Method and apparatus for channel noise reduction and shaping in a wireless communication system

Abstract

The present disclosure relates to a 5G communication system or a 6G communication system for supporting higher data rates beyond a 4G communication system such as long term evolution (LTE). The present application relates to a method performed by an electronic apparatus and an electronic apparatus, the method including: determining first channel state information; partitioning the first channel state information into a plurality of pieces of information in a frequency domain according to at least one partition scale; obtaining second channel state information by performing noise reduction and / or shaping through a neural network based on the plurality of pieces of information; and performing channel estimation based on the second channel state information, wherein if a number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a first partition scale, and if the number of the resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale.

Description

METHOD AND APPARATUS FOR CHANNEL NOISE REDUCTION AND SHAPING IN A WIRELESS COMMUNICATION SYSTEM

[0001] The present application relates to the field of the wireless communication, and in particular, the present application relates to a method performed by an electronic apparatus, and an electronic apparatus.

[0002] 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 5G (5th-generation) 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 6G (6th-generation) era, there have been ongoing efforts to develop improved 6G communication systems. For these reasons, 6G communication systems are referred to as beyond-5G systems.

[0003] 6G communication systems, which are expected to be commercialized around 2030, will have a peak data rate of tera (1,000 giga)-level bps and a radio latency less than 100μsec, and thus will be 50 times as fast as 5G communication systems and have the 1 / 10 radio latency thereof.

[0004] In order to accomplish such a high data rate and an ultra-low latency, it has been considered to implement 6G communication systems in a terahertz band (for example, 95GHz to 3THz bands). It is expected that, due to severer path loss and atmospheric absorption in the terahertz bands than those in mmWave bands introduced in 5G, technologies capable of securing the signal transmission distance (that is, coverage) will become more crucial. It is necessary to develop, as major technologies for securing the coverage, radio frequency (RF) elements, antennas, novel waveforms having a better coverage than orthogonal frequency division multiplexing (OFDM), beamforming and massive multiple input multiple output (MIMO), full dimensional MIMO (FD-MIMO), array antennas, and multiantenna transmission technologies such as large-scale antennas. In addition, there has been ongoing discussion on new technologies for improving the coverage of terahertz-band signals, such as metamaterial-based lenses and antennas, orbital angular momentum (OAM), and reconfigurable intelligent surface (RIS).

[0005] 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 collison 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 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 mecahnisms 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.

[0006] 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.

[0007] Embodiments of the present disclosure is to provide an apparatus and method for effectively providing a service in a wireless communication system.

[0008] An object of the present application is to be able to at least resolve one of technical defects in current communication methods to better satisfy communication requirements. In order to realize said object, the present application provides the following technical solutions.

[0009] In an embodiment, a method performed by an electronic apparatus in a wireless communication system is provided. The method includes: determining first channel state information; partitioning the first channel state information into a plurality of pieces of information in a frequency domain according to at least one partition scale; obtaining second channel state information by performing noise reduction and shaping through a neural network based on the plurality of pieces of information; and performing channel estimation based on the second channel state information, wherein if a number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a first partition scale, and wherein if the number of resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale.

[0010] In an embodiment, an electronic apparatus in a wireless communication system is provided. The electronic apparatus includes at least one transceiver; at least one processor communicatively coupled to the at least one transceiver; and at least one memory, communicatively coupled to the at least one processor, storing instructions executable by the at least one processor individually or in any combination to cause the electronic apparatus to determine first channel state information, partition the first channel state information into a plurality of pieces of information in a frequency domain according to at least one partition scale, obtain second channel state information by performing noise reduction and shaping through a neural network based on the plurality of pieces of information, and perform channel estimation based on the second channel state information, wherein if a number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a first partition scale, wherein if the number of resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale, and wherein each partition scale is a number of resource blocks that can be processed by different neural networks at a time.

[0011] According to a first aspect of embodiments of the present application, there is provided a method performed by a network node, including: determining first channel state information; partitioning the first channel state information into a plurality of pieces of information in a frequency domain according to at least one partition scale; obtaining second channel state information by performing noise reduction and / or shaping through a neural network based on the plurality of pieces of information; performing channel estimation based on the second channel state information, wherein if a number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a first partition scale, and if the number of the resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale.

[0012] Alternatively, the method further includes: determining whether the first channel state information satisfies a predetermined constraint condition, wherein if the first channel state information satisfies the predetermined constraint condition, the first channel state information is partitioned into the plurality of pieces of information, wherein when the first channel state information satisfies a first constraint condition related to a signal-to-noise ratio (SNR) and / or a second constraint condition related to a signal stationarity, it is determined that the first channel state information satisfies the predetermined constraint condition.

[0013] Alternatively, the determining whether the first channel state information satisfies the first constraint condition related to the signal-to-noise ratio includes: estimating the signal-to-noise ratio related to the first channel state information according to the first channel state information; and determining whether the first channel state information satisfies the first constraint condition according to a result of comparison of the estimated signal-to-noise ratio to a corresponding threshold.

[0014] Alternatively, the estimating the signal-to-noise ratio related to the first channel state information according to the first channel state information includes: estimating the signal-to-noise ratio related to the first channel state information in the frequency domain or a transform domain different from the frequency domain according to the first channel state information, wherein the determining whether the first channel state information satisfies the first constraint condition according to the result of comparison of the estimated signal-to-noise ratio to the corresponding threshold includes: determining that the first channel state information satisfies the first constraint condition if the estimated signal-to-noise ratio is less than the corresponding threshold.

[0015] Alternatively, the determining whether the first channel state information satisfies the second constraint condition related to the signal stationarity includes: determining a signal stationarity of the first channel state information according to the first channel state information, wherein the signal stationarity is a standard deviation or variance of a derivative of the first channel state information, or an integral of the square of a second-order derivative of the first channel state information; determining that the first channel state information satisfies the second constraint condition if the signal stationarity is greater than a corresponding threshold.

[0016] Alternatively, each partition scale is a number of resource blocks that can be processed by different neural networks at a time.

[0017] Alternatively, the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale includes: if the number of resource blocks of the first channel state information cannot be divided evenly by one partition scale selected from the at least one partition scale, extending or repeating at least a part of the first channel state information so that the extended or repeated first channel state information can be divided evenly by the one partition scale, and partitioning the extended or repeated first channel state information into the plurality of pieces of information according to the one partition scale.

[0018] Alternatively, the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale includes: determining a first extension length; extending both ends of the first channel state information outward by the first extension length, respectively; partitioning the extended first channel state information into the plurality of pieces of information in the frequency domain according one partition scale selected from the at least one partition scale, wherein at least a part of two adjacent ends of two adjacent pieces of information among the plurality of pieces of information is overlapping information.

[0019] Alternatively, the determining the first extension length includes: determining the first extension length according to the number of resource blocks of the first channel state information and the one partition scale.

[0020] Alternatively, the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale includes: partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to different-sized partition scales.

[0021] Alternatively, an operation of the partitioning needs to satisfy following constraints: a number of the partitioned plurality of pieces of information is minimum and a number of resource blocks that need to be extended during the partition is minimum.

[0022] According to a second aspect of the embodiments of the present application, there is provided a network node, including: a transceiver for transmitting and receiving a signal; and a processor coupled to the transceiver and configured to perform the above-described method.

[0023] According to a third aspect of the embodiments of the present application, there is provided a computer readable storage medium storing instructions which, when executed by at least one processor, cause the at least one processor to execute the above-described method.

[0024] The advantageous effects brought by the technical solutions provided by the embodiments of the present application will be illustrated hereinafter with reference to the specific alternative embodiments, or may be learned from the description of the embodiments, or may be acquainted through implementation of the embodiments.

