Communication device in wireless communication system and method performed by the communication device

Abstract

The present disclosure relates to a communication method and system for converging a 5th-Generation (5G) communication system or a 6th-Generation (6G) communication system for supporting higher data rates beyond a 4th-Generation (4G) system. According to an exemplary embodiment, a method performed by a communication device in a wireless communication system comprises: determining, based on a neural network related measurement, whether to transmit a first message for requesting a signal for a neural network calibration; and if the first message is transmitted, receiving at least two signals for the neural network calibration, wherein the signals for the neural network calibration are transmitted through one antenna port, and time domain resources and / or frequency domain resources carrying the signals for the neural network calibration satisfy a first condition.

Description

COMMUNICATION DEVICE IN WIRELESS COMMUNICATION SYSTEM AND METHOD PERFORMED BY THE COMMUNICATION DEVICE

[0001] The present disclosure relates to a communication device in a wireless communication system and a method performed by the communication device.

[0002] Considering the development of mobile 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, multiantenna transmission technologies including radio frequency (RF) elements, antennas, novel waveforms having a better coverage than OFDM, beamforming and massive MIMO, full dimensional MIMO (FD-MIMO), array antennas, and 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 (UE transmission) and a downlink (node B 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 nodes B and the like and enabling network operation optimization and automation and the like; an use of AI in wireless communication for improvement of overall network operation by considering AI from the initial phase of developing technologies for 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 (MEC, clouds, and the like) over the network.

[0006] It is expected that such research and development of 6G communication systems will bring the next hyper-connected experience to every corner of life. Particularly, it is expected that services such as truly immersive XR, high-fidelity mobile hologram, and digital replica could be provided through 6G communication systems.

[0007] Considering the development of wireless communication from generation to generation, the technologies have been developed mainly for services targeting humans, such as voice calls, multimedia services, and data services. Following the commercialization of 5th-generation (5G) communication systems, it is expected that the number of connected devices will exponentially grow. Increasingly, these will be connected to communication networks. Examples of connected things may include vehicles, robots, drones, home appliances, displays, smart sensors connected to various infrastructures, construction machines, and factory equipment. Mobile devices are expected to evolve in various form-factors, such as augmented reality glasses, virtual reality headsets, and hologram devices. In order to provide various services by connecting hundreds of billions of devices and things in the 6th-generation (6G) era, there have been ongoing efforts to develop improved 6G communication systems.

[0008] 6G communication systems, which are expected to be commercialized around 2030, have various significantly improved metrics compared to the current 5G communication systems. The peak data rate will reach at least 50 Gbit / s, and the user experienced data rate will reach at least 300 Mbit / s, the air-interface latency will be less than 1 ms, and the air-interface reliability will reach 10^(-5). In addition to the above basic communication metrics, the 6G communication systems will also have sensing capabilities, AI-related capabilities, better security, better interoperability and better sustainability.

[0009] In order for the 6G communication systems to fulfill the above metrics, more advanced air-interface technologies and network technologies need to be developed. The evolution of extreme Multiple Input Multiple Output (extreme MIMO) has been already under consideration, including the use of ultra-large scale antenna arrays, the development and evolution of distributed antenna systems, and the design of MIMO air-interface algorithms assisted by Artificial Intelligence (AI). This technology enables higher spectral efficiency, greater coverage, and precise localization and sensing capabilities. Additionally, for technologies that contribute to improve high-frequency band coverage, including metamaterial-based lenses and antennas, new antenna architectures, and reconfigurable intelligent surface (RIS), etc., they also need to be better evolved and developed.

[0010] In order to meet some of newly added functions of the 6G communication systems, new technologies need to be developed in the terms of network energy saving, air-interface security, and network security, meanwhile the feasibility of fusion technologies such as Integrated Sensing and Communication, needs to be studied.

[0011] Moreover, in order to improve the spectral efficiency and the overall network performances, the following technologies have been developed for 6G communication systems: a full-duplex technology for enabling an uplink transmission and a downlink transmission to simultaneously use the same frequency resource at the same time; a network technology for utilizing satellites, high-altitude platform stations (HAPS), and the like in an integrated manner; an improved network structure for supporting mobile base stations and the like and enabling network operation optimization and automation and the like; a dynamic spectrum sharing technology via collision avoidance based on a prediction of spectrum usage; an use of artificial intelligence (AI) in wireless communication for improvement of overall network operation by utilizing AI from a designing phase for developing 6G and internalizing end-to-end AI support functions; and a next-generation distributed computing technology for overcoming the limit of user equipment (UE) computing ability through reachable super-high-performance communication and computing resources (such as mobile edge computing (MEC), clouds, and the like) over the network. In addition, through designing new protocols to be used in 6G communication systems, developing mechanisms for implementing a hardware-based security environment and safe use of data, and developing technologies for maintaining privacy, attempts to strengthen the connectivity between devices, optimize the network, promote softwarization of network entities, and increase the openness of wireless communications are continuing.

[0012] 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.

[0013] The technical problem addressed by the present disclosure is that, in a wireless communication system, unnecessary signaling overhead, resource waste, and latency may occur if calibration signals for a neural network are requested and transmitted without consideration of actual measurement results. In addition, the accuracy and reliability of neural network calibration may be degraded if such calibration signals are transmitted through arbitrary antenna ports or over non-optimized time and frequency resources.

[0014] According to an aspect of the present disclosure, a method performed by a first communication device in a wireless communication system is provided. The method may comprise: determining based on a neural network related measurement, whether to transmit a first message based on a measurement related to a neural network, the first message being used to for requesting a signal for a neural network calibration; and if the first message is transmitted, receiving at least two signals for the neural network calibration in a situation where the first message is sent, wherein the signals for the neural network calibration are transmitted through a given one antenna port, and a time domain resources and / or a frequency domain resources carrying the signals for the neural network calibration satisfy a first condition.

[0015] In an exemplary implementation, it may be determined that the first message is to be transmitted, if the neural network related measurement satisfies at least one of: a measurement result of the neural network being greater than a first threshold; the measurement result of the neural network being less than a second threshold; and a neural network-based channel estimation and / or subsequent data reception by the first communication device being failed.

[0016] In an exemplary implementation, the first message may comprise at least one of: request information for the signal for the neural network calibration; and information on a result of the neural network related measurement.

[0017] In an exemplary implementation, the at least two signals for the neural network calibration may comprise one configured reference signal, an antenna port for transmission of remaining signal(s) in the at least two signals for the neural network calibration may be determined based on an antenna port of the configured reference signal, and the time domain resources and / or the frequency domain resources of the remaining signal(s) and the configured reference signal satisfy the first condition.

[0018] In an exemplary implementation, the first condition may comprise at least one of: a difference value between a maximum value and a minimum value in indices of time domain resources of the at least two signals for the neural network calibration being less than a third threshold; a difference value between a maximum value and a minimum value in indices of frequency domain resources of the at least two signals for the neural network calibration being less than a fourth threshold; an index of a time domain resource of a first signal in the at least two signals for the neural network calibration being a first value, and an interval between a time domain resource of a second signal in the at least two signals for the neural network calibration and the time domain resource of the first signal being a predefined second value; and an index of a frequency domain resource of the first signal in the at least two signals for the neural network calibration being a predefined third value, and an interval between a frequency domain resource of the second signal in the at least two signals for the neural network calibration and the frequency domain resource of the first signal being a predefined fourth value.

[0019] In an exemplary implementation, the first communication device may further receive configuration information. The configuration information may comprise at least one of: mapping information between the signal for the neural network calibration and the neural network, information on time domain resources and / or frequency domain resources used by the signals for the neural network calibration, information for determining an antenna port for a transmission of the signal for the neural network calibration, and information on a sequence for generating the signal for the neural network calibration.

[0020] In an exemplary implementation, the request information for the signal for the neural network calibration may comprise: information on an antenna port number associated with the signal; an identifier (ID) of the neural network; and / or information on an antenna port number associated with a configured reference signal.

[0021] In an exemplary implementation, the first message comprises information on a first measurement result and indication information for requesting the signal for the neural network calibration. Alternatively, the first message comprises information on a second measurement result and the indication information for requesting the signal for the neural network calibration. Alternatively, the first message comprises the information on the first measurement result and the information on the second measurement result. Here, the first measurement result is obtained based on a reference signal for a channel estimation, and the second measurement result is a channel measurement result obtained based on the first measurement result and the neural network.

