random access memory (RAM), and another part of the memory 206 could include a Flash memory or other read only memory (ROM).

[0068] Although Figure 1b illustrates one example of gNB 102, various changes may be made to Figure 1b. For example, the gNB 102 could include any number of each component shown in Figure 1b. As a particular example, an access point could include a number of interfaces 207, and the controller / processor 205 could support routing functions to route data between different network addresses. As another particular example, while shown as including a single instance of TX processing circuitry 203 and a single instance of RX processing circuitry 204, the gNB 102 could include multiple instances of each (such as one per RF transceiver). Also, various components in Figure 1b could be combined, further subdivided, or omitted and additional components could be added according to particular needs.

[0069]

[0070] Figure 1c illustrates an example user equipment according to embodiments of the present disclosure.

[0071] The embodiment of the UE 116 illustrated in Figure 1c is for illustration only, and the UEs 111-115 and 117-119 of Figure 1a could have the same or similar configuration. However, UEs come in a wide variety of configurations, and Figure 1c does not limit the scope of the present disclosure to any particular implementation of a UE.

[0072] As shown in Figure 1c, the UE 116 includes an antenna 301, a radio frequency (RF) transceiver 302, TX processing circuitry 303, a microphone 304, and receive (RX) processing circuitry 305. The UE 116 also includes a speaker 306, a controller or processor 307, an input / output (I / O) interface (IF) 308, an input device 309, a touchscreen display 310, and a memory 311. The memory 311 includes an OS 312 and one or more applications 313.

[0073] The RF transceiver 302 receives, from the antenna 301, an incoming RF signal transmitted by a gNB of the network 100. The RF transceiver 302 down-converts the incoming RF signal to generate an IF or baseband signal. The IF or baseband signal is sent to the RX processing circuitry 305, which generates a processed baseband signal by filtering, decoding, and / or digitizing the baseband or IF signal. The RX processing circuitry 305 transmits the processed baseband signal to the speaker 306 (such as for voice data) or to the processor 307 for further processing (such as for web browsing data).

[0074] The TX processing circuitry 303 receives analog or digital voice data from the microphone 304 or other outgoing baseband data (such as web data, e-mail, or interactive video game data) from the processor 307. The TX processing circuitry 303 encodes, multiplexes, and / or digitizes the outgoing baseband data to generate a processed baseband or IF signal. The RF transceiver 302 receives the outgoing processed baseband or IF signal from the TX processing circuitry 303 and up-converts the baseband or IF signal to an RF signal that is transmitted via the antenna 301.

[0075] The processor 307 can include one or more processors or other processing devices and execute the OS 312 stored in the memory 311 in order to control the overall operation of the UE 116. For example, the processor 307 could control the reception of forward channel signals and the transmission of reverse channel signals by the RF transceiver 302, the RX processing circuitry 305, and the TX processing circuitry 303 in accordance with well-known principles. In some embodiments, the processor 307 includes at least one microprocessor or microcontroller.

[0076] The processor 307 is also capable of executing other processes and programs resident in the memory 311, such as processes for CSI reporting on uplink channel. The processor 307 can move data into or out of the memory 311 as required by an executing process. In some embodiments, the processor 307 is configured to execute the applications 313 based on the OS 312 or in response to signals received from gNBs or an operator. The processor 307 is also coupled to the I / O interface 308, which provides the UE 116 with the ability to connect to other devices, such as laptop computers and handheld computers. The I / O interface 308 is the communication path between these accessories and the processor 307.

[0077] The processor 307 is also coupled to the touchscreen display 310. The user of the UE 116 can use the touchscreen display 310 to enter data into the UE 116. The touchscreen display 310 may be a liquid crystal display, light emitting diode display, or other display capable of rendering text and / or at least limited graphics, such as from web sites.

[0078] The memory 311 is coupled to the processor 307. Part of the memory 311 could include RAM, and another part of the memory 311 could include a Flash memory or other ROM.

[0079] Although Figure 1c illustrates one example of UE 116, various changes may be made to Figure 1c. For example, various components in Figure 1c could be combined, further subdivided, or omitted and additional components could be added according to particular needs. As a particular example, the processor 307 could be divided into multiple processors, such as one or more central processing units (CPUs) and one or more graphics processing units (GPUs). Also, while Figure 1c illustrates the UE 116 configured as a mobile telephone or smartphone, UEs could be configured to operate as other types of mobile or stationary devices.

[0080] At least some of the functions in the apparatus or electronic device provided in the embodiments of the present disclosure may be implemented by an AI model. For example, at least one of a plurality of modules of the apparatus or electronic device may be implemented through the AI model. The functions associated with the AI can be performed through a non-volatile memory, a volatile memory, and a processor.

[0081] The processor may include one or more processors. At this time, the one or more processors may be general-purpose processors such as a central processing unit (CPU), an application processor (AP), etc., or a pure graphics processing unit, such as a graphics processing unit (GPU), a visual processing unit (VPU), and / or an AI specialized processor, such as a neural processing unit (NPU).

[0082] The one or more processors control the processing of input data according to predefined operating rules or artificial intelligence (AI) models stored in the non-volatile memory and the volatile memory. The predefined operating rules or AI models are provided by training or learning.

[0083] Here, providing, by learning, refers to obtaining the predefined operating rules or AI models having a desired characteristic by applying a learning algorithm to a plurality of learning data. The learning may be performed in the apparatus or electronic device itself in which the AI according to the embodiments is performed, and / or may be implemented by a separate server / system.

[0084] The AI models may include a plurality of neural network layers. Each layer has a plurality of weight values. Each layer performs the neural network computation by computation between the input data of that layer (e.g., the computation results of the previous layer and / or the input data of the AI models) and the plurality of weight values of the current layer. Examples of neural networks include, but are not limited to, a convolutional neural network (CNN), a deep neural network (DNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bi-directional recurrent deep neural network (BRDNN), generative adversarial networks (GANs), and deep Q-networks.

[0085] The learning algorithm is a method of training a predetermined target apparatus (e. g., a robot) by using a plurality of learning data to enable, allow, or control the target apparatus to make a determination or prediction. Examples of the learning algorithm include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.

[0086] In accordance with the present disclosure, at least one step (e.g., determining a first BF algorithm, performing a per-antenna power constraint (PAPC), etc.) in the method performed by the base station in the communication system may be implemented by using an artificial intelligence model. The processor of the electronic device may preprocess data to convert the data into a form suitable for use as an input to the artificial intelligence model. The artificial intelligence model may be obtained by training. Here, "obtained by training" means that predefined operating rules or artificial intelligence models configured to perform desired features (or purposes) are obtained by training a basic artificial intelligence model with multiple pieces of training data by a training algorithm.

[0087] A beamforming (BF) technology is one of key technologies in the 5G / 6G wireless communication system. It is a signal preprocessing technology based on an antenna array, which generates directional beams by adjusting a weight of each array element in the antenna array, so that an obvious array gain can be obtained.

[0088] Therefore, the beamforming provides many significant advantages for a 5G / 6G network, such as increasing the signal strength and coverage, improving a signal-to-noise ratio (SNR), increasing a bandwidth and data rate, improving energy efficiency, and improving network capacity, and so on.

[0089] How to achieve the best beamforming performance in wireless channels in different channel environments is a technical problem to be enhanced.

[0090] The embodiments of the present disclosure provide a communication method, a base station, a user equipment, a storage medium, and a program product, and can also be understood as an AI-based intelligent beamforming scheme, which realizes an adaptation of the BF algorithm in different wireless channel environments, and determines the best BF algorithm in each wireless channel environment accurately and quickly by using an AI network, so as to improve the spectral efficiency.

[0091] In order to make the objectives, technical schemes and advantages of the present disclosure clearer, the technical schemes in the embodiments of the present disclosure and the technical effects achieved by the technical schemes of the present disclosure will be described below by several exemplary implementations with reference to the drawings. It should be pointed out that the following implementations can be referred to, learned from or combined with each other, and the same terms, similar features and similar implementation steps in different implementations will not be repeated again.