[0025] Embodiments of the present disclosure provides an apparatus and method for effectively providing a service in a wireless communication system.

[0026] In order to illustrate and understand technical solutions of embodiments of the present application more clearly and easily, accompanying drawings that need to be used in the description of the embodiments of the present application will be briefly introduced below.

[0027] FIG. 1 illustrates an example wireless network according to embodiments of the present disclosure.

[0028] FIG. 2 illustrates an example base station according to embodiments of the present disclosure.

[0029] FIG. 3 illustrates an example user equipment according to embodiments of the present disclosure.

[0030] FIG. 4 is a flowchart illustrating a method performed by an electronic apparatus according to an exemplary embodiment of the present disclosure.

[0031] FIG. 5 is a diagram illustrating an example of a frequency-domain blocking method according to an exemplary embodiment of the present disclosure.

[0032] FIG. 6 is a diagram illustrating another example of the frequency-domain blocking method according to an exemplary embodiment of the present disclosure.

[0033] FIG. 7 is a diagram illustrating another example of the frequency-domain blocking method according to an exemplary embodiment of the present disclosure.

[0034] FIG. 8 is a flowchart illustrating a method performed by an electronic apparatus according to another exemplary embodiment of the present disclosure.

[0035] FIG. 9 is a flowchart illustrating a process of training a second neural network according to an exemplary embodiment of the present disclosure.

[0036] FIG. 10 is a schematic diagram illustrating a process of determining a filter window according to an exemplary embodiment of the present disclosure.

[0037] FIG. 11 is a diagram illustrating an example of a transform-domain adaptive method according to an exemplary embodiment of the present disclosure.

[0038] FIG. 12 is a diagram illustrating an example of selecting resource blocks (RBs) according to an exemplary embodiment of the present disclosure.

[0039] FIG. 13 is a block diagram illustrating an electronic apparatus according to an exemplary embodiment of the present disclosure.

[0040] Before undertaking the Mode for Invention below, it may be advantageous to set forth definitions of certain words and phrases used throughout this patent document. The term "couple" and its derivatives refer to any direct or indirect communication between two or more elements, whether those elements are in physical contact with one another. The terms "transmit", "receive", and "communicate", as well as derivatives thereof, encompass both direct and indirect communication. The terms "include" and "comprise", as well as derivatives thereof, mean inclusion without limitation. The term "or" is inclusive, meaning and / or. The phrase "associated with", as well as derivatives thereof, means to include, be included within, interconnect with, contain, be contained within, connect to or with, couple to or with, be communicable with, cooperate with, interleave, juxtapose, be proximate to, be bound to or with, have, have a property of, have a relationship to or with, or the like. The term "controller" means any device, system or part thereof that controls at least one operation. Such a controller may be implemented in hardware or a combination of hardware and software and / or firmware. The functionality associated with any particular controller may be centralized or distributed, whether locally or remotely. The phrase "at least one of", when used with a list of items, means that different combinations of one or more of the listed items may be used, and only one item in the list may be needed. For example, "at least one of: A, B, and C" includes any of the following combinations: A, B, C, A and B, A and C, B and C, and A and B and C. Likewise, the term "set" means one or more. Likewise, the term "set" means one or more. Accordingly, a set of items can be a single item or a collection of two or more items.

[0041] Moreover, various functions described below can be implemented or supported by one or more computer programs, each of which is formed from computer readable program code and embodied in a computer readable medium. The terms "application" and "program" refer to one or more computer programs, software components, sets of instructions, procedures, functions, objects, classes, instances, related data, or a portion thereof adapted for implementation in a suitable computer readable program code. The phrase "computer readable program code" includes any type of computer code, including source code, object code, and executable code. The phrase "computer readable medium" includes any type of medium capable of being accessed by a computer, such as read only memory (ROM), random access memory (RAM), a hard disk drive, a compact disc (CD), a digital video disc (DVD), or any other type of memory. A "non-transitory" computer readable medium excludes wired, wireless, optical, or other communication links that transport transitory electrical or other signals. A non-transitory computer readable medium includes media where data can be permanently stored and media where data can be stored and later overwritten, such as a rewritable optical disc or an erasable memory device.

[0042] Definitions for other certain words and phrases are provided throughout this patent document. Those of ordinary skill in the art should understand that in many if not most instances, such definitions apply to prior as well as future uses of such defined words and phrases.

[0043] 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.

[0044] FIGS. 1-3 below describe various embodiments of the present disclosure implemented in wireless communications systems. The descriptions of FIGS. 1-3 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.

[0045] FIG. 1 illustrates an example wireless network according to embodiments of the present disclosure. The embodiment of the wireless network shown in FIG. 1 is for illustration only. Other embodiments of the wireless network 100 could be used without departing from the scope of the present disclosure.

[0046] As shown in FIG. 1, 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.

[0047] 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; 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 include 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.

[0048] 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).

[0049] 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.

[0050] 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.

[0051] Although FIG. 1 illustrates one example of a wireless network, various changes may be made to FIG. 1. 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.

[0052] FIG. 2 illustrates an example base station according to embodiments of the present disclosure. The embodiment of the gNB 102 illustrated in FIG. 2 is for illustration only, and the gNBs 101 and 103 of FIG. 1 could have the same or similar configuration. However, gNBs come in a wide variety of configurations, and FIG. 2 does not limit the scope of the present disclosure to any particular implementation of a gNB.

[0053] As shown in FIG 2, 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.

[0054] 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.

[0055] 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.

[0056] 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.

[0057] 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.

[0058] 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.

[0059] 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.

[0060] 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).

[0061] Although FIG. 2 illustrates one example of gNB 102, various changes may be made to FIG. 2. For example, the gNB 102 could include any number of each component shown in FIG. 2. 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). For example, various components in FIG. 2 could be combined, further subdivided, or omitted and additional components could be added according to particular needs.

[0062] FIG. 3 illustrates an example user equipment according to embodiments of the present disclosure. The embodiment of the UE 116 illustrated in FIG. 3 is for illustration only, and the UEs 111-115 and 117-119 of FIG. 1 could have the same or similar configuration. However, UEs come in a wide variety of configurations, and FIG. 3 does not limit the scope of the present disclosure to any particular implementation of a UE.

[0063] As shown in FIG. 3, 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.

[0064] 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).

[0065] 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.

[0066] 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.

[0067] 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.

[0068] 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.

[0069] 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.

[0070] Although FIG. 3 illustrates one example of UE 116, various changes may be made to FIG. 3. For example, various components in FIG. 3 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 FIG. 3 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.

[0071] In a 6G wireless communication system, a channel estimation module serves as an important component of a receiver because it provides information of a channel state during signal transmission, which has a crucial impact on performance of the receiver, and thus a quality of a channel estimation algorithm directly affects performance of subsequent equalization and demodulation. High-precision channel information is also a basis for implementation of technologies such as signal decision detection, beamforming and the like.

[0072] Channel estimation algorithms commonly used in the industry have their own advantages and disadvantages. For example, a Least Square (LS) channel estimation algorithm is widely used due to its low complexity and a fact that it does not require any statistical channel information as priori information, but its high channel estimation error limits the performance of the receiver, especially in environments with low signal-to-noise ratios. A Minimum Mean Square Error (MMSE) channel estimation algorithm can greatly improve the performance of the channel estimation algorithm in a low signal-to-noise ratio environment, but its implementation process requires matrix inversion calculation, and high complexity limits its application in actual landing. A Linear Minimum Mean Square Error (LMMSE) channel estimation algorithm uses statistical characteristics of a channel (such as a channel autocorrelation matrix) to improve accuracy of channel estimation. Compared with the traditional LS channel estimation algorithm, the LMMSE channel estimation algorithm can provide better performance under low signal-to-noise ratio (SNR) conditions. Although the LMMSE channel estimation algorithm simplifies the computation compared with the MMSE channel estimation algorithm, in practical applications, the implementation of the LMMSE channel estimation algorithm still requires certain computing resources, especially when processing channel estimation in massive MIMO systems.