[0022] In an exemplary implementation, the first communication device may further receive configuration information. The configuration information may indicate relevant information of the second reference signal that has a multiplexing association with the first reference signal, and may include at least one of the following: information on whether the first reference signal is multiplexed, information on whether at least one second reference signal is transmitted or used, information on whether at least one second reference signal is mapped to the same time-domain and / or frequency-domain resource as the first reference signal, and / or information on the multiplexing manner of at least one second reference signal in the time-domain and / or frequency-domain resources.

[0023] In an implementation exemplary, the first communication device may further receive configuration information. The configuration information may indicate relevant information on the manner that the first reference signal uses the time-frequency resources, and may include at least one of the following: information on whether at least one first reference signal exclusively occupies the time-frequency resource, and / or information on the multiplexing manner of at least one first reference signal in the time-domain and / or frequency-domain resources.

[0024] In an exemplary implementation, the first reference signal may be, for example, DMRS or CSI-RS, but is not limited thereto. The second reference signal may be a reference signal of the same type as the first reference signal but using different antenna ports.

[0025] In an exemplary implementation, the first reference signal can be split into a first calibration signal and a second calibration signal. The neural network in the communication device may be updated through the first calibration signal and the second calibration signal.

[0026] In an implementation exemplary, the first communication device may further receive configuration information. The configuration information may include the structure and computer parameters of one or more neural networks. In an implementation exemplary, the first communication device may obtain the above configuration information based on preset information. In an implementation exemplary, the first communication device may select a neural network based on preset information and configuration information including indication information. In an implementation exemplary, the first communication device may report a measurement result related on the neural network’s performance, and determine the neural network to be used subsequently based on the received indication information.

[0027] According to another aspect of the present disclosure, a method performed by a second communication device in a wireless communication system is provided. The method may comprise: receiving a first message for requesting a signal for a neural network calibration, wherein the transmission of the first message is determined by a first communication device based on a neural network related measurement.

[0028] In an exemplary implementation, the transmission of the first message may be determined by the first communication device if the neural network related measurement satisfies at least one of: a measurement result of the neural network being greater than a first threshold; the measurement result of the neural network being less than a second threshold; and a neural network-based channel estimation and / or subsequent data reception by the first communication device being failed.

[0029] In an exemplary implementation, the first message may comprise at least one of: request information for the signal for the neural network calibration; and information on a result of the neural network related measurement.

[0030] In an exemplary implementation, the at least two signals for the neural network calibration may comprise one configured reference signal, an antenna port for a transmission of remaining signal(s) in the at least two signals for the neural network calibration may be determined based on an antenna port of the configured reference signal, and time domain resources and / or frequency domain resources of the remaining signal and the configured reference signal satisfy the first condition.

[0031] In an exemplary implementation, the first condition may comprise at least one of: a difference value between a maximum value and a minimum value in indices of time domain resources of the at least two signals for the neural network calibration being less than a third threshold; a difference value between a maximum value and a minimum value in indices of frequency domain resources of the at least two signals for the neural network calibration being less than a fourth threshold; an index of a time domain resource of a first signal in the at least two signals for the neural network calibration being a first value, and an interval between a time domain resource of a second signal in the at least two signals for the neural network calibration and the time domain resource of the first signal being a predefined second value; and an index of a frequency domain resource of the first signal in the at least two signals for the neural network calibration being a predefined third value, and an interval between a frequency domain resource of the second signal in the at least two signals for the neural network calibration and the frequency domain resource of the first signal being a predefined fourth value.

[0032] In an exemplary implementation, the method may further comprise: receiving configuration information. The configuration information may comprise at least one of: mapping information between the signal for the neural network calibration and the neural network, information on a time domain resource and / or frequency domain resource for the signal for the neural network calibration, information for determining an antenna port for a transmission of the signals for the neural network calibration, and information on a sequence for generating the signal for the neural network calibration.

[0033] In an exemplary implementation, the request information for the signal for the neural network calibration may comprise: information on an antenna port number associated with the signal; an identifier (ID) of the neural network; and / or information on an antenna port number associated with a configured reference signal.

[0034] In an exemplary implementation, the first message comprises information on a first measurement result and indication information for requesting the signal for the neural network calibration. Alternatively, the first message comprises information on a second measurement result and the indication information for requesting the signal for the neural network calibration. Alternatively, the first message comprises the information on the first measurement result and the information on the second measurement result. Here, the first measurement result is obtained based on a reference signal for a channel estimation, and the second measurement result is a channel measurement result obtained based on the first measurement result and the neural network.

[0035] According to another aspect of the present disclosure, a method performed by a communication device in a wireless communication system is provided. The method may comprise: performing a channel estimation based on at least two signals for a neural network calibration, to obtain a first estimation result and a second estimation result of a first channel estimation; obtaining a first estimation result of a second channel estimation through a neural network based on the first estimation result of the first channel estimation; determining a first error based on the second estimation result of the first channel estimation and the first estimation result of the second channel estimation; and updating the neural network based on the first error, wherein the signals for the neural network calibration are transmitted through one antenna port, and time domain resources and / or frequency domain resources carrying the signals for the neural network calibration satisfy a first condition.

[0036] In an exemplary implementation, the method may further comprise: obtaining a second estimation result of the second channel estimation through the neural network based on the second estimation result of the first channel estimation; determining a second error based on the first estimation result of the first channel estimation and the second estimation result of the second channel estimation; and updating the neural network based on the first error and the second error.

[0037] In an exemplary implementation, the updating the neural network based on the first error and the second error may comprise: updating the neural network based on the first error, and then re-updating the updated neural network based on the second error; updating the neural network based on the second error, and then re-updating the updated neural network based on the first error; or determining a total error based on the first error and the second error, and updating the neural network based on the total error.

[0038] In an exemplary implementation, the signals for the neural network calibration may comprise one configured reference signal, an antenna port for a transmission of remaining signal(s) in the at least two signals may be determined based on an antenna port of the configured reference signal, and the time domain resources and / or the frequency domain resources of the remaining signal(s) and the configured reference signal may satisfy the first condition.

[0039] In an exemplary implementation, the first condition may comprise at least one of: a difference value between a maximum value and a minimum value in indices of time domain resources of the at least two signals for the neural network calibration being less than a third threshold; a difference value between a maximum value and a minimum value in indices of frequency domain resources of the at least two signals for the neural network calibration being less than a fourth threshold; an index of a time domain resource of a first signal in the at least two signals for the neural network calibration being a first value, and an interval between a time domain resource of a second signal in the at least two signals for the neural network calibration and the time domain resource of the first signal being a predefined second value; and an index of a frequency domain resource of the first signal in the at least two signals for the neural network calibration being a predefined third value, and an interval between a frequency domain resource of the second signal in the at least two signals for the neural network calibration and the frequency domain resource of the first signal being a predefined fourth value.

[0040] In an exemplary implementation, the first communication device may further receive configuration information. Here, the configuration information may comprise at least one of: mapping information between a signal for the neural network calibration and the neural network, information on time domain resources and / or frequency domain resources for the signals for the neural network calibration, information for determining an antenna port for a transmission of the signal(s) for the neural network calibration, and information on a sequence for generating the signal for the neural network calibration.

[0041] In an exemplary implementation, the method may further comprise: obtaining, based on a first received signal comprising at least one reference signal, first channel estimation information on a resource element (RE) occupied by the at least one reference signal; obtaining second channel estimation information corresponding to the first channel estimation information through the updated neural network based on the first channel estimation information; and determining channel state information on all REs occupied by the first received signal based on the second channel estimation information.

[0042] In an exemplary implementation, the obtaining second channel estimation information corresponding to the first channel estimation information through the updated neural network based on the first channel estimation information may comprise: obtaining the second channel estimation information through the updated neural network based on the first channel estimation information and auxiliary information, wherein the auxiliary information comprises at least one of information on the RE occupied by the at least one reference signal and information on a channel environment.

[0043] According to another aspect of the present disclosure, a communication device in a wireless communication system is provided. The communication device may comprise: a transceiver, configured to transmit and receive a signal; and a processor, coupled to the transceiver, and configured to perform any of the above methods.

[0044] According to another aspect of the present disclosure, a computer readable storage medium is provided. The computer readable storage medium stores a computer executable instruction. When the computer executable instruction is executed by a processor, the processor may perform the above methods.

[0045] According to the present disclosure, a communication device determines, based on neural network related measurements, whether to transmit a calibration request message, and if the message is transmitted, receives calibration signals through a specific antenna port and over time and / or frequency resources satisfying predetermined conditions. As a result, wireless resources can be used more efficiently, the accuracy and reliability of neural network calibration can be improved, and overall system performance and quality of service experienced by end users can be enhanced.