[0092] Figure 1d is a flowchart of a method performed by a base station in a communication system according to an embodiment of the present disclosure.

[0093] An embodiment of the present disclosure provides a method performed by a base station (BS) in a communication system. As shown in Figure 1d, the method comprises the following steps.

[0094] In step S101, wireless channel information is acquired, the wireless channel information comprising first information related to losses of candidate BF algorithms, the first information being determined based on a precoding matrix indicator (PMI), singular values and singular vectors, the singular values and the singular vectors being determined based on a sounding reference signal (SRS) measurement related channel matrix.

[0095] In the embodiment of the present disclosure, the wireless channel information refers to related information obtained by counting, calculating (or measuring) or processing the signals transmitted on the wireless channels for at least one SRS period, and may further include wireless channel parameters. Optionally, the wireless channel information may also be referred to as wireless environment information or wireless environment parameters, or wireless parameters for short.

[0096] As an example, the embodiment of the present disclosure can be applied in a time-division duplex (TDD) downlink single-user multiple-input multiple-output (MIMO) communication scenario, so the wireless channel information may refer to channel matrix related information of a MIMO system.

[0097] In the embodiment of the present disclosure, the wireless channel information comprises first information related to losses of candidate BF algorithms, and the first information related to the losses of candidate BF algorithms can measure the performance loss of each BF algorithm in the current wireless channel and be used for determining a first BF algorithm with highest spectral efficiency in step S102.

[0098] In the embodiment of the present disclosure, the acquiring the first information related to the losses of candidate BF algorithms may specifically comprise: determining the first information related to the losses of candidate BF algorithms based on a precoding matrix indicator (PMI) and a sounding reference signal (SRS) measurement related channel matrix (SRS H).

[0099] The process can also be understood as a data preprocessing process or a feature extraction process. Secondary features for measuring the performance loss of each BF algorithm are extracted by a signal processing technology, thus improving the algorithm classification performance (e.g., accuracy) and reducing the complexity of the AI model.

[0100] The PMI may be measured from a downlink channel by the UE based on a channel state information-reference signal (CSI-RS) and reported to the BS through CSI.

[0101] The SRS H is a result of channel estimation by the BS using an uplink (UL) SRS, , where is the number of transmit antennas of the base station, and is the number of receive antennas of the terminal. The specific numerical values of and may be subject to the actual implementation. For example, it is possible that , but it is not limited thereto. Based on the TDD channel reciprocity, the SRS H can also be understood as that the results of estimation of downlink channels are the same.

[0102] The PMI and the SRS H can also be understood as downlink channel information or downlink channel feature.

[0103] Optionally, singular value decomposition (SVD) may be performed on the channel matrix (SRS H) to obtain singular values and singular vectors. The first information related to the losses of candidate BF algorithms is determined based on the PMI, the singular values and the singular vectors.

[0104] The SVD of the SRS H may be as shown in the following math figure 1:

[0105]

[0106] where is a transpose matrix ofV, from which a diagonal matrix and orthogonal matricesUandVcan be solved. The diagonal matrix includes at least one singular value , and each of the orthogonal matrices and includes at least one singular vector.

[0107] In step S102, the first BF algorithm is determined from at least two candidate BF algorithms based on the wireless channel information by using a first artificial intelligence (AI) model.

[0108] The candidate BF algorithms refer to any algorithm that can be used for MIMO beamforming according to an architecture of acquiring the wireless channel information by the BS. Optionally, the candidate BF algorithms provided in the embodiment of the present disclosure include, but not limited to, at least one of:

[0109] (1) PMI-based BF algorithm

[0110] For example, the BS transmits a downlink channel state information-reference signal (CSI-RS) to the UE, the UE estimates downlink channel state information (CSI), and the UE reports the CSI to the BS. The PMI-based BF algorithm is to select, by the BS, a precoding matrix from a predetermined codebook by using the CSI reported by the UE, so that a BF weight can be obtained.

[0111] (2) SRS-based BF algorithm

[0112] For example, the BS allocates an uplink (UL) SRS resource to the UE, the UE transmits an UL SRS to the BS based on the UL SRS resource, and the BS estimates the downlink channels by using the UL SRS (based on the TDD channel reciprocity). The SRS-based BF algorithm is to calculate, by the BS, the BF weight by using the result of channel estimation of the SRS.

[0113] The SRS-based BF algorithm includes at least one of:

[0114] (2-1) an SRS-based zero forcing (ZF) algorithm; and

[0115] (2-2) an SRS-based singular value decomposition (SVD) algorithm.

[0116] Optionally, the SRS-based SVD algorithm may adopt a low-complexity SVD algorithm, for example, but not limited to, a Lanczos Bipartite Jacobian (LBJ)-SVD algorithm or the like.

[0117] In the embodiment of the present disclosure, the first BF algorithm refers to the optimal BF algorithm for current SRS period dynamically selected from the at least two candidate BF algorithms, and can achieve the highest spectral efficiency.

[0118] Optionally, the at least two candidate BF algorithms may refer to all candidate BF algorithms, for example, the PMI-based BF algorithm, the SRS-based ZF algorithm, the SRS-based SVD algorithm and the like. Alternatively, the at least two candidate BF algorithms may refer to at least two candidate BF algorithms obtained after preliminarily screening all candidate BF algorithms. Those skilled in the art can make extensions according to actual conditions, and the number of candidate BF algorithms and the specific calculation method will not be limited in the embodiments of the present disclosure.

[0119] Considering that the performance of different BF algorithms depends on the channel and environment and the wireless channels are always changing rapidly in a real environment due to the mobility of user equipments and the mobility of electromagnetic scatterers. For example, line of sight (LOS) and non-line of sight (NLOS) wireless transmissions occur alternately, and each candidate BF algorithm has its own advantages in different situations. For example, the SRS-based SVD algorithm has better performance. For another example, the SRS-based ZF algorithm has low complexity and low hardware implementation difficulty. For example, the SRS-based BF algorithm has no quantization error, and has better performance than the PMI-based BF algorithm in most cases. For another example, in some cases, the PMI-based BF algorithm has better performance than the SRS-based BF algorithm, such as for a user equipment at a cell edge (because the uplink power of the user equipment is limited, the coverage of the uplink SRS is smaller than that of the downlink CSI-RS, and the error of SRS channel estimation is high, resulting in poor quality of SRS channel estimation), and high spatial correlation channels (such as a LOS scenario) and other situations.

[0120] In the embodiment of the present disclosure, in order to achieve the best beamforming performance in all wireless channels, the best first BF algorithm in each wireless channel situation is accurately and rapidly determined by using the first AI model.

[0121] In the embodiment of the present disclosure, the first AI model can also be understood as a BF algorithm classification model and can adopt a lightweight AI classifier. In practical applications, those skilled in the art can set the specific AI model adopted in this step according to the actual situation. For example, it may mainly include a simple multilayer perceptron (MLP) classifier or the like. It will not be limited in the embodiment of the present disclosure.

[0122] In step S103, a first BF weight is determined based on the first BF algorithm.

[0123] In the embodiment of the present disclosure, after the first BF algorithm is determined, the BS may determine the first BF weight by a corresponding calculation method of the BF algorithm. The specific calculation method has no influence on the implementation of this scheme and will not be repeated here.

[0124] In step S104, downlink data is transmitted based on the first BF weight.

[0125] In the embodiment of the present disclosure, after the first BF weight is determined, the BS may apply the first BF weight to the downlink data to change the beam shape and direction so as to achieve the BF effect, as shown in Figure 2.

[0126]

[0127] Figure 2 is a schematic diagram of a BF effect according to an embodiment of the present disclosure.

[0128] Optionally, the embodiment of the present disclosure can be applied to a user equipment allocated with SRS resources. For example, the step S102 may specifically comprise: in a case where the user equipment is allocated with resources for SRS, determining the first BF algorithm from the at least two candidate BF algorithms based on the wireless channel information by using the first AI model.