[0073] In recent years, neural network-based channel estimation schemes have significantly improved the accuracy of channel estimation under the same reference signal overhead by learning the priori information in the channel structure. Similar to a super-resolution interpolation technology in image processing, these schemes use neural networks to learn a recovery process from received data to an original signal, which can more accurately learn a trend of channel changes, reasonably supplement a missing part of the channel, and correct channel distortion caused by interference or noise. Due to the rapid development of AI technology in recent years, technologies including Deep Learning (DL), Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), Graph Neural Network (GNN) and the like are widely applied to channel estimation. The neural network-based channel estimation schemes have performance advantages, but how to process channel state information flexibly is still a technical problem that needs to be solved.

[0074] FIG. 4 is a flowchart illustrating a method performed by an electronic apparatus according to an exemplary embodiment of the present disclosure. In the present application, the electronic apparatus may be a user terminal, a base station, a component of a base station (such as a Centralized Unit (CU), a Distributed Unit (DU), etc.) or the like.

[0075] As illustrated in FIG. 4, in step S410, first channel state information is determined. Specifically, the first channel state information associated with a reference signal is determined according to the received reference signal.

[0076] In some examples, the received reference signal may be a received signal of a time domain and / or a received signal of a frequency domain. The reference signal may be generated by a transmitter (e.g., a base station, or UE), for example, may be a DMRS signal, a CSI-RS signal. In addition, in the present application, the first channel state information associated with the reference signal may be determined by employing various methods, for example, at least one of following methods may be employed: an LS method, an MMSE method, an LMMSE method, and a Discrete Fourier Transform (DFT) method.

[0077] In step S420, the first channel state information is partitioned into a plurality of pieces of information in a frequency domain according to at least one partition scale. In the present application, this operation may also be referred to as pre-processing. In addition, each partition scale may be a number of resource blocks that can be processed by different neural networks at a time.

[0078] In an exemplary embodiment of the present application, the method may further include: determining whether the first channel state information satisfies a predetermined constraint condition. In an example, if the first channel state information satisfies the predetermined constraint condition, the first channel state information is partitioned into the plurality of pieces of information, wherein when the first channel state information satisfies a first constraint condition related to a signal-to-noise ratio (SNR) and / or a second constraint condition related to a signal stationarity, it is determined that the first channel state information satisfies the predetermined constraint condition. This will be described in detail below.

[0079] Specifically, the determining whether the first channel state information satisfies the first constraint condition related to the signal-to-noise ratio includes: estimating a signal-to-noise ratio related to the first channel state information according to the first channel state information; and determining whether the first channel state information satisfies the first constraint condition according to a result of comparison of the estimated signal-to-noise ratio to a corresponding threshold.

[0080] For example, the signal-to-noise ratio related to the first channel state information may be estimated in the frequency domain, and in particular, the signal-to-noise ratio related to the first channel state information may be estimated in the frequency domain according to the first channel state information. For example, noise estimation may be performed by assuming that a difference between adjacent elements (e.g., sub-carriers) in the frequency domain of the first channel state information mainly comes from noise. Specifically, the signal-to-noise ratio (SNR) related to the first channel state information may be estimated through following [Math Figure 1].

[0081]

[0082] Where, denotes channel state information on a kth sub-carrier, denotes an estimated noise variance for each sub-carrier, and p denotes a channel power of each sub-carrier. The SNR may be a linear expression in the [Math Figure 1], but the present application is not limited hereto, and the SNR may also be such a logarithmic expression . In addition, the above description is made by taking a dimension of a sub-carrier as an example. However, the present application is not limited hereto, and the signal-to-noise ratio related to the first channel state information may further be estimated according to other elements in the frequency domain.

[0083] In the case where the signal-to-noise ratio related to the first channel state information is estimated in the frequency domain, the determining whether the first channel state information satisfies the first constraint condition according to the result of comparison of the estimated signal-to-noise ratio to the corresponding threshold may include: determining that the first channel state information satisfies the first constraint condition if the estimated signal-to-noise ratio is less than the corresponding threshold. In addition, it is determined that the first channel state information does not satisfy the first constraint condition if the estimated signal-to-noise ratio is greater than or equal to the corresponding threshold. In the present application, if the electronic apparatus is a user terminal, the corresponding threshold may be specified by a high-level layer (such as a MAC layer, a base station or a component of the base station), or may be obtained by the electronic apparatus through an optimization operation. If the electronic apparatus is a base station or a component of the base station, the corresponding threshold may be specified by itself (for example, obtained through the optimization operation), or may be specified by a high-level layer (such as the MAC layer).

[0084] For another example, the signal-to-noise ratio related to the first channel state information may be estimated in a transform domain different from the frequency domain. In the present application, the transform domain may be one or a combination of more of following domains: a time-delay domain, a Doppler domain, and an angular domain, but the present application is not limited hereto. Specifically, the signal-to-noise ratio related to the first channel state information may be estimated in the transform domain different from the frequency domain according to the first channel state information. The process of estimating the signal-to-noise ratio related to the first channel state information is described by taking the transform domain being a time-delay domain as an example. For example, the SNR related to the first channel state information may be estimated in the time-delay domain through following [Math Figure 2].

[0085]

[0086] Where, denotes an expression of transforming the first channel state information to the time-delay domain. denotes a Fast Fourier Transform (FFT) length used while transforming the first channel state information to the time-delay domain. denotes a time-delay domain length of an effective multi-path channel, which may be expressed as wherein and denote a start position and an end position of the effective multi-path channel, respectively, and generally, and , wherein denotes a length of a cyclic prefix. In addition, values of and may also be obtained through a neural network. Therefore, represents an average power of noise at each sample, p is a power of the effective multi-path channel, and the SNR may be a linear expression in the above [Math Figure 2], or may also be such a logarithmic expression . In the present application, the effective multi-path channel refers to a multi-path channel that a signal actually goes through, rather than an extended multi-path channel that may exist after the frequency domain channel is transformed.

[0087] In the case where the signal-to-noise ratio related to the first channel state information is estimated in the transform domain different from the frequency domain, the determining whether the first channel state information satisfies the first constraint condition according to the result of comparison of the estimated signal-to-noise ratio to the corresponding threshold may include: determining that the first channel state information satisfies the first constraint condition if the estimated signal-to-noise ratio is less than the corresponding threshold. In addition, it is determined that the first channel state information does not satisfy the first constraint condition if the estimated signal-to-noise ratio is greater than or equal to the corresponding threshold. The corresponding threshold used here may be different from the corresponding threshold for comparison with the signal-to-noise ratio related to the first channel state information estimated in the frequency domain as described above. In addition, similar to the above description, in the present application, if the electronic apparatus is a user terminal, the corresponding threshold may be specified by a high-level layer (such as a MAC layer, a base station or a component of the base station), or may be obtained by the electronic apparatus through an optimization operation. If the electronic apparatus is a base station or a component of the base station, the corresponding threshold may be specified by itself (for example, obtained through the optimization operation), or may be specified by a high-level layer (such as the MAC layer).