[0046] FIG. 1 illustrates an example wireless network according to an embodiment of the present disclosure;

[0047] FIG. 2 illustrates an example base station according to an embodiment of the present disclosure;

[0048] FIG. 3 illustrates an example user equipment according to an embodiment of the present disclosure;

[0049] FIG. 4 is a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0050] FIG. 5 is a schematic diagram of a neural network according to an exemplary embodiment of the present disclosure;

[0051] FIG. 6 illustrates a process of obtaining channel state information based on a received reference signal according to an exemplary embodiment of the present disclosure;

[0052] FIG. 7 is a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0053] FIG. 8 is a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0054] FIGs. 9a, 9b and 9c illustrate calibration signals according to an exemplary embodiment of the present disclosure;

[0055] FIG. 10 illustrates a training process according to an exemplary embodiment of the present disclosure;

[0056] FIG. 11 illustrates a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0057] FIG. 12 illustrates a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0058] FIG. 13 is a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0059] FIG. 14 is a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure;

[0060] FIG. 15 illustrates a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure; and

[0061] FIG. 16 illustrates an exemplary structure of a communication device applicable to the present disclosure.

[0062] Before undertaking the DETAILED DESCRIPTION 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. Accordingly, a set of items can be a single item or a collection of two or more items.

[0063] 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.

[0064] 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.

[0065] 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.

[0066] 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.

[0067] 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.

[0068] 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.

[0069] 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 includes the UE 115 and the UE 116, as well as subscriber stations (SS, for example, UEs) 117, 118 and 119. In some embodiments, one or more of the gNBs 101-103 may communicate with each other and with the UEs 111-116 using existing wireless communication techniques, and one or more of the UE 111-119 may communicate directly with each other (e.g., UEs 117-119) using other existing or proposed wireless communication techniques.

[0070] 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).

[0071] 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.

[0072] 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.

[0073] 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.

[0074] 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.

[0075] 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.

[0076] 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 transmitted 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.

[0077] 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.

[0078] 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.

[0079] 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.

[0080] 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.

[0081] 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.

[0082] 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).

[0083] 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). Also, various components in FIG. 2 could be combined, further subdivided, or omitted and additional components could be added according to particular needs.

[0084] 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.

[0085] 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.

[0086] 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 transmitted 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).

[0087] 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.

[0088] 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.

[0089] 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.

[0090] 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.

[0091] 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.

[0092] 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.

[0093] Exemplary embodiments of the present disclosure are further described below in combination with the accompanying drawings.

[0094] The text and drawings are provided as examples only to help readers understand the present disclosure. The text and drawings are not intended to limit and should not be interpreted as limiting the scope of the present disclosure in any way. Although certain embodiments and examples have been provided, based on the content disclosed herein, it is obvious to those skilled in the art that modifications can be made to the illustrated embodiments and examples without departing from the scope of the present disclosure.

[0095] In a wireless communication system, the transmission of information (e.g., the transmission of information such as a physical downlink control channel (PDCCH), a physical downlink shared channel (PDSCH), a physical uplink shared channel (PUSCH), a physical uplink control channel (PUCCH)) may occur on at least one resource element (RE) of a physical resource, e.g., a plurality of time points, a plurality of frequency points, a plurality of antennas, and various combinations of time, frequency and antenna. The physical resource is a resource entity for a signal transmission in a communication system. For example, the physical resource may be a time domain physical resource, a frequency domain physical resource, an antenna domain physical resource, and the like. The state of the RE occupied by a signal will have a combined effect on the amplitude and phase of the signal transmitted through an entire wireless communication link. Generally, the combined effect is referred to as the channel state information on the corresponding RE in the present disclosure.

[0096] In order to meet the requirements of wireless data transmission, it is required to learn the channel state information of the RE occupied by the signal at a receiving end and / or a transmitting end. A general method of acquiring the channel state information of a certain RE is to transmit a reference signal (RS) on the RE and estimate the channel state information accordingly. For example, the reference signal is a transmitting signal composed of a generation sequence, and the content and the time point (time domain resource) and the frequency point (frequency domain resource) for transmitting of the reference signal are shared by both the receiving end and the transmitting end. The reference signal may alternatively be referred to as a pilot signal, a training signal, or the like.

[0097] Generally, the transmission of information occur on more than one RE. A method of acquiring the channel state information on all REs used for the transmission is to select some REs from the REs according to a certain rule to transmit a reference signal (RS), and estimate the channel state information on all REs accordingly. Specifically, according to the reference signal configuration information known to both the transmitting end and the receiving end, the transmitting end transmits the reference signal determined by the configuration information on the RE determined by the configuration information. The receiving end may obtain the channel state information from the corresponding RE according to the configuration information, and further estimate the channel state information on all the REs.

[0098] In an actual environment, due to the presence of noise and interference, there is an error between the channel state information estimation obtained through the method and the real channel influence, resulting in a reduction in the accuracy of data transmission. In order to meet increasing transmission rate requirements and improve the rate experience of terminals at a cell edge, a noise suppression algorithm may be added in the channel estimation process to reduce the error in the channel state information. However, the noise suppression algorithm has a high complexity of computation, and thus it is difficult to meet the processing delay requirements of the current 5G communication system.

[0099] For the above problems, the present disclosure proposes a neural network-based channel estimation process. A neural network is added in a channel estimation process to implement noise suppression, thereby effectively solving the above problems faced by a traditional channel estimation process.

[0100] In addition, when the system deploys the neural network, the neural network is mostly based on offline training, that is, the neural network in the system is acquired through the offline training before being deployed by the system. After the actual deployment of the system, the parameters of the neural network will not change. However, the actual environment where the communication device is will change over time, which causes the performance of the neural network to gradually decrease, thereby affecting the performance of the channel estimation process. In this regard, it is required to periodically and / or aperiodically update the parameters in the neural network in the communication device to ensure the noise reduction performance of the neural network. However, the parameter update to the neural network needs to rely on real data lables (e.g., real channel state information), which are difficult to acquire.

[0101] Further, the present disclosure further provides a calibration signal for updating a neural network and a neural network on-line update process based on the calibration signal, thereby ensuring the performance stability of the neural network in the channel estimation process.

[0102] FIG. 4 is a flowchart of a method performed by a communication device according to an embodiment of the present disclosure.

[0103] According to the embodiment, the communication device may be any device capable of receiving and transmitting a signal and performing a signal detection. For example, the communication device may be a base station or a terminal, but is not limited thereto. Alternatively, for example, the communication device may be a user equipment having a large scale of antenna ports, etc.

[0104] As shown in FIG. 4, in step S410, first information is obtained based on a received signal. Here, the received signal contains at least one reference signal, and the first information is a channel estimation result obtained based on the reference signal contained in the received signal. For example, the first information includes a channel estimation result corresponding to a channel on an RE occupied by the reference signal. For example, the channel estimation result contains the channel state information of a channel through which the reference signal passes, and an error caused by a factor such as noise and / or an interference due to the influence of an environment, a communication device and / or a signal processing process. In step S420, based on the first information, second information including the channel estimation result corresponding to the channel on the RE occupied by the reference signal is obtained through a corresponding first neural network. Here, the second information is closer, than the first information, to a real channel influence. In step S430, channel state information corresponding to channels on all REs occupied by the received signal is obtained based on the second information.

[0105] According to the above method, since the signal processing is performed through the neural network, the communication device can obtain the channel state information closer to the real channel influence, thereby improving the overall performance of the communication system. In addition, the above method is more robust to non-linear factors such as a timing offset and a frequency offset, and thus no significant performance attenuation will be caused by the non-linear factors.

[0106] The contents involved in the above steps S410-S430 will be respectively described below in detail.

[0107] According to an exemplary embodiment, the present disclosure relates to a neural network-based channel estimation method.

[0108] Referring to FIG. 4, in step S410, the first information is obtained based on the received signal. Here, the received signal contains at least one reference signal. The first information is the channel estimation result obtained based on the reference signal contained in the received signal, and includes the channel estimation result corresponding to the channel on the RE occupied by the reference signal. This channel estimation result may contain the channel state information of the channel through which the reference signal passes, and the error caused by the factor such as the noise and / or the interference due to the influence of the environment and / or the communication device and / or the signal processing process. According to the embodiment, the step may contain: extracting the reference signal from the received signal; and obtaining the first information based on the reference signal.

[0109] In some examples, in the process of obtaining the first information based on the reference signal, the communication device may obtain the first information according to a preset reference signal sequence and the reference signal. Specifically, for example, an algorithm that may be adopted may include, but not limited to, at least one of: 1) a least square method (LS); 2) a minimum mean square error method (MMSE); and 3) a linear minimum mean square error method (LMMSE).