[0129] For example, in one example, the BS may determine whether to allocate resources for SRS to the UE. If it is determined to allocate the resources for SRS to the UE, the resources for SRS are allocated to the UE. In this case, the SRS-based BF algorithm is generally used, but the SRS-based BF algorithm does not necessarily achieve the best beamforming performance. For example, in a LOS scenario, the SRS-based ZF algorithm has a higher nulling loss, and its performance is inferior to the PMI-based algorithm. Therefore, the AI-based BF scheme provided in the embodiment of the present application can be adopted to achieve the best beamforming performance. If it is determined to not allocate the resources for SRS to the UE, the resources for SRS are not allocated to the UE. In this case, the PMI-based BF algorithm can be used. Further, after the resources for SRS are allocated to the UE, the BS will determine whether to release the resources for SRS. After the UE's resources for SRS are released, the PMI-based BF algorithm will be used.

[0130] In the embodiment of the present disclosure, the performance granularity of the step S102 may be for each SRS period (e.g., SRS reporting period), for example, each time unit. The time unit may be specifically a slot, but it is not limited thereto. Those skilled in the art can set a time length of the SRS period according to the actual situation, and it will not be limited in the embodiment of the present disclosure. As an example, the step S102 may specifically comprise: determining the first BF algorithm corresponding to the current SRS period from the at least two candidate BF algorithms based on the wireless channel information by using the first AI model.

[0131] In the embodiment of the present disclosure, the performance granularity of the steps S103 and S104 may be for each precoding resource block group (PRG). As an example, the step S103 may specifically comprise: determining a first BF weight corresponding to each PRG, and the step S104 may specifically comprise: transmitting downlink data based on the first BF weight corresponding to each PRG.

[0132] It should be understood by those skilled in the art that the above several performance granularities are only schematic descriptions and do not constitute limitations on the embodiments of the present disclosure, and appropriate variations made based on these examples can also be applied to the present disclosure and shall fall into the scope of protection of the present disclosure.

[0133] In the embodiment of the present disclosure, the first information related to the losses of candidate BF algorithms includes:

[0134] (1) Information associated with the result of comparison of the maximum singular value and the minimum singular value in the singular values

[0135] In the embodiment of the present disclosure, the singular value after the channel SVD represents the energy intensity of each decomposed orthogonal layer. For example, the result of comparison of the maximum singular value and the maximum singular value is used to represent the channel energy concentration.

[0136] Optionally, the result of comparison of the maximum singular value and the minimum singular value includes a quotient (ratio) of the maximum singular value to the minimum singular value, for example, a value of the maximum singular value / minimum singular value. The channel energy concentration may be represented by using the value of the maximum singular value / minimum singular value. If the value of the maximum singular value / minimum singular value is larger, meaning that the energy is more concentrated on the main path, a channel correlation between different transport layers is higher (e.g., in the LOS scenario), and the algorithm performance is worse.

[0137] In the embodiment of the present disclosure, an energy concentration index may also be used to measure the nulling loss of the SRS-based ZF algorithm. The ZF algorithm can achieve an orthogonal effect of each layer and eliminate the interference between data streams in the UE. However, its disadvantage is that a false noise will be amplified by the ZF weight. This phenomenon is called nulling loss. The performance of the ZF algorithm is limited by the nulling loss. A larger nulling loss indicates higher spatial correlation of the channel and worse performance of the ZF algorithm. For example, in the LOS scenario, the correlation of each layer of the channel is higher, and the ZF algorithm will inevitably bring about a higher nulling loss in order to achieve the orthogonal effect of each layer. In other words, the nulling loss is related to the result of comparison of the maximum singular value and the minimum singular value and can be measured by using the result of comparison of the maximum singular value and the minimum singular value.

[0138] (2) Similarity between the singular vector and the PMI

[0139] Considering that the performance of the PMI-based BF algorithm is limited by the quantization error of the codebook, in the embodiment of the present disclosure, the quantization error of the PMI is measured by the similarity between the singular vector of channel decomposition and the PMI.

[0140] When the similarity between the singular vector and the PMI is higher, the position of the user equipment is closer to the pointing direction of the center of the quantized PMI beam. At this time, the quantization error of the PMI is smaller. Since the BF weight obtained by the SRS-based BF algorithm (e.g., the ZF algorithm or SVD algorithm) is a non-constant modulus matrix and needs to be subjected to PAPC, the power loss / orthogonal loss may be increased. However, since the PMI codebook is a constant modulus matrix and does not need to be subjected to PAPC, when the quantization error of the PMI is smaller, the PMI-based BF algorithm has better performance.

[0141] Optionally, the similarity between the singular vector and the PMI includes a cosine similarity between the singular vector and the PMI. However, it is not limited thereto, and other similarity calculation methods are also possible.

[0142] In the embodiment of the present disclosure, the first information may further include at least one of:

[0143] (3) First antenna power variance, which is associated with the BF weight determined by the SRS-based BF algorithm.

[0144] The power of the BF weight obtained by the SRS-based SVD algorithm on each antenna and the power of the BF weight obtained by the ZF algorithm on each antenna are not consistent. In order to satisfy the transmit power condition of antennas on the hardware, it is necessary to perform PAPC on the BF weight for power scaling to constrain the transmit power of each antenna.

[0145] Figure 3 is a schematic diagram of a proportional scaling per-antenna power constraint (PAPC) according to an embodiment of the present disclosure, Figure 4 is a schematic diagram of a power loss according to an embodiment of the present disclosure, and Figure 5 is a schematic diagram of a full-power PAPC according to an embodiment of the present disclosure.

[0146] The PAPC method comprises two modes, i.e., proportional power scaling and full-power.

[0147] 1. Proportional scaling PAPC: as shown in Figure 3, the antenna power with the highest power of the BF weight is scaled to 1, while the power of other antennas is scale in equal proportion, so that some antennas will transmit at a power lower than the maximum power. The proportional scaling PAPC maintains the orthogonality between the layers of the BF weight, but may bring about the power loss, as shown in Figure 4, so that the performance of the SVD algorithm or the ZF algorithm will be reduced.

[0148] 2. Full-power PAPC: as shown in Figure 5, the power of each antenna of the BF weight is scaled to 1, so that each antenna will transmit at the maximum power. The full-power PAPC fully utilizes the transmit power on each antenna, but may bring about the orthogonal loss (introducing an interlayer interference) and will reduce the performance of the SVD algorithm or the ZF algorithm.

[0149] In the embodiment of the present disclosure, the power / orthogonal loss caused by PAPC may be represented by using the antenna power variance of the BF weight.

[0150] Optionally, the BF weight is determined by the SRS-based BF algorithm. For example, for the SRS-based SVD algorithm, the BF weight is first columns of the decomposed orthogonal matrixV(the dimension is ), so that . For the SRS-based ZF algorithm, .

[0151] Further, the variance of the power on each antenna dimension of the BF weight is calculated to obtain the first antenna power variance. The smaller the variance of the first antenna power, the closer the antenna power is, the less the power / orthogonal power loss of the PAPC, and the better the performance of the algorithm.

[0152] (4) Second antenna power variance, which is associated with the BF weight after performing per-antenna power constraint (PAPC)

[0153] Optionally, this step may also calculate the second antenna power variance of the BF weight after performing the PAPC, for example, the antenna power variance of the BF weight after performing the proportional scaling PAPC. The principle is similar and will not be repeated.

[0154] In the embodiment of the present disclosure, the wireless channel information further includes at least one of: PMI, channel matrix (SRS H), rank indicator (RI), channel quality indicator (CQI), receive power for SRS, SRS measurement related signal to interference plus noise ratio (SINR).

[0155] The PMI, the CQI and the RI may be measured from a downlink channel by the UE based on a CSI-RS and reported to the BS through the CSI. The SRS H is the result of channel estimation by the BS using an UL SRS. The SINR may reflect the accuracy of SRS channel estimation.