[0088] In addition, the determining whether the first channel state information satisfies the second constraint condition related to the signal stationarity includes: determining the signal stationarity of the first channel state information according to the first channel state information; and determining that the first channel state information satisfies the second constraint condition if the signal stationarity is greater than the corresponding threshold. In the present application, the signal stationarity may represent stationary　state of a signal, which may be a standard deviation or variance of a derivative of the first channel state information, or an integral of the square of a second-order derivative of the first channel state information. For example, the integral of the square of the second-order derivative of the first channel state information may be calculated through following [Math Figure 3].

[0089]

[0090] Where, denotes a second-order derivative of the channel state information on a kth sub-carrier. In addition, it is determined that the first channel state information does not satisfy the second constraint condition if the estimated signal-to-noise ratio is less than or equal to the corresponding threshold. The corresponding threshold used here may be different from the corresponding threshold for comparison with the signal-to-noise ratio related to the first channel state information estimated in the frequency domain or the transform domain as described above. Similarly, in the present application, if the electronic apparatus is a user terminal, the corresponding threshold may be specified by a high-level layer (such as a MAC layer, a base station or a component of the base station), or may be obtained by the electronic apparatus through an optimization operation. If the electronic apparatus is a base station or a component of the base station, the corresponding threshold may be specified by itself (for example, obtained through the optimization operation), or may be specified by a high-level layer (such as the MAC layer).

[0091] In another example, if the first channel state information does not satisfy the predetermined constraint condition, the first channel state information is not partitioned, but the channel estimation is directly performed using the first channel state information according to a known channel estimation method.

[0092] In conclusion, the electronic apparatus may simply determine a current channel state through the above process of determining whether the first channel state information satisfies the predetermined constraint condition, and according to a result of the determination, decide whether to partition the first channel state information into the plurality of pieces of information and to start a channel noise reduction and / or shaping operation based on artificial intelligence (AI), and may further save computing power of the electronic apparatus to a certain extent, so that the electronic apparatus may have a larger space for adjustment while adjusting the computing power. The process of partitioning the first channel state information into the plurality of pieces of information in the frequency domain will be described in detail below with reference to the figures.

[0093] In the present application, the first channel state information may be partitioned into the plurality of pieces of information in the frequency domain according to a fixed partition scale (e.g., a first partition scale). For example, the first partition scale may be 3 resource blocks (RBs), that is, the first channel state information may be partitioned into the plurality of pieces of information of 3RBs in the frequency domain (e.g., a dimension of a sub-carrier). Since the number of RBs of the first channel state information may not be divided evenly by the first partition scale, or the number of RBs of the first channel state information is too large, the first channel state information may be partitioned in different forms according to the number of RBs of the first channel state information.

[0094] In an exemplary embodiment, if the number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to the first partition scale, and if the number of resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale, and wherein the first partition scale and the second partition scale each are the number of resource blocks that can be processed by different neural networks at a time. Herein, for the convenience of description, the neural network for noise reduction and / or shaping is referred to as a first neural network. Specifically, assuming that the number of RBs of the first channel state information is N, if N is less than the first predetermined threshold, it may be considered that the number of RBs of the first channel state information is not large, and at the moment, the first channel state information may be partitioned according to a smaller first partition scale (e.g., the first partition scale of a first value 3 RBs). If N is greater than the first predetermined threshold, it may be considered that the number of RBs of the first channel state information is too large, and if the first channel state information is still partitioned according to the smaller first partition scale (e.g., the first partition scale of the first value 3 RBs), the number of the obtained partitioned plurality of pieces of information is too large, which will seriously affect a reasoning complexity of the first neural network. Thus, if N is greater than the first predetermined threshold, the first channel state information may be partitioned into the plurality of pieces of information according to the larger second partition scale (e.g., the second partition scale of a second value 20 RBs) in the frequency domain, which may reduce the reasoning complexity of the first neural network. In the example, sizes (i.e., the first value and the second value) of the first partition scale and the second partition scale are only exemplary, but the present application is not limited hereto. The first neural networks corresponding to different partition scales may be trained and used according to the needs.

[0095] In an example, the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale includes: if the number of resource blocks of the first channel state information cannot be divided evenly by one partition scale selected from the at least one partition scale, extending or repeating at least a part of the first channel state information so that the extended or repeated first channel state information can be divided evenly by the one partition scale, and partitioning the extended or repeated first channel state information into the plurality of pieces of information according to the one partition scale. Wherein, the partition scale for partitioning the first channel state information may be selected from the at least one partition scale in accordance with the method of determining the partition scale according to a result of comparison between the number of resource blocks of the first channel state information and the first predetermined threshold as described above. However, the present application is not limited hereto, the electronic apparatus may select the partition scale for partitioning the first channel state information from the at least one partition scale according to a preset rule, and this may also be understood as selecting the first neural network to be used for noise reduction and / or shaping.

[0096] Specifically, as illustrated in FIG. 5, assuming that the number of RBs of the first channel state information is N, and one partition scale (e.g., the first partition scale) selected from the at least one partition scale is m (e.g., 3) RBs. If N can be divided evenly by 3, then the first channel state information can be partitioned into k = N / 3 pieces of information. If N cannot be divided evenly by 3, then at least a part of the first channel state information is extended or repeated so that the extended or repeated first channel state information can be divided evenly by 3. For example, it is extended or repeated by copying a part of the first channel state information and adding it to the first channel state information (such as adding it to one end, two ends or a certain position in the middle of the first channel state information), that is, repeating at least a part of the first channel state information at a specific position in the first channel state information, and for another example, it is extended by performing extrapolation, mirroring, and other operations on at least a part of the first channel state information. Specifically, for example, a longest part M that can be divided evenly by 3 is taken out from the first channel state information, and a remaining part of less than 3 RBs becomes 3 RBs through extension or repeating. However, the present application is not limited hereto. The extended or repeated at least part may not be the part that is less than m and remains after taking out the longest part M which can be divided evenly by m. The extended or repeated at least part may be any part before taking out the longest part M that can be divided evenly by m, or any part in the process of taking out the longest part M that can be divided evenly by m. The present application does not specifically define the extended or repeated at least part.

[0097] In another example, an overlapping extended blocking method may also be used while partitioning the first channel state information. In detail, the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale includes: determining a first extension length; extending both ends of the first channel state information outward by the first extension length, respectively; partitioning the extended first channel state information into the plurality of pieces of information in the frequency domain according to one partition scale selected from the at least one partition scale, wherein at least a part of two adjacent ends of two adjacent pieces of information among the plurality of pieces of information is overlapping information.

[0098] Specifically, the first extension length may be determined according to the number of resource blocks of the first channel state information and one partition scale, wherein the one partition scale is the number of resource blocks that can be processed by the first neural network at a time. As illustrated in FIG. 6, each of two outermost xs is the first extension length of the extended part, and each of middle xs is an overlapping part, wherein the extended part may be obtained by one of following methods: an extrapolation method, a repetition method, a cyclic method, a mirroring method, etc. In an example illustrated in FIG. 6, in order to facilitate reasoning, it is assumed that the length of the extended part is equal to that of the overlapping part, but the present application is not limited hereto. The length of the extended part may not be equal to that of the overlapping part. For example, the first extension length may be calculated according to following [Math Figure 4].

[0099]

[0100] Where, N denotes the number of RBs of the first channel state information; m denotes the number of RBs that can be processed by the first neural network at a time, that is, one partition scale; denotes rounding up the result of dividing N by m; and y denotes the number of RBs of a non-overlapping part. By solving [Math Figure 4], values of x and y may be obtained, as shown in following [Math Figure 5].