[0110] Here, an exemplary embodiment is given. For example, a received signal y carrying a PDSCH channel contains a DMRS signal corresponding to the PDSCH channel. First, according to the position of an RE occupied by the DMRS signal, the communication device combines the signal on the corresponding RE into a reference signal y_rs. Then, the communication device obtains a channel estimation h_1 from the reference signal y_rs by using various signal estimation algorithms such as a least square algorithm, and uses the channel estimation h_1 as the first information. Specifically, the communication device may use the quotient of the reference signal y_rs and a pre-configured reference signal sequence x_rs ,i.e., h_1= y_rs / x_rs, as the first information. In addition, it should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0111] In some examples, in the process of obtaining the first information based on the reference signal, some post-processing (e.g., an addition, an amplification, a noise reduction, filtering, whitening, and a fast Fourier transform) may alterantively be performed on the channel estimation result corresponding to the channel on the RE occupied by the reference signal, and the result obtained through the post-processing may be used as the first information.

[0112] In the present disclosure, “at least one” may represent one or a combination of two or more. For example, in a 5G system, the received signal contains reference signals of two different antenna ports. After the process in the foregoing embodiment, the first information contains the channel state information of the two different antenna ports. In this regard, it is required to separate the channel state information of the different antenna ports to respectively use the channel state information as the first information corresponding to the two different antenna ports. For example, it is possible to perform an addition and a subtraction on the results on two adjacent frequency domain resources, to obtain the information corresponding to the reference signal of each antenna port (i.e., the information corresponding to each antenna port) as the first information corresponding to the antenna port.

[0113] In some examples, the received signal may be a complete received signal containing the reference signal, or may be a partial signal containing the reference signal and extracted from the received signal. The received signal may include at least one of: a radio frequency signal received by an antenna port of the communication device; a digital domain signal obtained by performing a digital-to-analog conversion on the radio frequency signal; a signal obtained by performing specific processing (e.g., an amplification, a noise reduction, filtering, whitening, and a fast Fourier transform) on the digital domain signal.

[0114] In some examples, the received signal is a digital domain wireless signal received by the antenna port of the communication device, or a signal obtained by performing a noise reduction on the digital domain wireless signal through a band-pass filter.

[0115] In some examples, the received signal is a signal obtained after a signal received by the antenna port of the communication device passes through a spatial filter (e.g., an eigenvector matrix of a channel or a spatial discrete cosine transform (discrete Fourier transform (DFT)) matrix).

[0116] In some examples, the received signal is a signal obtained by performing whitening on the signal received by the antenna port of the communication device. For example, a possible whitening method is as follows: estimating an interference covariance matrix in a current environment based on the reference signal in the signal received by the antenna port of the communication device; performing a Cholesky decomposition or principal component analysis (PCA) decomposition on the interference covariance matrix; and multiplying the inverse matrix of the matrix after the decomposition by the signal received by the antenna port of the communication device, to obtain a received reference signal.

[0117] Next, in step S420, based on the first information, the second information is obtained through the corresponding first neural network. Here, the second information is the channel state information on the RE occupied by the reference signal, and is closer, than the first information, to the real channel effect.

[0118] In some examples, there may be one or more neural networks in the communication device. Based on one or more antenna ports contained in the first information, the corresponding first neural network is obtained for processing the first information. Specifically, based on the reference signal of each antenna port in the first information, the first information of the corresponding antenna port can be obtained in step S410. The antenna ports contained in the first information may be divided into one or more antenna port groups, each of which contains at least one antenna port. The first information corresponding to the antenna port in each antenna port group needs to be processed by using the first neural network corresponding to the group. Different antenna port groups may correspond to the same first neural network or different first neural networks.

[0119] In some embodiments, the first information corresponding to each antenna port in one antenna port group is independently processed using the corresponding first neural network. For example, the first information contains the information corresponding to three antenna ports (an antenna port 0, an antenna port 1 and an antenna port 2). The antenna port 0 and the antenna port 1 are divided into a group 0, and the antenna port 2 is divided into a group 1. The information corresponding to the antenna port 0 and the information corresponding to the antenna port 1 are respectively inputted to a first neural network (numbered 0 and corresponding to the group 0) for processing. The information corresponding to the antenna port 2 is inputted to a first neural network (numbered 1 and corresponding to the group 1) for processing. The combination of the outputs of all the neural networks constitutes the second information.

[0120] In some embodiments, the information corresponding to each antenna port in one antenna port group is combined to be processed using the corresponding first neural network, and then the processed information is decomposed into the information of each antenna port. For example, the antenna port 0 and the antenna port 1 contained in the first information are divided into the group 0. The information corresponding to the antenna port 0 and the information corresponding to the antenna port 1 are multiplied by a specific coefficient, and then combined into new information through an addition, and then the combined information is inputted to the first neural network for processing. The processed information may be separated by performing an addition and a subtraction on the results on two adjacent frequency domain resources, thus obtaining the information corresponding to each of the antenna port 0 and the antenna port 1.

[0121] According to an embodiment, the second information may be obtained based on the first information and auxiliary information.

[0122] In some examples, the auxiliary information may include the information on the RE occupied by the reference signal. For example, the auxiliary information may include at least one of: information on a corresponding time domain resource, information on a corresponding frequency domain resource, information on a corresponding antenna resource, information on a channel environment, position information (including an absolute position and / or a relative position) of the RE, structure information, and distribution information (as an example, an OFDM symbol position where the reference signal is; and as another example, a frequency domain interval between REs where the reference signal is).

[0123] In some examples, the auxiliary information may include the information on the channel environment. As an example, the information on the channel environment may include at least one of: a signal-to-noise ratio, multi-path delay spread information of a channel, and relevant time information of the channel. As an example, the signal-to-noise ratio may include SNR or SINR. The multi-path delay spread information of the channel may refer to a multi-path distribution rule of the channel, and the distribution may be preset by the installation personnel of the communication device, or may be obtained by the communication device by analyzing the long-time operation data of the communication device. As another example, the relevant time information of the channel may refer to a relationship between the correlation of the channels on the REs carrying the reference signal and the distances of the channels in time, for example, refer to a temporal correlation coefficient, and this parameter may be calculated after the communication device obtains a speed according to a rate sensor.

[0124] Alternatively, if at least one condition is satisfied, based on the first information, the second information is obtained using a noise suppression algorithm. The noise suppression algorithm used may be, for example, one or more of a sliding window average algorithm, a delay domain truncation algorithm, a filtering denoising algorithm, an SVD denoising algorithm, or the like.

[0125] Alternatively, the at least one condition may include, for example, at least one of: a failure in data reception; a failure of a subsequent signal decoding process; a signal-to-noise ratio of the signal after a subsequent equalization process not exceeding a first threshold; a noise intensity and / or signal-to-noise ratio in the second information exceeding a second threshold; a reception of indication signaling for no use of the neural network; and there being an indication that the neural network cannot work properly.

[0126] FIG. 5 is a schematic diagram of a neural network according to an embodiment of the present disclosure.

[0127] The neural network may be implemented jointly and / or independently by the same physical entity or different entity units. A physical entity (or entity unit) may be composed of hardware, software, or a combination thereof, and can be configured by a person skilled in the art according to actual needs.

[0128] As shown in FIG. 5, the neural network may include an information processing unit and a plurality of neural network layers. The information processing unit may convert input information and / or auxiliary information into at least one of a data vector, a matrix and a tensor that can be used for the calculation of the neural network layers. The neural network layers may include at least one of an input layer, a hidden layer / intermediate layer (intermediate hidden layer) and an output layer shown in FIG. 5. In addition, each neural network layer may consist of a plurality of neurons n, and the neurons may be activated using at least one activation function. As an example only rather than a limitation, the neurons may be activated using, for example, a tanh function, an ReLU function, an eLU function, an seLU function, a ceLU function, a preLU function, a geLU function, a LeakyReLU function, a Sigmoid function, a Softmax function, and a Softplus function. Each neural network layer may perform an operation on the input data thereof to implement, for example, a function such as a matrix transformation, a data dimension reduction, a data feature extraction, and a data feature combination. As an example only rather than a limitation, the plurality of neural network layers may constitute different neural network structures in series or in parallel, including, but not limited to, a multilayer perceptron (MLP), a multilayer perceptron mixer (MLP-mixer), a convolutional neural network (CNN), a deep neural network (DNN), a recursive neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recursive deep neural network (BRDNN), a generative adversarial network (GAN), a transform network (Transformer), and the like.