[0156]

[0157] Figure 6 is a schematic diagram of adaptive BF algorithm selection based on features on one RB according to an embodiment of the present disclosure, Figure 7 is a schematic diagram of adaptive BF algorithm selection based on features on multiple RBs according to an embodiment of the present disclosure, and Figure 8 is a schematic structure diagram of a second AI model according to an embodiment of the present disclosure.

[0158] In the embodiment of the present disclosure, the first information related to the losses of candidate BF algorithms is the first information corresponding to at least one resource block (RB). Optionally, the at least one RB is an RB corresponding to the central frequency of the broadband.

[0159] As an example, when the applied device requires an extremely low complexity scheme, a data preprocessing module may only extract features on one RB of the central frequency of the broadband. As shown in Figure 6, the broadband information (e.g., SRS H, RI, PMI, CQI, etc.) and the features (e.g., SRS power, SINR, channel energy concentration (e.g., the ratio of the maximum singular value to the minimum singular value, etc.), PMI quantization error (e.g., the similarity between the singular vector and the PMI), PAPC loss (e.g., the antenna power variance of the BF weight, etc.) on one RB of the central frequency extracted by the data preprocessing module are input into the first AI model (e.g., a MLP-based BF algorithm classifier), and an identity (ID) of the best BF algorithm among the candidate BF algorithms is output.

[0160] As another example, when the applied device requires a high-accuracy scheme, the data preprocessing module may extract features on multiple RBs of the central frequency of the broadband, for example, features on multiple RBs near the central frequency or features of multiple RBs distributed at equal intervals (e.g., 25%, 50%, 75%, etc.) using the central frequency as a baseline. Those skilled in the art can set the selection modes of multiple RBs according to the actual situation, and it will not be limited in the embodiment of the present disclosure. Specifically, if the wireless channel information at least includes the features on at least two RBs, the step S102 may specifically comprise: fusing the first information corresponding to the at least two RBs to obtain second information; and determining the first BF algorithm from at least two candidate BF algorithms based on at least the second information by using the first AI model.

[0161] For example, as shown in Figure 7, the first AI model may further include a 1D-CNN (convolutional neural networks) encoder corresponding to the RB dimension and a global pooling layer, which is used to fuse the RB-level input (the features on multiple RBs of the central frequency, e.g., SRS power, SINR, channel energy concentration (e.g., the ratio of the maximum singular value to the minimum singular value, etc.), PMI quantization error (e.g., the similarity between the singular vector and the PMI), PAPC loss (e.g., the antenna power variance of the BF weight) into a broadband-level feature (e.g., the second information). The broadband-level feature and the broadband-level input (e.g., SRS H, RI, PMI, CQI, etc.) are input into the MLP classifier to estimate the ID of the best BF algorithm among the candidate BF algorithms.

[0162] In the scheme of dynamically selecting the BF algorithm based on current wireless channel features provided in the embodiment of the present disclosure, by performing data preprocessing on the original channel information (PMI, SRS H, etc.), the feature of the performance loss of each BF algorithm is extracted, and the best BF algorithm for each SRS period is accurately and rapidly determined by using the first AI model, thereby maximizing the spectral efficiency.

[0163] In addition, it is also found by the inventor(s) of the present disclosure that the existing PAPC methods only depend on the input BF weight without considering the wireless channels and thus cannot achieve the best performance. For example, the proportional scaling PAPC cannot fully utilize the power. If the power difference between the antennas of the BF weight is larger, the power loss caused by the proportional scaling PAPC is higher, as shown in Figure 4. For another example, the full-power PAPC will destroy the interlayer orthogonality of the BF weight, resulting in a loss of orthogonality, so that the interlayer interference cannot be effectively suppressed. Moreover, in different wireless environments, the power loss and the orthogonal loss have different influences on the throughput performance.

[0164] Based on this, an embodiment of the present disclosure further provides an AI-based PAPC method. Specifically, if the first BF algorithm is an SRS-based BF algorithm, the step S104 may specifically comprise the following steps.

[0165] In step S201, a second BF weight after performing PAPC on the first BF weight is determined based on the first BF weight and the channel matrix (SRS H) by using a second AI model.

[0166] In step S202, downlink data is transmitted based on the second BF weight.

[0167] In the embodiment of the present disclosure, according to the wireless channels, the power of each antenna is scaled by using the second AI model to realize PAPC, thereby effectively balancing the power loss and the orthogonal loss and achieving the higher spectral efficiency in comparison to the existing PAPC methods.

[0168] In the embodiment of the present disclosure, an optional implementation is provided for the step S201. Specifically, the optional implementation may comprise the following steps.

[0169] In step S2011, PAPC is performed on the first BF weight to obtain a third BF weight.

[0170] For example, the proportional scaling PAPC is performed on the first BF weight calculated by the SRS-based BF algorithm. The antenna with the highest power of the first BF weight is scaled to 1, while the power of other antennas is scaled in equal proportion, so that the scaled third BF weight is output.

[0171] In step S2012, a power compensation coefficient is determined based on the third BF weight and the channel matrix (SRS H) by using the second AI model.

[0172] In the embodiment of the present disclosure, the power compensation coefficient is estimated based on the second AI model. The inputs of the second AI model are the SRS H and the third BF weight scaled in the step S2011. The output power compensation coefficient represents a compensation proportion for the power loss and has a value between 0 and 1 to indicate a balance between the power loss and the orthogonal loss, so it may also be referred to as a power scaling factor.

[0173] In the embodiment of the present disclosure, the second AI model may also be referred to as an AI PAPC model. In practical applications, those skilled in the art can set the specific AI model adopted in this step according to the actual situation, and it will not be limited in the embodiment of the present disclosure. In one example, as shown in Figure 8, the structure of the second AI model may adopt an MLP network to extract spatial features (the inputs are the SRS H and the third BF weight WPAPCscaled in the step S2011), and the power compensation coefficient alpha (all antennas) between 0 and 1 is estimated through an activation function (e.g., sigmoid).

[0174] In step S2013, the third BF weight is adjusted based on the power compensation coefficient to obtain the second BF weight.

[0175]

[0176] Figure 9 is a schematic diagram of an AI-based PAPC process according to an embodiment of the present disclosure, and Figure 10 is a schematic diagram of a training process of the second AI model according to an embodiment of the present disclosure.

[0177] In the embodiment of the present disclosure, the power compensation coefficient is used for the power compensation of the third BF weight scaled in the step S2011. Optionally, the third BF weight is adjusted based on the power compensation coefficient and an antenna power upper limit.

[0178] For example, as shown in Figure 9, an AI-based PAPC process may include: after the first BF weight WBFis calculated by the SRS-based BF algorithm, performing the proportional scaling PAPC to obtain the third BF weight WPAPCand the power loss on each antenna. AI factor estimation is performed in combination with the third BF weight WPAPCand the SRS H by using the second AI model to obtain the power compensation coefficient alpha. Power compensation is performed on the first BF weight WBFbased on the power loss on each antenna and the power compensation coefficient alpha, and the second BF weight WAI-PAPCfinally used for BF processing is output.

[0179] In the embodiment of the present disclosure, an optional implementation is provided for the power compensation. Specifically, a first proportion related to the antenna power associated with the third BF weight and a second proportion related to the antenna power upper limit may be obtained based on the power compensation coefficient, the antenna power corresponding to the third BF weight and the antenna power upper limit are fused, and the second BF weight is obtained based on the result of fusion.

[0180] As an example, the calculation of the result of fusion may be as shown in math figure 2:

[0181]

[0182] where i represents an antenna index, represents the power compensation coefficient alpha, represents the antenna power corresponding to the third BF weight scaled in the step S2011, and represents the antenna power upper limit.

[0183] When is 0, the first proportion is 1, and the antenna i does not need to be compensated, and when is 1, the second proportion is 1, the power loss of the antenna i is completely compensated, and the power of the antenna i reaches the upper limit 1.

[0184] It should be understood that, when the alpha on all antennas is all 0, the algorithm is equivalent to the proportional scaling PAPC, and when the alpha is all 1, the algorithm is equivalent to the full-power PAPC.