[0101]

[0102] Where, , if N can be divided evenly by m, and if N cannot be divided evenly by m, . Let , then , that is, it may be regarded as that the first channel state information is partitioned into k pieces of information, each of which has a size of the above partition scale m, i.e., may also be referred to as k the common bases of the first neural network. In this case, the above [Math Figure 5] may be simplified to following [Math Figure 6].

[0103]

[0104] For example, if the number of RBs of the first channel state information is N=18, and the common base of the first neural network is m=4, N cannot be divided evenly by m. According to the above [Math Figure 5], it may be obtained that x=1 / 3 RB, y=10 / 3 RB, that is, the first channel state information may be partitioned into k=5 pieces of information in the case where there are overlapping parts.

[0105] For another example, if the number of RBs of the first channel state information is N=24, and the common base of the first neural network is m=4, N can be divided evenly by m. According to the above [Math Figure 6], it may be obtained that x=0, y=4, that is, the first channel state information may be partitioned into k=6 pieces of information in the case where there is no overlapping part. In other words, at this time, the overlapping extended blocking method degenerates into the conventional blocking method described above.

[0106] The above-described method for determining the first extension length is only an example. The present application is not limited hereto, and the first extension length and / or a length of an internal overlapping part may be customized. In addition, it is preferable to ensure that a length of each piece of information of the plurality of pieces of information partitioned through the overlapping extended blocking method is equal.

[0107] In another example, while partitioning the first channel state information, a hybrid blocking method may be used, that is, the number of RBs that can be processed by different first neural networks at a time may be used in hybrid. In other words, a large common base and a small common base of different first neural networks may be used in hybrid. Specifically, the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale may include: partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to different-sized partition scales. Preferably, an operation of the partitioning needs to satisfy following constraints: a number of the partitioned plurality of pieces of information is minimum and a number of resource blocks that need to be extended during the partition is minimum, wherein, each partition scale is a number of resource blocks that can be processed by different first neural networks at a time.

[0108] Specifically, as illustrated in FIG. 7, if the number of RBs that can be processed by two first neural networks at a time is respectively 3 RBs and 20 RBs (that is, the common bases of the two first neural networks are respectively 3 RBs and 20 RBs), and the number of RBs of the first channel state information is 25 RBs (i.e., a frequency-domain length of the first channel state information is 25 RBs), according to the above-described hybrid blocking method, the channel state information may be directly partitioned into two parts of 20 RBs or split into nine parts of 3 RBs in the frequency domain. However, in consideration of reasonably utilizing computing resources, the first channel state information may be partitioned into one part of 20 RBs and two parts of 3 RBs in the frequency domain, but is not separately partitioned into the information of two parts of 20 RBs or nine parts of 3 RBs. That is, according to the above preferable method, the above partitioning operation needs to satisfy following constraints: the number of the partitioned plurality of pieces of information is minimum and the number of RBs that need to be extended while partitioning the first channel state information is minimum. When such a hybrid blocking method satisfies the constraints, the computing resources may be reasonably utilized to avoid following situations: when the first channel state information just exceeds a base size of the first neural network having a large common base, the first channel state information is partitioned into two pieces of information of the size of the large common base, leading to an increase in the computing amount. In other words, this method can select which first neural networks to use according to the dimension of the first channel state information (the number of RBs of the first channel state information) and the dimension (different common bases) of at least one first neural network (i.e. a first neural network group), so as to make more rational use of computing resources and avoid waste of resources.

[0109] In step S430, second channel state information is obtained by performing noise reduction and / or shaping through the first neural networks based on the plurality of pieces of information.

[0110] Specifically, the partitioned plurality of pieces of information are input to a corresponding first neural network for noise reduction and / or shaping, for example, if the first channel state information is partitioned into the plurality of pieces of information according to one partition scale, the partitioned plurality of pieces of information may be input to the first neural network corresponding to the partition scale, that is, input to the first neural network that can process the number of resource blocks corresponding to the partition scale at a time. If the first channel state information is partitioned into the plurality of pieces of information according to two different partition scales, the partitioned plurality of pieces of information may be input to the two first neural networks respectively corresponding to the two partition scales, that is, the information of a corresponding scale is input to one first neural network that can process the number of resource blocks corresponding to one of the partition scales, and the information of a corresponding scale is input to the other first neural network that can process the number of resource blocks corresponding to the other partition scale.

[0111] The second channel state information obtained by performing the noise reduction and / or shaping operation on the partitioned plurality of pieces of information through the first neural networks may have lower noise, and a frequency domain fluctuation is smoother, so that it is more conducive to the operations of subsequent equalization modules, etc. In addition, the above process may further enable the first neural network to adaptively process the first channel state information configured with more RBs without requiring more frequent replacement of neural network models.

[0112] In step S440, the channel estimation is performed based on the second channel state information. For example, the channel estimation may be performed using the second channel state information in accordance with a known channel estimation method.

[0113] FIG. 8 is a flowchart illustrating a method performed by an electronic apparatus according to another exemplary embodiment of the present application.

[0114] In step S810, first channel state information is determined. The operation of the step S810 is the same as that of the step S410, and thus will not be repeated here again.

[0115] In step S820, pre-processing is performed on the first channel state information.

[0116] Specifically, the method may further include: determining whether the first channel state information satisfies a predetermined constraint condition.

[0117] In an example, if the first channel state information satisfies the predetermined constraint condition, the pre-processing is performed on the first channel state information, wherein when the first channel state information satisfies a first constraint condition related to a signal-to-noise ratio and / or a second constraint condition related to a signal stationarity, it is determined that the first channel state information satisfies the predetermined constraint condition. If the first channel state information does not satisfy the predetermined constraint condition, the first channel state information is not pre-processed, but the channel estimation is directly performed using the first channel state information in accordance with a known channel estimation method. Since the similar description of the process of determining whether the first channel state information satisfies the predetermined constraint condition has been made above with reference to FIG. 4, it will not be repeated here again.

[0118] In another exemplary embodiment of the present application, the pre-processing may include performing noise reduction processing on the first channel state information.

[0119] In this case, in an example, the performing the pre-processing on the first channel state information may include: transforming the first channel state information from a frequency domain to a transform domain; determining a filter window according to the transformed first channel state information; filtering the transformed first channel state information through the filter window; and transforming a filtering result to the frequency domain.

[0120] For example, taking the transform domain being a time-delay domain as an example, in the case where the first channel state information is transformed to the time-delay domain through frequency domain Fourier transform for example, the filter window is determined according to the transformed first channel state information, and the transformed first channel state information is filtered through following [Math Figure 7].

[0121]

[0122] Where, denotes the first channel state information that has been transformed to the time-delay domain; denotes a filter window in the time-delay domain; denotes a windowed signal; and t denotes a time-delay position of a signal in the time-delay domain.

[0123] In the present application, the filter window may be determined through a neural network. In the present application, for the convenience of description, the neural network for determining the filter window is referred to as a second neural network. In an example, the determining the filter window according to the transformed first channel state information may include: based on the transformed first channel state information, determining a length of the filter window and a start position and an end position of the filter window through the second neural network, that is, determining the filter window in the above filtering process. Specifically, may be expressed as following [Math Figure 8].

[0124]

[0125] In the example, is a rectangular window, wherein the length of the filter window and the start position t0and the end position t1of the filter window are obtained by performing prediction by the second neural network. A training process of the second neural network is described with reference to FIG. 9 below.

[0126] FIG. 9 is a flow chart illustrating a process of training a second neural network according to an exemplary embodiment of the present disclosure.