[0129] For example, in the situation where the neural network shown in FIG. 5 is the neural network mentioned above, the input information may refer to channel state information, and the information processing unit may convert the channel state information and the auxiliary information into at least one of the vector, the matrix and the tensor used for the calculation of the neural network layers.

[0130] In some examples, the neural network may adopt a deep convolutional neural network. The channel state information corresponding to a plurality of samples can be stacked and arranged in a tensor form according to the positions of physical resources of the channel state information, to be inputted into the neural network. The neural network may output a piece of matrix data of a shape corresponding to transmitting and receiving antennas as feature information. In other embodiments of this example, the neural network may adopt a Transformer structure, in which the channel state information on a plurality of REs is superimposed with the position Tokens corresponding to the positions of the REs and then inputted into the neural network together with a pre-trained query Token, and the neural network outputs the data of a channel where the query Token is as the feature information.

[0131] As an example rather than a limitation, a set consisting of channel state information is defined as . First, the information processing unit processes the known channel state information through a numerical mapping method (e.g., normalization, and standardization), to make the numerical scale or distribution of the channel state information suitable for the calculation of the neural network. More specifically, taking the Minimum-Maximum (Min-Max) standardization as an example, it is possible to first calculate the maximum value and the minimum value in all elements in , respectively, and then process each element in as follows:

[0132]

[0133] After the above numerical mapping is performed, the set consisting of the processed element may be converted into at least one of the vector, the matrix and the tensor corresponding to the input size of the neural network layers.

[0134] In some examples, the information processing unit may alternatively perform position encoding on the input information, and convert the position-encoded information into at least one of the vector, the matrix and the tensor used for the calculation of the neural network. Alternatively, the position encoding may refer to an addition and / or a multiplication on the input information and the codewords in a predetermined codebook. As an example only rather than a limitation, in an example in which the input information refers to the channel state information, the position encoding performed using a codebook may include the following processing:

[0135] First, a set consisting of known channel state information is processed according to a numerical mapping method, to obtain a set , which contains several numerically mapped elements . Then, a position-encoded codebook is obtained according to the following equation.

[0136]

[0137] Here, n denotes the position of a physical resource (e.g., a time domain, a frequency domain and a space domain) where the channel state information is, d denotes a physical resource dimension of the channel state information, N is any positive number greater than the number of the elements contained in the set , and i denotes an index of the physical resource dimension. Then, the position encoding is performed on the known channel state information element through an equation to obtain a set . Finally, the set consisting of the processed element is converted into at least one of the vector, the matrix and the tensor corresponding to the input size of the neural network layers, to be inputted to the neural network layers.

[0138] As shown in FIG. 5, the information processing unit may perform processing (e.g., the above mapping and the above position encoding) on the inputted first information and the auxiliary information, such that the neural network may acquire priori information implied therein, to assist the neural network in improving the precision and flexibility of acquiring a prediction result. The way in which the feature information is obtained using the neural network can make the obtained feature information more accurate. Alternatively, the information processing unit may alternatively input its input directly to the neural network layer without any additional processing. In addition, it should be noted that the operation in the neural network unit is merely exemplary, and the present disclosure is not limited thereto.

[0139] After the second information is acquired, in step S430, the channel state information is obtained based on the second information. Here, the channel state information refers to the channel state information corresponding to the channels on all REs occupied by the received signal. According to an embodiment, the step may include: using an RE occupied by a reference signal as a reference RE and determining the channel state information corresponding to the channels on all the REs occupied by the received signal based on second information corresponding to a channel on the reference RE. For example, it is possible to select at least one reference RE for each RE occupied by the received signal, where the reference RE should be an RE carrying the reference signal; obtain the channel state information corresponding to a channel on the selected reference RE from the second information; and determine the channel state information corresponding to a channel on each RE occupied by the received signal, based on the channel state information corresponding to the channel on the reference RE.

[0140] Here, an exemplary embodiment is given. The REs occupied by the currently received signal are #0-#11, where the reference signal is on the REs #2 and #7. A terminal first selects the REs #2 and #7 as reference REs. Then, the terminal obtains the second information and corresponding to the REs. Finally, the terminal calculates the channel state information of an RE according to the serial number of the RE. Specifically, the channel state information on the RE numbered x is . In addition, it should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0141] In some examples, it is possible to first obtain the channel state information of all REs at the same time / frequency domain position based on the second information of the reference RE, and then use the obtained channel state information as new second information, thereby further obtaining the channel state information on other REs.

[0142] In some examples, the channel state information on all the REs may be obtained directly through the second information of the reference RE. For example, in an NR system, an interpolation matrix W can be obtained through an MMSE algorithm by means of a correlation matrix between the REs. Based on the interpolation matrix W, the channel state information may be represented by , where is a vector representation corresponding to the second information.

[0143] In some examples, it is possible to implement the above interpolation process using a first-order interpolation algorithm in a frequency domain. The specific process includes the following steps. First, for an RE having a subcarrier numbered x, two REs having the subcarriers adjacent to the subcarrier and numbered x1and x2are found from the REs carrying the reference signal. Then, the channel state information and on the two REs is obtained. Finally, the channel state information on a target RE is calculated.

[0144] In some examples, it is possible to respectively implement the above interpolation process using the first-order interpolation algorithm in a frequency domain and a time domain. The specific process includes the following steps. First, for an RE positioned on a y-th OFDM symbol and having a subcarrier numbered x, the REs positioned on y1-th y2-th OFDM symbols adjacent to the y-th OFDM symbol and having subcarriers numbered x are found from the REs carrying the reference signal. Then, the channel state information and on the two REs is obtained through the above interpolation process in the frequency domain. Finally, the channel state information on a target RE is calculated, where and are preset coefficients.

[0145] In FIG. 6 an exemplary embodiment is given to better illustrate the process of obtaining channel state information based on a received reference signal. As shown in FIG. 6, first, in block 601, first information is obtained through an LS estimation based on a demodulation reference signal (DMRS), and noise suppression filtering is performed on the first information to eliminate irrelevant interference information. Then, in block 602, the first information is split into two groups based on odd and even serial numbers in a frequency domain dimension, to be respectively inputted into a lightweight U-net neural network composed of a plurality of convolutional layers, and the two groups of information outputted are reorganized according to their serial numbers, and then smoothed in a frequency domain to be used as an output of the neural network. Finally, the output of the neural network is used as second information, and in blocks 603 and 604, the subsequent interpolation process in frequency and time domains are performed. It should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0146] According to an exemplary embodiment, the present disclosure further relates to an update to the neural network. For example, the neural network may be updated using a calibration signal.

[0147] According to an exemplary embodiment, the present disclosure further relates to the transmitting and / or reception of the calibration signal.

[0148] FIG. 7 is a flowchart of a method performed by a communication device according to an embodiment of the present disclosure.

[0149] According to the embodiment, the communication device may be any device capable of acquiring a signal and performing a signal detection. For example, the communication device may be a base station or a terminal, but is not limited thereto. Alternatively, for example, the communication device may even be a user equipment having a large scale of antenna ports, etc.

[0150] As shown in FIG. 7, in step S710, it is determined whether to transmit a first message for a calibration signal for a neural network calibration, based on a neural network related measurement. In step S720, if the first message is transmitted, at least two calibration signals are received. Here, the received at least two calibration signals are transmitted through the same antenna port, and a time domain resource and / or a frequency domain resource carrying the at least two calibration signals satisfy a first condition. Each calibration signal may be associated with one reference signal and / or one neural network. The association information (or mapping information) between the calibration signal and the reference signal and / or the neural network, the antenna port for a transmission of the calibration signal(s), the time domain resources and / or the frequency domain resources may be from the preset information of the communication device, or may be from the configuration information received by the communication device from the outside.

[0151] The contents involved in the above steps S710-S720 will be respectively described below in detail.

[0152] Referring to FIG. 7, in step S710, it is determined whether to transmit the first message for the calibration signal for the neural network calibration, based on the neural network related measurement. According to an embodiment, this step may specifically contain: obtaining a result of a neural network performance measurement; and determining whether to transmit indication information for a requirement for the calibration signal based on the above result.

[0153] In some examples, the result of the neural network performance measurement may refer to the measurement value or measurement result obtained based on the steps described below with reference to FIG. 11.

[0154] In some examples, the result of the neural network performance measurements may be obtained through the success or failure of the communication device in the reception of data. As an example, the success or failure of the data reception is used as a result. As another example, the bit error rate (BER), the block error rate (BLER), or the like in the reception of data is used as a result.