[0185] Further, the way of obtaining the second BF weight based on the result of fusion may be as shown in math figure 3:

[0186]

[0187] where i represents an antenna index, represents the antenna power corresponding to the third BF weight scaled in the step S2011, and represents the third BF weight scaled in the step S2011.

[0188] In the embodiment of the present disclosure, the trained second AI model can estimate a better power compensation coefficient than the existing PAPC, thereby achieving higher spectral efficiency.

[0189] In the embodiment of the present disclosure, an optional implementation is provided for the training of the second AI model. Specifically, the second AI model is learnt online by reinforcement learning. Reinforcement learning has been successfully applied to tasks in many fields. Each episode of the task to which the reinforcement learning is applied generally has a same starting point and similar state space. However, in this task, traffics and channels of the user equipments are always changing, so the same starting point and similar state space cannot be satisfied. Therefore, an embodiment of the present disclosures provides a data collection process for the episode and a mechanism of collecting the baseline performance of the current channel and applying the performance after AI-PAPC in a step, thereby reducing the influence on the reinforcement learning process from change in traffics and channels, and ensuring that the trained AI PAPC is better than the existing PAPC algorithms. Specifically, the second AI model may be trained in the following way:

[0190] performing at least one episode of training steps on the third AI model, wherein the training samples of each episode include a first BF weight sample, an SRS channel matrix sample and a downlink data sample, and the training steps of each episode comprise:

[0191] determining a first power compensation coefficient sample based on the first BF weight sample and the SRS channel matrix sample of the episode by using the third AI model, obtaining a Gaussian distribution by using the first power compensation coefficient sample as a mean and a set probability as a variance, and sampling based on the Gaussian distribution to obtain a second power compensation coefficient sample; and determining a second BF weight sample based on the second power compensation coefficient sample, and determining a first average spectral efficiency after performing the BF processing on downlink data sample based on the second BF weight sample;

[0192] performing PAPC on the first BF weight sample to obtain a third BF weight sample, and determining a second average spectral efficiency after performing the BF processing on the downlink data sample based on the third BF weight sample;

[0193] updating the third AI model based on the first average spectral efficiency and the second average spectral efficiency; and

[0194] performing the training steps of a next episode on the updated third AI model by using the training samples of the next episode until a training end condition is satisfied, so as to obtain the second AI model.

[0195] The third AI model can be understood as the second AI model before or during training.

[0196] Optionally, the average spectral efficiency refers to the number of bits correctly transmitted by the downlink data sample (e.g., physical downlink shared channel (PDSCH)) divided by the number of occupied RBs.

[0197] Optionally, the training samples of each episode may be different. For example, the training samples of each episode may be collected in real time (e.g., being trained online), or the training samples of each episode come from communication data of different time units collected in advance (e.g., being trained offline).

[0198] In the embodiment of the present disclosure, the goal of the reinforcement learning is to maximize the average spectral efficiency of the user equipment.

[0199] In other words, the related definitions of the entities and interactions of the reinforcement learning are as follows:

[0200] Agent: a power compensation coefficient estimation model in the AI PAPC, including the third AI model.

[0201] Environment: the wireless channels, the communication modules of the base station and the user equipment, the beamforming process other than the Agent.

[0202] Episode (which can also be understood as an action set): including a round of interaction and data collection process of the Agent and Environment. An episode length may be defined as the number of qualified steps (training steps). The training data collected in the episode come from the communication process of the base station and the same user equipment.

[0203] State: the channel matrix sample output by the environment in each step and the BF weight (first BF weight sample) calculated by the selected BF algorithm.

[0204] Action: first power compensation coefficient sample alpha' for AI PAPC. In the training process of reinforcement learning, the Agent generates the process of exploration to be included in the Action. Specifically, the third AI model estimates the first power compensation coefficient sample alpha', and then constructs a Gaussian distribution by using the alpha' as a mean and a preset probability as a variance. The second power compensation coefficient sample alphasampledmay be sampled and obtained based on the Gaussian distribution, and the Action of this step is generated using the alphasampled. may be greater at the beginning. In order to make the alphasampledhave a uniform value in the range of 0 to 1, with the increase of the training steps, may be gradually decreased, and the probability that alphasampledtakes the value alpha gradually increased, until a training end condition is satisfied (for example, the model converges).

[0205] Reward: it is defined as the average spectral efficiency estimated using the communication technology after the third AI model is applied (power compensation is performed using the coefficient sampled_alpha of the Action). In online learning, there are multiple user equipments in a cell, and the TDD adopts a combined feedback mechanism for multiple downlink slots. The user equipment can only obtain a small amount of acknowledge characters (ACK) / negative acknowledgement (NACK) feedback in one step, so that it is not enough to reflect the real block error rate (BLER) level of the user equipment. In the embodiment of the present disclosure, the Reward and Baseline are defined by using the average spectral efficiency in the episode.

[0206] Advantage (advantage function): it is defined as the result of comparison (e.g., Reward-Baseline) of the first average spectral efficiency and the second average spectral efficiency. The Advantage represents the performance deviation between the Actions selected by the Agent and the existing PAPC in the episode channel. When the average spectral efficiency converted from the coefficient of the Action is higher than that of the existing PAPC, the Advantage is positive; otherwise, the Advantage is negative.

[0207] In the embodiment of the present disclosure, the training process of the second AI model further includes a data collection process. Specifically, as shown in Figure 10, after the training is started, the base station collects the training data of N episodes based on the current third AI model, including the state and action of each step and the advantage of each episode. Further, the training of the third AI model is performed, and the parameters of the third AI model are updated. The base station re-collects the training data based on the updated third AI model, and performs training again. This cycle is repeated until the training converges.

[0208] Figure 11 is a schematic diagram of a data collection process of each episode according to an embodiment of the present disclosure, and Figure 12 is a schematic diagram of an interaction and data collection process in a step according to an embodiment of the present disclosure.

[0209] Specifically, as shown in Figure 11, the data collection process of each episode may comprise the following steps.

[0210] In step 11.1, a user equipment for data collection is screened.

[0211] Optionally, the screening condition may include the following: the user equipment has a continuous data stream in future, for example, the amount of buffered data reported by the downlink BO (radio link control (RLC)) reaches a first threshold; the time-varying characteristic of the channel of the user equipment is low, for example, in the recent SRS reports, the correlation of the channel of the central frequency point reaches a second threshold.

[0212] If there are user equipments satisfying the condition currently, a user equipment is selected randomly or according to a specific rule from the user equipments satisfying the condition to start the data collection of the Episode; otherwise, it waits for a period of time before determination.

[0213] In step 11.2, the Agent interacts with the Environment for one step and collects data.

[0214] In order to reduce the influence on the training of reinforcement learning from the change in channels, in the embodiment of the present disclosure, each step is divided into two adjacent periods T, where the first period T1 is used to measure the Baseline performance in the current channel, and the second period T2 is used to measure the performance after the application of the action in the current channel. Specifically, the existing PAPC (e.g., proportional scaling or full-power PAPC) is applied in T1, and the performance data is collected for the calculation of Baseline; and, the Action generated by the Agent is applied in T2, and the performance data is collected for the calculation of Reward. The length of the period T may be designed in such a way that if the configuration of the user equipment is wideband SRS, T is the SRS period; and, if the configuration of the user equipment is sub-band SRS, T is the SRS period multiplied by the number of sub-bands. The interaction and data collection process in the Step is shown in figure 12.

[0215] Period T1: the user equipment transmits an SRS, and the base station estimates an SRS H and applies the selected BF algorithm and the existing PAPC to calculate a BF weight (e.g., the third BF weight sample) . The base station applies this BF weight into the PDSCH of the user equipment, and the user equipment feeds back ACK / NACK. The number of bits correctly transmitted by the PDSCH and the number of occupied RBs (e.g., the second average spectral efficiency) in T1 are counted for the calculation of Baseline.