[0127] As illustrated in FIG. 9, in step S910, a second label is obtained by transforming a first label corresponding to channel state information to a transform domain. For example, taking the transform domain being a time-delay domain as an example, assuming that a label corresponding to channel state information is , a second label is obtained by transforming to the time-delay domain through inverse Fourier transform.

[0128] In step S920, each transform-domain sample of the second label is compared to a second predetermined threshold, and a filter window value is set for a position of each transform-domain sample. If a certain transform-domain sample of the second label is greater than or equal to the second predetermined threshold, a filter window value corresponding to a position of the transform-domain sample is set to a first filter window value, and otherwise set to a second filter window value. For example, taking the transform domain being a time-delay domain as an example, each time-delay-domain sample of the second label is compared to the second predetermined threshold , and if a certain time-delay-domain sample of the second label is greater than or equal to the second predetermined threshold , a filter window value corresponding to a position of the time-delay-domain sample is set to a first filter window value 1, and otherwise set to a second filter window value 0. For example, the above process may be expressed as [Math Figure 9].

[0129]

[0130] In step S930, a third label of the filter window corresponding to the channel state information is determined by counting a sum of power of channels at the positions having the first filter window value. For example, taking the transform being a time-delay domain as an example, a shape of the filter window is modified by counting a sum of multi-path power of the channels at the positions of . Specifically, the modified window may be obtained by retaining continuous rectangular windows with at least 90% summed power and deleting the remaining rectangular windows, so that the third label of the filter window corresponding to the channel state information may be determined. For example, in conjunction with FIG. 10, the above process may be expressed as [Math Figure 10].

[0131]

[0132] In step S940, the second neural network is trained based on the third label. Specifically, the third label of the filter window corresponding to the channel state information may be obtained through the above steps S910 to S930, then the second neural network may be trained by a binary classification method. The trained second neural network may be used to predict the length of the filter window and the start position and the end position of the filter window. In addition, the examples are enumerated for illustration by taking the transform domain being the time-delay domain as an example, but the present application is not limited hereto, and transform domain may further be a Doppler domain, an angular domain, or the like.

[0133] The above method for predicting the filter window through the second neural network may enable the predicted filter window to retain effective information under various channel conditions more generally, and remove the noise in the transform domain on the premise of retaining the effective information as much as possible. For example, an effective time-delay path under various channel conditions may be retained in the time-delay domain, and then the noise in the time-delay domain may be removed as much as possible. In addition, if the first channel state information has time offset, this method may have better time offset robustness under the premise of ensuring performance.

[0134] In the present application, in addition to predicting the filter window through the second neural network, other methods may also be used to determine the filter window. Specifically, the determining the filter window according to the transformed first channel state information may include: determining noise power according to the first channel state information that has benn transformed to the transform domain, determining signal power of each transform domain position according to the first channel state information that has been transformed to the transform domain, and determining the filter window according to the noise power and the signal power of the each transform domain position.

[0135] For example, taking the transform domain being a time-delay domain as an example, the first channel state information may be transformed to the time-delay domain through frequency domain Fourier transform, and then the filter window is determined through following [Math Figure 11].

[0136]

[0137] Where, is the noise power calculated in the time-delay domain, and is the signal power at the time-delay domain position t. Since a channel impulse response in the time-delay domain is not continuous, there may be blank moments between the impulse responses, and there is an influence of noise at these moments. The filter window determined by the above method may adaptively adjust a size of the filter window according to the signal-to-noise ratios at different time-delay domain positions, to ensure that effective signals are not cut off together to the greatest extent while removing the noise between the impulse responses as much as possible.

[0138] In another example, the performing the pre-processing on the first channel state information may include: transforming the first channel state information to a transform domain; performing convolution processing on the transformed first channel state information; and transforming a convolution result to a frequency domain. For example, taking the transform domain being a time-delay domain as an example, the first channel state information is transformed to the time-delay domain through frequency domain Fourier transform, a convolution window is set to , and the convolution processing performed on the transformed first channel state information may be expressed as following [Math Figure 12].

[0139]

[0140] Where, denotes the first channel state information that has been transformed to the time-delay domain; denotes a convolution-processed signal. The description is made by taking the transform domain being the time-delay domain as an example, but the present application is not limited hereto. The noise reduction processing may also be performed through the convolution processing in other transform domains.

[0141] In another exemplary embodiment of the present application, the pre-processing may include performing noise reduction processing and scale change on the first channel state information.

[0142] Specifically, the performing the pre-processing on the first channel state information may include: transforming the first channel state information from a frequency domain to a transform domain; performing noise reduction on the transformed first channel state information; and transforming the transformed first channel state information subjected to the noise reduction to the frequency domain according to a transformation size corresponding to the number of resource blocks that can be processed by the first neural network at a time. That is, the pre-processing may be the noise reduction processing and scale change performed on the transformed first channel state information, wherein the scale change operation enables a scale of finally obtained information to correspond to the number of resource blocks that can be processed by the first neural network at a time.

[0143] For example, as illustrated in FIG. 11, assuming that the number of RBs of the first channel state information is N, the first channel state information is transformed from the frequency domain to the time-delay domain through inverse Fourier transform. Then, the noise reduction is performed on the transformed first channel state information in the time-delay domain. In the present application, the noise reduction processing may be various types of noise reduction processing such as the convolution operation, the filtering operation performed using the filter window, etc. as described above, for example, [Math Figure 7] and [Math Figure 12]. Since the detailed description has been made above, it will not be repeated here. Thereafter, the transformed first channel state information subjected to the noise reduction is transformed to the frequency domain through the Fourier transform. At this moment, the Fourier transformed size corresponds to the number of RBs N1 that can be processed by the first neural network at a time. That is, the Fourier transformed size is a size N1 of a common base of the first neural network, so that the scale of the channel state information that has been transformed to the frequency domain corresponds to the common base of the first neural network.

[0144] In step S830, the second channel state information is obtained by performing noise reduction and / or shaping by the first neural network based on the pre-processed first channel state information. Specifically, if the noise reduction processing is performed on the first channel state information by only adopting the above-described noise reduction operation (without performing scale change), the first neural network may be selected to perform the noise reduction and / or shaping according to the needs. If the first channel state information is pre-processed by using the above-described noise reduction and scale change operation, the first neural network corresponding to the transformation size used while changing the scale may be used to perform noise reduction and / or shaping. In step S840, the channel estimation is performed based on the second channel state information, for example, the channel estimation may be performed using the second channel state information in accordance with a known channel estimation method.

[0145] Various method steps described above with reference to FIG. 4 to FIG. 11 may be used in combination with each other. For example, after the first channel state information is partitioned into the plurality of pieces of information in the frequency domain, the noise reduction method described above with reference to the step S820 is used to perform noise reduction processing on each partitioned piece of information, and then the first neural network corresponding to the scale of each piece of information is selected to perform noise reduction and / or shaping. For another example, after the noise reduction processing is performed on the first channel state information according to the noise reduction method described above with reference to the step S820, the first channel state information subjected to the noise reduction processing may be partitioned into the plurality of pieces of information using the partition method described above with reference to the step S420, and further the second channel state information is obtained by performing the noise reduction and / or shaping through the first neural network based on the plurality of pieces of information. Some examples in which the various method steps described above with reference to FIG. 4 to FIG. 11 are partially combined with each other are enumerated only in an example form here, and other combined methods are also included within the scope of the present application.

[0146] In addition, in the present application, the first neural network for noise reduction and / or shaping operation mendtioned above may be trained and perform an inferring operation separately from the second neural network for determining the filter window, and may also be trained and / or perform the inferring operation with the second neural network as an entire neural network.