[0155] In some examples, the communication device decides whether to transmit the indication information for the requirement for the calibration signal based on one or more results of the neural network performance measurement. In some embodiments, the communication device directly decides whether to transmit the indication information for the requirement for the calibration signal based on the results of the neural network performance measurement. For example, if the reception for a neural network-based channel estimation and / or subsequent data fails, the indication information for the requirement for the calibration signal should be transmitted. In some embodiments, the communication device decides whether to transmit the indication information for the requirement for the calibration signal based on a comparison between the result of the neural network performance measurement and a threshold. As an example, if the measurement result (e.g., the BER or BLER) in the reception of data is higher than a preset threshold, the indication information for the requirement for the calibration signal should be transmitted. As another example, if the measurement result (e.g., a noise reduction gain) of the neural network is lower than a threshold, the indication information for the requirement for the calibration signal should be transmitted. In some embodiments, the communication device decides whether to transmit the indication information for the requirement for the calibration signal based on the plurality of results of the neural network performance measurement. For example, if the data reception for the neural network-based channel estimation result fails and the noise reduction gain of the neural network remains lower than the threshold, the indication information for the requirement for the calibration signal should be transmitted. In some examples, the indication information for the requirement for the calibration signal is explicitly transmitted. As an example, the request information for the calibration signal is directly transmitted on the corresponding physical resource. As another example, the corresponding indication information is added to the relevant measurement report. For example, the indication information for requesting the signal for the neural network calibration may be added to the measurement result corresponding to the input of the neural network, or the indication information for requesting the signal for the neural network calibration may be added to the measurement result corresponding to the output of the neural network.

[0156] In some examples, the indication information for the requirement for the calibration signal is implicitly transmitted. As an example, the information on the measurement result of the neural network may be transmitted. As another example, in a plurality of reported CQIs (channel quality indications), the same CQI and / or the inverse CQI (i.e., the CQI corresponding to the input of the neural network should be higher than and / or equal to the CQI corresponding to the output of the neural network) are used. For example, the CQI corresponding to the input of the neural network and the CQI corresponding to the output of the neural network may be respectively reported. If the CQI corresponding to the output of the neural network is greater than the CQI corresponding to the input of the neural network, it indicates that the neural network improves the channel estimation quality. If the CQI corresponding to the output of the neural network is equal to or less than the CQI corresponding to the input of the neural network, it indicates that the neural network fails to improve the channel estimation quality and even reduces the channel estimation quality. In this situation, the neural network needs to be calibrated. It should be understood that the transmitted / reported measurement result of the neural network is not limited to the CQI, and may alternatively be other measurement results related to the channel quality.

[0157] In some examples, the above indication information may further include an antenna port used by the calibration signal. In some embodiments, the information (e.g., the antenna port number) of the calibration signal may be directly added to the indication information. In some embodiments, the information of the neural network (e.g., the ID of the neural network) associated with the calibration signal is added to the indication information. In some embodiments, the information (e.g., the antenna port number) of the reference signal associated with the calibration signal is added to the indication information.

[0158] Alternatively, the communication device may not need to decide whether to transmit the indication information for the requirement for the calibration signal based on the one or more results of the neural network performance measurement. In some examples, if reporting a measurement result, the communication device may add the indication information for indicating whether the calibration signal is required. For example, if the indication information for indicating whether the calibration signal is required is a first value, it indicates that the calibration signal is required. If the indication information for indicating whether the calibration signal is required is a second value, it indicates that the calibration signal is not required. For example, the indication information for requesting the signal for the neural network calibration may be added to the measurement result corresponding to the input of the neural network, or the indication information for requesting the signal for the neural network calibration may be added to the measurement result corresponding to the output of the neural network. In some examples, the communication device may respectively report the measurement (e.g., CQI) corresponding to the input of the neural network and the measurement corresponding to the output of the neural network. If the CQI corresponding to the output of the neural network is greater than the CQI corresponding to the input of the neural network, it indicates that the neural network improves the channel estimation quality. If the CQI corresponding to the output of the neural network is equal to or less than the CQI corresponding to the input of the neural network, it indicates that the neural network fails to improve the channel estimation quality and even reduces the channel estimation quality. In this situation, the neural network needs to be calibrated. It should be understood that the transmitted / reported measurement result of the neural network is not limited to the CQI, and may alternatively be other measurement results related to the channel quality.

[0159] In step S720, if the first message is transmitted, the at least two calibration signals are received. Here, the received at least two calibration signals are transmitted through the same antenna port, and the time domain resource and / or the frequency domain resource carrying the at least two calibration signals satisfy the first condition. As described above, each calibration signal may be associated with one reference signal and / or one neural network.

[0160] Alternatively, the at least two signals for the neural network calibration include one configured reference signal, the antenna port for a transmission of remaining signal(s) is determined based on the antenna port of the configured reference signal, and the time domain resources and / or the frequency domain resources of the remaining signal and the configured reference signal satisfy the first condition.

[0161] For example, the first condition includes at least one of: a difference value between a maximum value and a minimum value of indices of the time domain resources and / or the frequency domain resources being less than a third threshold; or there being an interval as a second value between the time domain resources and / or the frequency domain resources with a first value as a reference. The first condition may be determined based on the configuration information of the communication device.

[0162] Alternatively, before the communication device receives the calibration signals, the method further includes transmitting / receiving configuration information on the calibration signals. Here, the configuration information contains the information on the reference signal and / or the neural network associated with the calibration signals. The content related to the configuration information will be described in detail below.

[0163] Through the above method, the neural network can be updated online according to the real channel state. In the situation where the performance of the neural network gradually degrades due to the change of the environment of the communication device over time, the on-line update to the neural network can ensure the noise reduction performance of the neural network, resulting in a stable channel estimation result.

[0164] FIG. 8 is a flowchart of a method performed by a communication device according to an embodiment of the present disclosure.

[0165] According to the embodiment, the communication device may be any device capable of acquiring a signal and performing a signal detection. For example, the communication device may be a base station or a terminal, but is not limited thereto. Alternatively, for example, the communication device may even be a user equipment having a large scale of antenna ports, etc.

[0166] As shown in FIG. 8, in step S810, based on a received calibration signal, a first value and a second value of first information on the calibration signal are acquired. In step S820, a first value of second information corresponding to the first value of the first information and / or a second value of the second information corresponding to the second value of the first information are obtained through a first neural network. In step S830, a first error is determined based on the second value of the first information and the first value of the second information, and / or a second error is determined based on the first value of the first information and the second value of the second information. In step S840, the first neural network is updated based on a first error and / or a second error.

[0167] According to an embodiment, the calibration signal may be contained in a received signal, for example, the received signal may be of the same type as the received signal described in step S410.

[0168] Considering that a plurality of calibration signals use the same antenna port and are close in time-frequency position, the calibration signals have approximately consistent corresponding channel state information but are different in noise process. Therefore, by using the channel state information of the calibration signals as training data of a network, it is possible to force the network to converge to a noiseless expected output.

[0169] According to the above method, based on the received signal only, the communication device can implement the online update to the neural network inside the communication device without performing a channel state information transmission process. This scheme not only avoids the significant increase in the system complexity that is caused by the online update to the neural network, but also achieves the performance robustness in changing channel environments.

[0170] The contents involved in the above steps S810-S840 will be respectively described below in detail.

[0171] Referring to FIG. 8, in step S810, based on the received calibration signal, the first value and the second value of the first information on the calibration signal are acquired. Specifically, this step may include: extracting at least one calibration signal from a received signal, and dividing the calibration signal into one or more calibration signal groups based on an antenna port used by the calibration signal; and obtaining a first value and a second value of first information on each calibration signal based on the calibration signal groups.

[0172] According to an embodiment, the calibration signal is a reference signal, and positioned on the specific physical resource described below. The antenna port used by the calibration signal is determined based on the reference signal and / or the neural network associated with the calibration signal. The reference signal and / or the neural network refer to the reference signal and / or the neural network used in step S410 and / or step S420. The communication device performs a comparison according to a transmitting sequence pre-configured by the communication device and the received calibration signal, thus obtaining corresponding channel state information. According to an embodiment, in step S810, the communication device may respectively perform the process described in step S410 on each calibration signal contained in the calibration signal groups, thus acquiring a corresponding channel estimation result. In the present disclosure, "at least one" may represent one, or a combination of two or more.

[0173] According to an embodiment, a calibration signal group may contain two calibration signals using the same antenna port, that is, the two calibration signals are associated with the same reference signal / or neural network. The antenna port used by the calibration signals is determined by the associated reference signal / or neural network. The calibration signals in each calibration signal group may be the same, partially the same, and / or different.