[0216] Period T2: the user equipment transmits an SRS, and the base station estimates an SRS H and applies the selected BF algorithm to calculate a BF weight (e.g., the first BF weight sample) . The State (SRS H, the first BF weight sample) is input into the Agent, and the Action (the BF weight with scaled antenna power, e.g., the third BF weight sample ) is output. The base station applies this BF weight into the PDSCH of the user equipment, and the user equipment feeds back ACK / NACK. The number of bits correctly transmitted by the PDSCH and the number of occupied RBs (e.g., the first average spectral efficiency) in T2 are counted for the calculation of Reward.

[0217] In step 11.3, it is checked whether the step data is qualified.

[0218] Optionally, the check condition may include the following: the user equipment channels of T1 and T2 are close enough, for example, the average value (averaging among RBs) of the SRS H correlation values of T1 and T2 reaches a third threshold; and, the user equipment transmits downlink data in T1 and T2, for example, the number of RBs occupied by the PDSCH of the user equipment in T1 and T2 is greater than the smallest fourth threshold.

[0219] In step 11.4, if the data of the step is qualified, the episode length is increased by 1. When the episode length reaches the target length, the episode is successfully collected; otherwise, the process returns to the step 2 to continuously collect data.

[0220] In step 11.5, if the data of the step is unqualified, it is further determined whether the continuous unqualified step data reaches a fifth threshold. If the continuous unqualified step data does not reach the fifth threshold, the process returns to the step 2 to continuously collect data. If the unqualified step data reaches the fifth threshold, it indicates that the channel environment changes sharply or the transmission of the downlink data ends. At this time, the episode can end in advance. At the end of the episode, if the episode length reaches the minimum length, the episode is collected successfully; otherwise, the episode is collected unsuccessfully, and the collected data is discarded.

[0221] In step 11.6, the Episode Advantage is calculated.

[0222] Specifically, Baseline is the ratio of the sum of the number of bits (e.g., ) correctly transmitted in all T1 in the episode to the sum of the number of occupied RBs (e.g., ), as shown by math figure 4; and, Reward is the ratio of the sum of the number of bits (e.g. ) correctly transmitted in all T2 in the episode to the sum of the number of occupied RBs (e.g., ), as shown by math figure 5.

[0223]

[0224]

[0225]

[0226] In the embodiment of the present disclosure, the second AI model learns toward the direction of maximizing expected Advantage (increasing the probability of occurrence of Action with high Advantage), and finally converges to the alpha estimation strategy superior to the existing PAPC algorithm.

[0227] It should be understood by those skilled in the art that, the above training mode is only schematic description and does not constitute limitations on the embodiments of the present disclosure, and appropriate variations made based on these examples can also be applied to the present disclosure. For example, the data collection process shown in Figure 11 may also be performed in advance to collect data for offline training. For another example, it is also possible to directly use the data of one period T to measure the Baseline performance in the current channel and the performance after the application of Action in the current channel or the like, which shall fall into the scope of protection of the present disclosure.

[0228] In the embodiment of the present disclosure, in a case where it is difficult to find the best power scaling coefficient label since there are a large number of 5G / 6G Massive-MIMO antennas and the power scaling factor search space of each antenna is continuous, a method of learning the second AI model online by reinforcement learning is proposed, wherein the advantage function of the reinforce learning is designed by the communication technology, and model training is performed with the goal of maximizing the equivalent channel capacity, so that the model effectively converges to the locally optimal power scaling coefficient, which exceeds the spectral efficiency of the existing PAPC method.

[0229] Figure 13 is a flowchart of an AI-based adaptive beamforming scheme according to an embodiment of the present disclosure.

[0230] Based on at least one of the above embodiments, as shown in Figure 13, an embodiment of the present disclosure provides a complete process of an AI-based adaptive beamforming scheme. The process may comprise the following steps.

[0231] 1) Data preprocessing. The wireless channel information (e.g., PMI, RI, SRS H, etc.; and SRS power, SINR, etc.) from various BF algorithms are input simultaneously. SVD is performed on the SRS H on one or more RBs of the central frequency, and secondary features for measuring the performance loss of the BF algorithm in the current wireless channel are extracted based on the decomposed singular value and singular vector or the related parameters of the ZF algorithm. By constructing the secondary features of the original wireless channel information, it is convenient for AI to determine a BF method suitable for the current scenario more rapidly and accurately.

[0232] 2) BF algorithm classification based on the AI model. Based on the original wireless channel information and the extracted secondary features, the first AI model estimates the BF algorithm (e.g., the first BF algorithm) with highest spectral efficiency in the current wireless channel, and outputs the Id index of this algorithm. Candidate BF methods include, but not limited to, a PMI-based BF algorithm, an SRS-based ZF algorithm, an SRS-based SVD algorithm and the like. The SRS-based SVD algorithm may use a low-complexity LBJ-SVD algorithm.

[0233] 3) BF weight calculation. The BF algorithm corresponding to the input algorithm Id index is run to calculate the BF weight (e.g., the first BF weight) of each PRG. If it is the PMI-based BF algorithm, the step 4) is skipped, and the BF weight finally applied by beamforming is output.

[0234] 4) AI-based PAPC. The BF weight (e.g., the first BF weight) calculated by the SRS-based BF algorithm and the SRS H are input, and the second AI model adjusts the power scaling of each antenna according to the wireless channel and outputs the BF weight (e.g., the second BF weight) finally applied by beamforming.

[0235] The inventor(s) of the present disclosure has (have) conducted system-level simulation based on the proposed intelligent BF scheme, and the simulation scenario is an unlicensed mobile access (UMa) LOS / NLOS mixed scenario. The simulation results show that the average throughput performance of single-user MIMO of the intelligent BF scheme provided in the embodiments of the present disclosure is 9% higher than that of the existing SVD BF, 13% higher than that of the existing ZF BF, and 16% higher than that of the existing PMI BF.

[0236] An embodiment of the present disclosure further provides a method performed by a base station in a communication system, and the method comprises the following steps.

[0237] In step S301, an SRS is acquired.

[0238] For example, it is an UL SRS transmitted by a UE based on a UL SRS resource.

[0239] In step S302, a fourth BF weight is determined based on the SRS.

[0240] For example, the BS calculates the fourth BF weight based on an SRS-based ZF algorithm or an SRS-based SVD algorithm by using the result of channel estimation of the SRS.

[0241] In step S303, a fifth BF weight after performing PAPC on the fourth BF weight is determined based on the fourth BF weight and the SRS measurement related channel matrix (SRS H) by using a fourth AI model.

[0242] In the embodiment of the present disclosure, the AI PAPC can be applied to any one of the existing SRS-based algorithms. According to the wireless channels, the power of each antenna is scaled by using the fourth AI model to realize PAPC, thereby effectively balancing the power loss and the orthogonal loss and achieving higher spectral efficiency.

[0243] In step S304, downlink data is transmitted based on the fifth BF weight.

[0244] In the embodiment of the present disclosure, after the fifth BF weight is determined, the BS may apply the fifth BF weight to the downlink data to change the beam shape and direction so as to achieve the BF effect.

[0245] In the embodiment of the present disclosure, an optional implementation is provided for the step S303. Specifically, the optional implementation may comprise the following steps.

[0246] In step S3031, PAPC is performed on the fourth BF weight to obtain a sixth BF weight.

[0247] For example, proportional scaling PAPC is performed on the fourth BF weight calculated by the SRS-based BF algorithm. The antenna with the highest power of the fourth BF weight is scaled to 1, while the power of other antennas is scaled in equal proportion, so that the scaled sixth BF weight is output.

[0248] In step S3032, a power compensation coefficient is determined based on the sixth BF weight and the channel matrix (SRS H) by using the fourth AI model.

[0249] In the embodiment of the present disclosure, the power compensation coefficient is estimated based on the fourth AI model. The inputs of the fourth AI model are the SRS H and the sixth BF weight scaled in the step S3031. The output power compensation coefficient represents the compensation proportion for the power loss and has a value between 0 and 1 to indicate a balance between the power loss and the orthogonal loss, so it may also be referred to as a power scaling factor.