[0147] For example, the first neural network and the second neural network can be updated online separately or as a whole. Since the neural network needs to make generalizability adjustment on the number of RBs for different scheduling, in the present application, the number of RBs of the input channel state information may have a certain randomness when the parameters of the neural network are updated each time. Specifically, when the first neural network and the second neural network are trained separately or as a whole, training data may be input according to the fixed number of RBs or the random number of RBs. If the training data is input according to the fixed number of RBs, the fixed number of RBs may be a pre-configured length. If the training data is input according to the random number of RBs, one of a start point and an end point of each set of the number of RBs may be configured as a random integer. In addition, in the present application, when the first neural network and the second neural network are trained separately or as a whole after the number of RBs is configured, the RBs input to the neural network each time may be selected from the training data according to one or a combination of more of the following selecting manners: a first manner of continuously selecting RBs and a second manner of selecting RBs at equal interval.

[0148] In an exemplary embodiment of the present application, the manner of the combination of the first manner and the second manner is preferably selected to select the RBs input to the neural network each time from the training data, which may greatly expand the generalization of the neural network and improve the universality of the neural network in different time-delay channels.

[0149] In another exemplary embodiment of the present application, the one of or combination of more of the above selecting manners may be selected from the first manner and the second manner according to a channel condition, wherein the channel condition is distribution of paths on which the time-delay domain impulse response of the training data is greater than a third predetermined threshold during a previous period of time. For example, if the distribution of the selected paths is concentrated, it indicates that the multi-path effect is low at this time and the channel frequency domain correlation is high, thereby preferably selecting the first manner of continuously selecting RBs. If the distribution of the selected paths is more dispersed, it indicates that the multi-path effect is higher at this time and the channel frequency domain correlation is lower, thereby preferably selecting the second manner of selecting RBs at equal interval. For example, as illustrated in FIG. 12, if each time the neural network is updated online, a resource of W RBs is scheduled at a high-level layer, and an electronic apparatus as a receiver selects a random number of RBs for the online update of the model, the electronic apparatus may randomly determine a starting point a of the number of RBs and / or an end point b of the number of RBs and select RBs using the first manner of continuously selecting RBs, thereby ultimately resulting in the acquisition of input data with a length of L=b-a, and utilizing these pieces of acquired input data to train the neural network. However, the present application is not limited hereto. The RBs, which are input to the neural network each time, may also be selected from the training data in accordance with a third manner of randomly selecting RBs, or the RBs, which are input to the neural network each time, may be selected in accordance with a combination of a plurality of manners among the first manner, the second manner, and the third manner.

[0150] In addition, in the present application, while updating the model (e.g., the first neural network and / or the second neural network), an update period thereof may be at least one of following manners: periodic update, semi-periodic update, and trigger-based update. The periodic update refers to updating a model by a period of , and the period may be configured by a high-level layer. The semi-periodic update refers to that an electronic apparatus as a receiver starts to update a model periodically in response to receiving a high-level layer instruction, and terminates the periodic update in response to receiving another high-level layer instruction after a period of time. The trigger-based update refers to that an electronic apparatus as a receiver updates a model once only when it receives a high-level layer instruction for indicating model update.

[0151] In addition, in the present application, the electronic apparatus may further decide whether to update the model according to a current computing power load, for example, if the current computing power load of the electronic apparatus is greater than a fourth predetermined threshold, the electronic apparatus may decide not to update the model, and if the current computing power load of the electronic apparatus is less than or equal to the fourth predetermined threshold, the electronic apparatus may decide to increase an update frequency of the model. In the present application, the fourth predetermined threshold may be set according to a practical situation. In this way, the computing resources of the electronic apparatus may be utilized more reasonably.

[0152] FIG. 13 is a block diagram illustrating an electronic apparatus 1300 according to an exemplary embodiment of the present disclosure.

[0153] As illustrated in FIG. 13, the electronic apparatus 1300 may include a transceiver 1320 and a processor 1310, wherein the processor 1310 is coupled to the transceiver 1320, and is configured to perform the methods described above with reference to FIG. 4 to FIG. 12. For details of operations of the methods, reference may be made to the descriptions of FIG. 4 to FIG. 12, which will not be repeated here again.

[0154] In addition, according to the embodiments of the present disclosure, there may further be provided an electronic apparatus, including: at least one processor; and at least one memory storing computer executable instructions, wherein the computer executable instructions, when executed by the at least one processor, cause the at least one processor to execute the above-described methods.

[0155] As an example, the electronic apparatus may be a PC computer, a tablet device, a personal digital assistant, a smart phone, or other device capable of executing the above indication set. Here, the electronic apparatus does not have to be a single electronic apparatus, and may also be any aggregate of devices or circuits that can execute the above-mentioned indications (or indication sets) individually or jointly. The electronic apparatus may also be a part of an integrated control system or a system manager, or may configured as a portable electronic device that is interconnected with the local or remote (e.g., via wireless transmission) via interfaces.

[0156] In the electronic apparatus, the processor may include a central processing unit (CPU), a graphics processing unit (GPU), a programmable logic device, a dedicated processor system, a microcontroller, or a microprocessor. As an example rather than limitation, the processor may also include an analog processor, a digital processor, a microprocessor, a multi-core processor, a processor array, a network processor, and the like.

[0157] The processor may run instructions or codes stored in the memory, wherein the memory may also store data. Instructions and data may also be transmitted and received over a network via a network interface device, wherein the network interface device may use any known transmission protocol.

[0158] The memory may be integrated with the processor, for example, RAM or flash memory is arranged in an integrated circuit microprocessor or the like. In addition, the memory may include a separate device, such as an external disk drive, a storage array, or any other storage device that can be used by a database system. The memory and the processor may be operatively coupled, or may communicate with each other, for example, through an I / O port, a network connection or the like, so that the processor can read files stored in the memory.

[0159] In addition, the electronic apparatus may also include a video display (such as a liquid crystal display) and a user interaction interface (such as a keyboard, a mouse, a touch input device, etc.). All components of the electronic apparatus may be connected to each other via a bus and / or a network.

[0160] According to the embodiments of the present disclosure, there there may further be provided a computer readable storage medium storing instructions, wherein the instructions, when executed by at least one processor, cause the at least one processor to execute the above-described methods according to the exemplary embodiments of the present application. Examples of the computer-readable storage medium here include: Read Only Memory (ROM), Random Access Programmable Read Only Memory (PROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Random Access Memory (RAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), flash memory, non-volatile memory, CD-ROM, CD-R, CD+R, CD-RW, CD+RW, DVD-ROM, DVD-R, DVD+R, DVD- RW, DVD+RW, DVD-RAM, BD-ROM, BD-R, BD-R LTH, BD-RE, blu-ray or optical disc storage, Hard Disk Drive (HDD), Solid State Drive (SSD), card storage (such as, multimedia cards, secure digital (SD) cards or extreme speed digital (XD) cards), magnetic tapes, floppy disks, magneto-optical data storage devices, optical data storage devices, hard disks, solid state disks, and any other devices that are configured to store computer programs and any associated data, data files and data structures in a non-transitory manner and provide the computer programs and any associated data, data files and data structures to the processor or computer so that the processor or computer may execute the computer programs. The instructions or computer programs in the above-mentioned computer-readable storage medium may run in an environment deployed in a computer apparatus such as a client, a host, a proxy device, a server, or the like. In addition, in one example, the computer program and any associated data, data files and data structures are distributed on networked computer systems, so that the computer programs and any associated data, data files and data structures are stored, accessed, and executed in a distributed manner through one or more processors or computers.