[0174] In some examples, the reference signal with which the calibration signal is associated may also be used as a calibration signal. That is, the two calibration signals contained in the calibration signal group may refer to an existing reference signal in the system and / or a newly added reference signal associated with the reference signal. The newly added reference signal may be a retransmission of the existing reference signal in the above system. The reference signal in the present disclosure includes, but not limited to, at least one of: a demodulation reference signal (DMRS), a channel state information reference signal (CSI-RS), a phase tracking reference signal (PT-RS) and a sounding reference signal (SRS).

[0175] In some embodiments, a DMRS signal and an additional DMRS signal in a PDSCH channel are used as two calibration signals to constitute one calibration signal group. In this situation, the first value and the second value of the first information obtained in step S810 may refer to a channel estimation result obtained based on the DMRS signal and a channel estimation result obtained based on the additional DMRS signal.

[0176] In some embodiments, the calibration signals may refer to the DMRS signal in the PDSCH channel and the retransmission of the DMRS signal on a next OFDM symbol, as shown in FIG. 9a. In this situation, the first value and the second value of the first information obtained in step S810 may refer to a channel estimation result obtained based on the DMRS signal and a channel estimation result obtained based on the retransmitted DMRS signal.

[0177] It should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0178] In some examples, the two calibration signals in the calibration signal group may be newly added reference signals configured specially, and the newly added reference signals coexist with various existing reference signals in the system. The reference signal associated with the calibration signals may be determined through the relevant configuration information. For example, some REs originally used for data transmission are changed to REs for carrying a new calibration signal, as shown in FIG. 9b.

[0179] In some examples, the two calibration signals in the calibration signal group may be a mixture of the above two cases. For example, the original demodulation reference signal in the data transmission and the newly added dedicated reference signal are simultaneously used as calibration signals to constitute the calibration signal group, as shown in FIG. 9c.

[0180] In some examples, a plurality of different channel estimation results may be obtained based on one calibration signal. That is, based on one calibration signal, the first value and the second value of the first information on the calibration signal can be acquired. Specifically, after a first channel estimation result is obtained based on the calibration signal, an other channel estimation result such as a second channel estimation result (i.e., the second value of the first information) may be obtained through a noise suppression algorithm such as an SVD decomposition algorithm, a Kalman filtering algorithm and a wavelet filtering algorithm based on the first channel estimation result (i.e., the first value of the first information). The above plurality of pieces of channel state information may be combined in pairs or all, to be used to perform subsequent steps. For example, after the first channel estimation result is obtained based on the calibration signal, the first channel estimation result is wavelet-decomposed into a plurality of scales of wavelet coefficients according to a preset wavelet dictionary. Then, threshold processing is performed on each scale of wavelet coefficient, and the coefficient smaller than a threshold is zeroed. Finally, the processed wavelet coefficients are reconstructed, and the reconstruction result is used as the second channel estimation result. The information used in the subsequent steps may be the first channel estimation result, the second channel estimation result, or the combination thereof.

[0181] Referring to FIG. 8, in step S820, the first value and / or the second value (corresponding to the second information in step S420 of FIG. 4) of the second information corresponding to the first value and / or the second value of the first information are obtained through the neural network. According to an embodiment, this step specifically includes: determining the neural network based on an antenna port corresponding to a calibration signal group; and obtaining the first value of the second information corresponding to the first value of the first information and / or the second value of the second information corresponding to the second value of the first information through the neural network.

[0182] In some examples, the calibration signal group is directly associated with the corresponding neural network, i.e., all calibration signals contained in the calibration signal group are associated with one neural network. This neural network is the neural network used in this step.

[0183] In some examples, the corresponding neural network is determined through the reference signal associated with the calibration signal. Specifically, the reference signals associated with all calibration signals contained in the calibration signal group are first obtained. Then, the antenna port group (e.g., the antenna port group in step S420) to which the antenna ports of the above reference signals belong is obtained. The first neural network corresponding to this antenna port group is the neural network used in this step. That is, for the first information obtained based on the calibration signal and the first information obtained based on the reference signal associated with the calibration signal, the same neural network should be used to obtain the second information.

[0184] Here, an exemplary embodiment is given. The first value and the second value of the first information are respectively and . By using the first value of the first information as a sample, the first value of the second information can be obtained through the neural network. By using the second value of the first information as a sample, the second value of the second information can be obtained through the neural network.

[0185] In step S830, the first error is determined based on the first value of the first information and the second value of the second information, and / or the second error is determined based on the second value of the first information and the first value of the second information. For example, in the above exemplary embodiment, the corresponding first error is calculated based on the second value of the first information and the first value of the second information. The corresponding second error is calculated based on the first value of the first information and the second value of the second information.

[0186] In step S840, the neural network is updated based on the first error and / or the second error. For example, based on the total error , the neural network may be updated using an ADAM algorithm. It should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0187] In some examples, the process of obtaining an error includes: inputting a channel estimation result as a sample into the neural network, and obtaining the error based on an output of the neural network and a label. Here, a channel estimation result not used as the sample will be used as the label for the training process of the neural network. For example, a channel estimation result is used as a sample to be inputted into the neural network, and the error is obtained based on the output of the neural network and the label .

[0188] In some examples, the error may be a Euclidean distance, an L2 norm, a cosine similarity, a KL divergence, and the like. In some embodiments, the error may be the MSE between the output of the neural network and the label. For example, if the channel estimation result is a vector, if the output result of the neural network is denoted by , and the label is denoted by , then the error may be , where represents an L2 norm operation. In some embodiments, the error may be the cosine similarity between the output of the neural network and the label. For example, if the channel estimation result is a vector, if the output of the neural network is denoted by , and the label is denoted by , then the error may be , where represents an L2 norm operation.

[0189] In some examples, the error may be a total error after an operation is performed on a plurality of errors, and the communication device updates the neural network based on the total error. For example, the total error is obtained by performing weighted averaging on the errors according to a specific coefficient.

[0190] In some examples, the communication device respectively updates the neural network based on the errors. As an example, the neural network is first updated through the first error. Then, the neural network is updated through the second error. As another example, the neural network is updated only through the first error.

[0191] In some examples, the first information contains three or more channel estimation results (each of which is from the corresponding calibration signal), and the channel estimation results may be combined in pairs by the the communication device to perform the above neural network update step. For example, the first information contains a first value, a second value and a third value of the first information, and the first information is processed by the neural network to obtain the first value, the second value and the third value of the corresponding second information. For example, a first error (e.g., an error determined based on the first value of the first information and the second value of the second information, or an error determined based on the second value of the first information and the first value of the second information, or a total error based on the above errors) is obtained according to the first channel estimation result and the second channel estimation result. A second error (e.g., an error determined based on the second value of the first information and the third value of the second information, or an error determined based on the third value of the first information and the second value of the second information, or a total error based on the above errors) is obtained according to the second channel estimation result and the third channel estimation result. A third error (e.g., an error determined based on the first value of the first information and the third value of the second information, or an error determined based on the third value of the first information and the first value of the second information, or a total error based on the above errors) is obtained according to the third channel estimation result and the first channel estimation result. The three errors are averaged to obtain a total error, and the neural network is updated based on the total error.

[0192] In some examples, an error will be backpropagated layer by layer according to a backpropagation rule, to obtain the error corresponding to each parameter in the neural network. Then, the update amount of the parameter is calculated based on the error of the parameter by using an optimization algorithm, and the parameter is accordingly updated. In some examples, the optimization algorithm may be a gradient descent algorithm, a random gradient descent algorithm, a small-batch gradient descent algorithm, a momentum gradient descent algorithm, a Nesterov gradient acceleration algorithm, an Adagrad method, an AdaDelta method, and an Adam algorithm. Specifically, as an example only rather than a limitation, the common machine learning training algorithm such as derivative clipping, regularization, model pruning and model quantization may alternatively be selected to be used in a network parameter weight update process.

[0193] In FIG. 10, a specific embodiment is given to better illustrate the above training process. The actual training process is as shown in the drawing. In the training process, an optimization formula used to train the neural network may be described as

[0194]

[0195] Here, is a function used to calculate the error amount, and are respectively the first value and the second value of the first information, which may be channel estimation values respectively obtained based on the reference signals on the REs in a set S1 and a set S2, where and , I representing all REs for the transmission of the reference signals. For example, the set S1 refers to an RE that has a subcarrier of which the serial number is an odd number and that is used to carry one calibration signal, and the set S2 refers to an RE that has a subcarrier of which the serial number is an even number and that is used to carry the other calibration signal.