[0250] In the embodiment of the present disclosure, the structure of the fourth AI model may refer to the above description of the second AI model and will not be repeated here.

[0251] In step S3033, the sixth BF weight is adjusted based on the power compensation coefficient to obtain the fifth BF weight.

[0252] In the embodiment of the present disclosure, the power compensation coefficient is used for the power compensation of the sixth BF weight scaled in the step S3031. Optionally, the sixth BF weight is adjusted based on the power compensation coefficient and an antenna power upper limit.

[0253] In the embodiment of the present disclosure, an optional implementation is provided for the power compensation. Specifically, a third proportion of the antenna power corresponding to the sixth BF weight and a fourth proportion of the antenna power upper limit may be obtained based on the power compensation coefficient, the antenna power corresponding to the third BF weight and the antenna power upper limit are fused based on the third proportion and the fourth proportion, and the fifth BF weight is obtained based on the result of fusion.

[0254] For example, the specific calculation process may refer to the above description of math figure 2 and math figure 3 and will not be repeated here.

[0255] In the embodiment of the present disclosure, the trained fourth AI model can estimate a better power compensation coefficient than the existing PAPC, thereby achieving higher spectral efficiency.

[0256] In the embodiment of the present disclosure, the fourth AI model is learnt online by reinforcement learning. Specifically, the training mode of the fourth AI model, the related definitions of the entities and interactions of the reinforcement learning, and the data collection process may refer to the above description and will not be repeated here.

[0257] Figure 14 is a flowchart of a method performed by a user equipment in a communication system according to an embodiment of the present disclosure.

[0258] In an embodiment of the present disclosure, a method performed by a user equipment in a communication system is further provided. As shown in Figure 14, the method may specifically comprise the following steps.

[0259] In step 401, a reference signal is transmitted, the reference signal being used by a base station to determine first information related to losses of candidate BF algorithms, the first information being determined based on a PMI, singular values and singular vectors, the singular values and the singular vectors being determined based on an SRS measurement related channel matrix.

[0260] In step 402, downlink data is received, the downlink data being transmitted by the base station based on a first BF weight, the first BF weight being determined based on a first BF algorithm, the first BF algorithm being determined from at least two candidate BF algorithms by the base station based on the first information by using a first AI model.

[0261] Optionally, the candidate BF algorithms include at least one of:

[0262] a PMI-based BF algorithm; and

[0263] an SRS-based BF algorithm;

[0264] wherein the SRS-based BF algorithm includes at least one of:

[0265] an SRS-based zero forcing (ZF) algorithm; and

[0266] an SRS-based singular value decomposition (SVD) algorithm.

[0267] Optionally, the first information includes:

[0268] information associated with the result of comparison of the maximum singular value and the minimum singular value in the singular value; and

[0269] a similarity between the singular vector and the PMI.

[0270] Optionally, the first information further includes at least one of:

[0271] a first antenna power variance, which is associated with the BF weight determined by the SRS-based BF algorithm; and

[0272] a second antenna power variance, which is associated with the BF weight after performing per-antenna power constraint (PAPC).

[0273] Optionally, the result of comparison of the maximum singular value and the minimum singular value includes a quotient of the maximum singular value to the minimum singular value; and / or,

[0274] the similarity between the singular vector and the PMI includes a cosine similarity between the singular vector and the PMI.

[0275] Optionally, the wireless channel information further includes at least one of:

[0276] the PMI, a channel matrix, a rank indicator (RI), a channel quality indicator (CQI), receive power for an SRS, and an SRS measurement related signal to interference plus noise ratio (SINR).

[0277] Optionally, in a case where the user equipment is allocated with resources for SRS, the first BF algorithm is determined from the at least two candidate BF algorithms by the base station based on the wireless channel information by using the first AI model.

[0278] Optionally, the downlink data is transmitted by the base station based on the second BF weight, and the second BF weight is a BF weight after performing PAPC on the first BF weight determined by the base station based on the first BF weight and the channel matrix by using the second AI model.

[0279] Optionally, the second BF weight is obtained by adjusting the third BF weight based on the power compensation coefficient, the power compensation coefficient is determined by the base station based on the third BF weight and the channel matrix by using the second AI model, and the third BF weight is obtained by performing PAPC on the first BF weight.

[0280] Optionally, the second BF weight is determined based on the result of fusing the antenna power corresponding to the third BF weight with the antenna power upper limit based on the first proportion related to the antenna power associated with the third BF weight and the second proportion related to the antenna power upper limit, and the first proportion and the second proportion are obtained based on the power compensation coefficient.

[0281] Optionally, the second AI model is trained in the following way:

[0282] performing at least one episode of training steps on the third AI model, wherein the training samples of each episode comprise a first BF weight sample, a channel matrix sample and a downlink data sample, and the training steps of each episode comprise:

[0283] determining a first power compensation coefficient sample based on the first BF weight sample and the channel matrix sample of the episode by using the third AI model, obtaining a Gaussian distribution by using the first power compensation coefficient sample as a mean and a set probability as a variance, and sampling based on the Gaussian distribution to obtain a second power compensation coefficient sample; and determining a second BF weight sample based on the second power compensation coefficient sample, and determining a first average spectral efficiency after performing BF processing on the downlink data sample based on the second BF weight sample;

[0284] performing PAPC on the first BF weight sample to obtain a third BF weight sample, and determining a second average spectral efficiency after performing the BF processing on the downlink data sample based on the third BF weight sample;

[0285] updating the third AI model based on the first average spectral efficiency and the second average spectral efficiency; and

[0286] performing the training steps of a next episode on the updated third AI model by using the training samples of the next episode until a training end condition is satisfied, so as to obtain the second AI model.

[0287] The actions executed in the method performed by a user equipment in a communication system provided in the embodiments of the present disclosure correspond to the steps of the method performed by a base station in a communication system provided in the embodiments of the present disclosure, and the detailed functional description and beneficial effects of the steps can specifically refer to the above description of the corresponding method and will not be repeated here.

[0288] An embodiment of the present disclosure provides an electronic device, comprising: a transceiver configured to transmit and receive signals; and a processor coupled to the transceiver and configured to implement the steps in the above method embodiments. Optionally, the electronic device may be a base station, and the processor is configured to implement the steps in the embodiments of the method performed by the base station. The detailed functional description and the achieved beneficial effects can specifically refer to the above description of the embodiments of the method performed by the base station and will not be repeated here. Optionally, the electronic device may be a user equipment, and the processor is configured to implement the steps in the embodiments of the method performed by the user equipment. The detailed functional description and the achieved beneficial effects can specifically refer to the above description of the embodiments of the method performed by the user equipment and will not be repeated here. In practical applications, the base station or the user equipment can be understood as different network nodes.

[0289] The embodiments of the present disclosure further comprise an electronic device comprising a processor and, optionally, a transceiver and / or memory coupled to the processor configured to perform the steps of the method provided in any of the optional embodiments of the present disclosure.

[0290] Figure 15 is a schematic structure diagram of an electronic device according to an embodiment of the present disclosure.

[0291] As shown in Figure 15, the electronic device 4000 shown in Figure 15 may include a processor 4001 and a memory 4003. The processor 4001 is connected to the memory 4003, for example, through a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, and the transceiver 4004 may be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that, in practical applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation to the embodiments of the present disclosure. Optionally, the electronic device may be a first network node, a second network node or a third network node.

[0292] The processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logical blocks, modules and circuits described in connection with this disclosure. The processor 4001 can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0293] The bus 4002 may include a path to transfer information between the components described above. The bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus or the like. The bus 4002 can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 15, but it does not mean that there is only one bus or one type of bus.

[0294] The memory 4003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, and can also be EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, compact disk storage (including compressed compact disc, laser disc, compact disc, digital versatile disc, blue-ray disc, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium capable of carrying or storing computer programs and capable of being read by a computer, without limitation.

[0295] The memory 4003 is used for storing computer programs for executing the embodiments of the present disclosure, and the execution is controlled by the processor 4001. The processor 2401 is configured to execute the computer programs stored in the memory 4003 to implement the steps shown in the foregoing method embodiments.