[0161] It should be explained that terms "first", "second", "third", "fourth", "1", "2" and the like (if exist) in the description and the claims as well as the drawings of the present application are used to distinguish similar objects, and are not necessary to be used to describe specific order or sequence. It should be understood that the data used in this way may be interchanged under appropriate circumstances so that the embodiments of the present application described herein may be implemented in an order other than those illustrated or described herein.

[0162] It should be understood that although various operation steps are indicated by arrows in the flowcharts of the embodiments of the present application, the order in which these steps are implemented is not limited to the order indicated by the arrows. The implementation steps in each of the flowcharts may be carried out in other orders according to the requirements in some application scenarios of the embodiments of the present application, unless specifically stated. In addition, some or all of the steps in each flowchart may include a plurality of sub steps or stages based on an actual implementation scenario. Some or all of these sub steps or stages may be executed at the same time, and each sub step or stage in these sub steps or stages may also be executed at different times respectively. In scenarios with different execution times, an execution order of these sub-steps or stages may be flexibly configured according to needs, and not be limited by the embodiments of the present application.

[0163] Only optional embodiments of some of the embodiment scenarios of the present application are described above, and it should be indicated that for those ordinarily skilled in the art, without departing from the technical conception of the solution of the present application, the employment of other similar means of implementation based on the technical ideas of the present application also pertain to the scope of protection of the embodiments of the present application.

Claims

1.A method performed by an electronic apparatus in a wireless communication system, the method comprising:determining first channel state information;partitioning the first channel state information into a plurality of pieces of information in a frequency domain according to at least one partition scale;obtaining second channel state information by performing noise reduction and shaping through a neural network based on the plurality of pieces of information; andperforming channel estimation based on the second channel state information,wherein if a number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a first partition scale, andwherein if the number of resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale.2.The method of claim 1, further comprising:determining whether the first channel state information satisfies a predetermined constraint condition,wherein if the first channel state information satisfies the predetermined constraint condition, the first channel state information is partitioned into the plurality of pieces of information, andwherein when the first channel state information satisfies a first constraint condition related to a signal-to-noise ratio and a second constraint condition related to signal stationarity, it is determined that the first channel state information satisfies the predetermined constraint condition.3.The method of claim 2, wherein the determining whether the first channel state information satisfies the first constraint condition related to the signal-to-noise ratio comprises:estimating the signal-to-noise ratio related to the first channel state information in the frequency domain or a transform domain different from the frequency domain according to the first channel state information; anddetermining that the first channel state information satisfies the first constraint condition if the estimated signal-to-noise ratio is less than the corresponding threshold.4.The method of claim 2, wherein the determining whether the first channel state information satisfies the second constraint condition related to the signal stationarity comprises:determining a signal stationarity of the first channel state information according to the first channel state information, wherein the signal stationarity is a standard deviation or variance of a derivative of the first channel state information, or an integral of the square of a second-order derivative of the first channel state information; anddetermining that the first channel state information satisfies the second constraint condition if the signal stationarity is greater than a corresponding threshold.5.The method of claim 1, wherein each partition scale is a number of resource blocks that can be processed by different neural networks at a time.6.The method of claim 1, wherein the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale comprises:if the number of resource blocks of the first channel state information cannot be divided evenly by one partition scale selected from the at least one partition scale, extending or repeating at least a part of the first channel state information so that the extended or repeated first channel state information can be divided evenly by the one partition scale, and partitioning the extended or repeated first channel state information into the plurality of pieces of information according to the one partition scale.7.The method of claim 1, wherein the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to the at least one partition scale comprises:determining the first extension length according to the number of resource blocks of the first channel state information and the one partition scale;extending both ends of the first channel state information outward by the first extension length, respectively; andpartitioning the extended first channel state information into the plurality of pieces of information in the frequency domain according to one partition scale selected from the at least one partition scale,wherein at least a part of two adjacent ends of two adjacent pieces of information among the plurality of pieces of information is overlapping information.8.The method of claim 1, wherein the partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to at least one partition scale comprises:partitioning the first channel state information into the plurality of pieces of information in the frequency domain according to different-sized partition scales,wherein a number of the partitioned plurality of pieces of information is minimum and a number of resource blocks that need to be extended during the partition is minimum.9.An electronic apparatus in a wireless communication system, the electronic apparatus comprising:at least one transceiver;at least one processor communicatively coupled to the at least one transceiver; andat least one memory, communicatively coupled to the at least one processor, storing instructions executable by the at least one processor individually or in any combination to cause the electronic apparatus to:determine first channel state information,partition the first channel state information into a plurality of pieces of information in a frequency domain according to at least one partition scale,obtain second channel state information by performing noise reduction and shaping through a neural network based on the plurality of pieces of information, andperform channel estimation based on the second channel state information,wherein if a number of resource blocks of the first channel state information is less than a first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a first partition scale,wherein if the number of resource blocks of the first channel state information is greater than or equal to the first predetermined threshold, the first channel state information is partitioned into the plurality of pieces of information in the frequency domain according to a second partition scale, wherein the second partition scale is greater than the first partition scale, andwherein each partition scale is a number of resource blocks that can be processed by different neural networks at a time.10.The electronic apparatus of claim 9, wherein the instructions executable by the at least one processor individually or in any combination further cause the electronic apparatus to:determine whether the first channel state information satisfies a predetermined constraint condition,wherein if the first channel state information satisfies the predetermined constraint condition, the first channel state information is partitioned into the plurality of pieces of information, andwherein when the first channel state information satisfies a first constraint condition related to a signal-to-noise ratio and a second constraint condition related to signal stationarity, it is determined that the first channel state information satisfies the predetermined constraint condition.11.The electronic apparatus of claim 10, wherein the instructions executable by the at least one processor individually or in any combination further cause the electronic apparatus to:estimate the signal-to-noise ratio related to the first channel state information in the frequency domain or a transform domain different from the frequency domain according to the first channel state information, anddetermine that the first channel state information satisfies the first constraint condition if the estimated signal-to-noise ratio is less than the corresponding threshold.12.The electronic apparatus of claim 10, wherein the instructions executable by the at least one processor individually or in any combination further cause the electronic apparatus to:determine a signal stationarity of the first channel state information according to the first channel state information, wherein the signal stationarity is a standard deviation or variance of a derivative of the first channel state information, or an integral of the square of a second-order derivative of the first channel state information, anddetermine that the first channel state information satisfies the second constraint condition if the signal stationarity is greater than a corresponding threshold.13.The electronic apparatus of claim 9, wherein the instructions executable by the at least one processor individually or in any combination further cause the electronic apparatus to:if the number of resource blocks of the first channel state information cannot be divided evenly by one partition scale selected from the at least one partition scale, extend or repeat at least a part of the first channel state information so that the extended or repeated first channel state information can be divided evenly by the one partition scale, and partition the extended or repeated first channel state information into the plurality of pieces of information according to the one partition scale.14.The electronic apparatus of claim 9, wherein the instructions executable by the at least one processor individually or in any combination further cause the electronic apparatus to:determine the first extension length according to the number of resource blocks of the first channel state information and the one partition scale,extend both ends of the first channel state information outward by the first extension length, respectively, andpartition the extended first channel state information into the plurality of pieces of information in the frequency domain according to one partition scale selected from the at least one partition scale,wherein at least a part of two adjacent ends of two adjacent pieces of information among the plurality of pieces of information is overlapping information.15.The electronic apparatus of claim 9, wherein the instructions executable by the at least one processor individually or in any combination further cause the electronic apparatus to:partition the first channel state information into the plurality of pieces of information in the frequency domain according to different-sized partition scales,wherein a number of the partitioned plurality of pieces of information is minimum and a number of resource blocks that need to be extended during the partition is minimum.

Images