[0296] Embodiments of the present disclosure provide a computer-readable storage medium having a computer program stored on the computer-readable storage medium, the computer program, when executed by a processor, implements the steps and corresponding contents of the foregoing method embodiments.

[0297] Embodiments of the present disclosure also provide a computer program product including a computer program, the computer program when executed by a processor realizing the steps and corresponding contents of the preceding method embodiments.

[0298] The terms "first", "second", "third", "fourth", "1", "2", etc. (if present) in the specification and claims of this disclosure and the accompanying drawings above are used to distinguish similar objects and need not be used to describe a particular order or sequence. It should be understood that the data so used is interchangeable where appropriate so that embodiments of the present disclosure described herein can be implemented in an order other than that illustrated or described in the text.

[0299] It should be understood that while the flow diagrams of embodiments of the present disclosure indicate the individual operational steps by arrows, the order in which these steps are performed is not limited to the order indicated by the arrows. Unless explicitly stated herein, in some implementation scenarios of embodiments of the present disclosure, the implementation steps in the respective flowcharts may be performed in other orders as desired. In addition, some, or all of the steps in each flowchart may include multiple sub-steps or multiple phases based on the actual implementation scenario. Some or all of these sub-steps or stages can be executed at the same moment, and each of these sub-steps or stages can also be executed at different moments separately. The order of execution of these sub-steps or stages can be flexibly configured according to requirements in different scenarios of execution time, and the embodiments of the present disclosure are not limited thereto.

[0300] The above text and accompanying drawings are provided as examples only to assist the reader in understanding the present disclosure. They are not intended and should not be construed as limiting the scope of the present disclosure in any way. Although certain embodiments and examples have been provided, based on what is disclosed herein, it will be apparent to those skilled in the art that the embodiments and examples shown may be altered without departing from the scope of the present disclosure. Employing other similar means of implementation based on the technical ideas of the present disclosure also fall within the scope of protection of embodiments of the present disclosure.

Claims

1.A method performed by a base station in a wireless communication system, the method comprising:receiving, from a terminal, a sounding reference signal (SRS) for obtaining a channel matrix;generating first information on each performance loss in a channel for a plurality of candidate beamforming (BF) algorithms, based on a precoding matrix indicator (PMI) and the channel matrix for the SRS;determining a BF algorithm with highest gain among the plurality of the candidate BF algorithms based on the first information, the plurality of the candidate BF algorithms including a PMI based BF algorithm and an SRS-based BF algorithm;determining a BF weight based on the BF algorithm with the highest gain; andtransmitting, to a terminal, downlink data based on the BF weight.2.The method of claim 1,wherein the SRS based BF algorithm includes at least one of an SRS zero forcing (ZF) algorithm and an SRS based singular value decomposition (SVD) algorithm, andwherein the first information includes at least one of information on a channel energy concentration, information on a quantization error of the PMI, and information on a loss of per antenna power constraint (PAPC).3.The method of claim 1,wherein the first information is related to at least one resource block (RB) on a central frequency of a broadband provided by the base station,wherein the BF algorithm with the highest gain is determined further based on second information for the broadband, by using a first artificial intelligence (AI) model, andwherein the second information includes at least one of the PMI, the channel matrix, a rank indicator (RI), and a channel quality indicator (CQI).4.The method of claim 3,wherein in case that the first information is related to a plurality of RBs on the central frequency of the broadband, broadband scale information is generated by fusing the first information for each RB among the plurality of RBs,wherein the broadband scale information is used for the determining a BF algorithm with a highest gain, andwherein the first AI model includes a convolutional neural network (CNN) encoder corresponding to the RB dimension and a global pooling layer used to fuse the first information for the each RB.5.The method of claim 1, further comprising:in case that the BF algorithm with a highest gain is the SRS based BF algorithm, performing a per antenna power constraint (PAPC) on the BF weight;generating a scaled BF weight based on the PAPC;determining a power compensation coefficient for each antenna in the base station based on the scaled BF weight and the channel matrix, by using a second artificial intelligence (AI) model trained based on a reinforcement learning; andupdating the BF weight used for a transmission of the downlink data, based on the power compensation coefficient.6.The method of claim 5, further comprising:determining a first average spectral efficiency for the power compensation coefficient based on the PAPC with the second AI model;determining a second average spectral efficiency for the power compensation coefficient based on the PAPC without the second AI model; andupdating the second AI model based on the comparison between the first average spectral efficiency and the second average spectral efficiency.7.A method performed by a terminal in a wireless communication system, the method comprising:transmitting, to a base station, a sounding reference signal (SRS) used for obtaining a channel matrix; andreceiving, from a base station, a downlink data based on a beamforming (BF) weight,wherein the BF weight is determined based on a BF algorithm with a highest gain, andwherein the BF algorithm with the highest gain is determined based on a precoding matrix indicator (PMI) transmitted from the terminal and the channel matrix.8.A base station in a wireless communication system, the base station comprising:a transceiver; anda processor coupled to the transceiver and configured to:receive, from a terminal, a sounding reference signal (SRS) for obtaining a channel matrix;generate first information on each performance loss in a channel for a plurality of candidate beamforming (BF) algorithms, based on a precoding matrix indicator (PMI) and the channel matrix for the SRS;determine a BF algorithm with highest gain among the plurality of the candidate BF algorithms based on the first information, the plurality of the candidate BF algorithms including a PMI based BF algorithm and an SRS-based BF algorithm;determine a BF weight based on the BF algorithm with the highest gain; andtransmit, to a terminal, downlink data based on the BF weight.9.The base station of claim 8,wherein the SRS based BF algorithm includes at least one of an SRS zero forcing (ZF) algorithm and an SRS based singular value decomposition (SVD) algorithm, andwherein the first information includes at least one of information on a channel energy concentration, information on a quantization error of the PMI, and information on a loss of per antenna power constraint (PAPC).10.The base station of claim 8,wherein the first information is related to at least one resource block (RB) on a central frequency of a broadband provided by the base station,wherein the BF algorithm with the highest gain is determined further based on second information for the broadband, by using a first artificial intelligence (AI) model, andwherein the second information includes at least one of the PMI, the channel matrix, a rank indicator (RI), and a channel quality indicator (CQI).11.The base station of claim 10,wherein in case that the first information is related to a plurality of RBs on the central frequency of the broadband, broadband scale information is generated by fusing the first information for each RB among the plurality of RBs,wherein the broadband scale information is used for the determining a BF algorithm with a highest gain, andwherein the first AI model includes a convolutional neural network (CNN) encoder corresponding to the RB dimension and a global pooling layer used to fuse the first information for the each RB.12.The base station of claim 8, wherein the processor is further configured to:perform a per antenna power constraint (PAPC) on the BF weight, in case that the BF algorithm with a highest gain is the SRS based BF algorithm,generate a scaled BF weight based on the PAPC,determine a power compensation coefficient for each antenna in the base station based on the scaled BF weight and the channel matrix, by using a second artificial intelligence (AI) model trained based on a reinforcement learning, andupdate the BF weight used for a transmission of the downlink data, based on the power compensation coefficient.13.The base station of claim 12, wherein the processor is further configured to:determine a first average spectral efficiency for the power compensation coefficient based on the PAPC with the second AI model,determine a second average spectral efficiency for the power compensation coefficient based on the PAPC without the second AI model, andupdate the second AI model based on the comparison between the first average spectral efficiency and the second average spectral efficiency.14.A terminal in a wireless communication system, the method comprising:a transceiver; anda processor coupled to the transceiver and configured to:transmit, to a base station, a sounding reference signal (SRS) used for obtaining a channel matrix, andreceive, from a base station, a downlink data based on a beamforming (BF) weight,wherein the BF weight is determined based on a BF algorithm with a highest gain, andwherein the BF algorithm with the highest gain is determined based on a precoding matrix indicator (PMI) transmitted from the terminal and the channel matrix.

Images