[0196] or and , both can be modeled as

[0197]

[0198] Here, represents additive white Gaussian noise (AWGN), and represents a distortion caused by a non-ideal factor of the system. If the used error amount is an L2 norm, the error amount can be converted to

[0199]

[0200] Here, is a noise variance and is a mean square error (MSE) between and . Clearly, as a channel dimension (e.g., time, frequency and space dimensions) N increases, will decrease significantly, and can be eliminated by ensuring the distance between REs in the set S1 and the set S2, and the subsequent training process. For example, in the training process, the sets S1 and S2 are exchanged to realize the training for bidirectional mapping.

[0201] Therefore, the above optimization formula can be simplified to

[0202]

[0203] Furthermore, the neural network trained after is further processed through the step added in the model will achieve the effect achieved by performing training based on a real channel h and an optimization formula .

[0204] It should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0205] As mentioned above, the association information between the calibration signal and the reference signal and / or the neural network may be contained in the configuration information of the communication device. The configuration information may be the preset information in the communication device, or may be the information received by the communication device from the outside. According to an embodiment, if the foregoing structure is positioned in a terminal-side communication device, the communication device may receive the configuration information. According to an embodiment, if the foregoing structure is positioned in a base station-side communication device, the communication device may transmit the configuration information.

[0206] In some examples, the configuration information and / or the preset information include at least one of: mapping information between the calibration signal and the neural network, information on a time domain resource and / or frequency domain resource used by the calibration signal, information for determining an antenna port for a transmission of the calibration signal, and information on a sequence for generating the calibration signal.

[0207] In some examples, the relevant configuration information may be received through radio resource control signaling (RRC) and / or downlink control information (DCI). Alternatively, the relevant configuration information may alternatively be preset in the communication device.

[0208] The radio resource control signaling (RRC) contains the specific configuration information with respect to the calibration signal, and the specific configuration information may specifically include: the reference signal and / or the neural network associated with the calibration signal. According to an embodiment, the reference signal associated with the calibration signal may be a DMRS signal, a PT-RS signal in a PUSCH and / or PDSCH, or may be a CSI-RS and / or an SRS signal. The neural network associated with the calibration signal is the neural network used in steps S820 and S420.

[0209] In some examples, the antenna port of the calibration signal is determined based on the neural network associated with the calibration signal. Specifically, the communication device should assume that the channel through which the calibration signal passes is similar to or associated with the channel through which the reference signal corresponding to the neural network in step S420 passes. In some embodiments, there is a one-to-one mapping relationship between the serial number of the neural network and the antenna port of the calibration signal. Based on the serial number of the neural network required to be updated, the communication device obtains the antenna port of the corresponding calibration signal by looking up a table, or the like.

[0210] In some examples, there is a one-to-one mapping relationship between the antenna port of the calibration signal and the antenna port of the reference signal associated with the calibration signal. The mapping relationship may be contained in the configuration information. In some embodiments, the calibration signal and the reference signal associated with the calibration signal are transmitted under the same antenna port. In some embodiments, the mapping relationship may refer to an interval-based increment or decrement. For example, if the antenna port of the DMRS given in the configuration information includes antenna ports 0, 1, 4 and 5, then the reference signal associated with the calibration signal of the antenna port 0 is the DMRS of the antenna port 0, the reference signal associated with the calibration signal of the antenna port 1 is the DMRS of the antenna port 1, the reference signal associated with the calibration signal of the antenna port 2 is the DMRS of the antenna port 4, and the reference signal associated with the calibration signal of the antenna port 3 is the DMRS of the antenna port 5. In some embodiments, the mapping relationship may be obtained by checking an index in a first table. For example, the reference signal associated with the calibration signal of an antenna port n is the DMRS of an antenna port p, p being the value corresponding to the index n in the first table.

[0211] The radio resource control signaling (RRC) may further contain: a time-frequency resource occupied by the calibration signal and a generation parameter of the sequence used by the calibration signal. Specifically, the time-frequency resource occupied by the calibration signal may be an absolute value based on a time-frequency resource grid, or may be a relative offset based on the reference signal associated with the calibration signal.

[0212] In some examples, the radio resource control signaling (RRC) may contain antenna port multiplexing related information. Specifically, the orthogonal code used by the calibration signal (e.g., the OCC used by the DMRS) may be given in the configuration information. Based on the orthogonal code, a plurality of different calibration signals may be transmitted through their corresponding antenna ports over the same time-frequency resource.

[0213] In some examples, the radio resource control signaling (RRC) contains an approach to processing the calibration signal with respect to an additional DMRS. In some embodiments, the transmission of the calibration signal is only based on the DMRS signal. For example, the calibration signal appears only in the same OFDM symbol of the DMRS, and there is no calibration signal in the OFDM symbol where the additional DMRS is. In some other embodiments, the transmission of the calibration signal will be based on the DMRS signal and its corresponding additional DMRS. Specifically, the time-frequency resource for the transmission of the calibration signal will be respectively calculated using the DMRS and the additional DMRS based on the same configuration parameter. For example, the OFDM symbol used by the calibration signal and given in the configuration information is a symbol positioned behind the DMRS associated with the calibration signal. In this situation, if the DMRS is positioned on the second OFDM symbol and the additional DMRS is positioned on the eleventh OFDM symbol, then the calibration signal is positioned on the third and twelfth OFDM symbols.

[0214] In some examples, the radio resource control signaling (RRC) may contain a resource type related parameter, for example, an aperiodic resource, a semi-periodic resource, a periodic resource and a corresponding time slot parameter.

[0215] In an exemplary embodiment, if the calibration signal is configured to use the semi-periodic resource or the aperiodic resource, the calibration signal will indicate its opportunity of occurrence based on additional signaling. The calibration signal occurs in certain time slots and / or OFDM symbols in this transmission only if a calibration signal transmitting indication appears in the configuration information in the signaling for scheduling this transmission. The specific position of the calibration signal may be directly indicated by the configuration information in the signaling (e.g., if the configuration information in the signaling is 0, it indicates that there is no calibration signal in the current transmission, and if the configuration information is 1, it indicates that the calibration signal is positioned after the DMRS by one OFDM symbol), or may be obtained through other configuration information (e.g., a transmission opportunity interval given in the semi-periodic resource).

[0216] In an exemplary embodiment, if the calibration signal is configured to use the semi-periodic resource or the periodic resource, the calibration signal will occur on the allocated time-frequency resource in the corresponding time slot according to the corresponding time slot parameter.

[0217] As an example, if being configured to use the periodic resource, the calibration signal will appear, according to a configured time interval T and a configured time slot offset T0, on the corresponding time slot n*T+T0 (n is a natural number). Here, different calibration signals may be configured with different time intervals T and different time slot offsets T0, thereby implementing the alternate transmission for the calibration signals.

[0218] As another example, the calibration signal is configured to use the semi-periodic resource. Before being configured to be turned off through signaling, the calibration signal will occur on each scheduled PDSCH transmission time slot according to the configured time-frequency resource.

[0219] In some examples, a special CSI-RS signal may be configured as a calibration signal. In this situation, the configuration information may include: a generation ID of a sequence used by the CSI-RS signal, a time-frequency resource, and a used antenna port. Specifically, the antenna port used for the transmission of the special CSI-RS signal is the antenna port in the configuration information. However, if steps S420 and S820 are performed based on this calibration signal, the corresponding neural network should be determined based on the DMRS signal associated with the calibration signal. Meanwhile, the antenna port of the CSI-RS resource should be frequency-divided or time-divided.

[0220] In some examples, a special SRS signal on one or more OFDM symbols may be used as a calibration signal. In this situation, the configuration information may include: a generation ID of a sequence used by the signal, a number of OFDM symbols occupied by a time domain, a frequency domain position related parameter, and / or an antenna port used by the signal. It should be noted that, unlike the existing SRS, the time-frequency resource of the special SRS signal does not allow the multi-antenna port multiplexing transmission, and should not be configured as intra-slot frequency hopping in which the repetition factor is an odd number. In some examples, the difference value between the maximum value and the minimum value of the resources (time domain resources and / or frequency domain resources) of the calibration signal is less than a preset threshold. In some examples, there is an interval as a second value between resources (time domain resources and / or frequency domain resources) of the calibration signal with a first value as a reference.

[0221] Here, several exemplary embodiments of the configuration information are given.

[0222] Example 1: The calibration signal may refer to a DMRS signal (including an additional DMRS signal) associated with the calibration signal and a repetition of the DMRS signal (and the additional DMRS signal). Two calibration signals will use identical transmitting sequences (i.e., both the sequences are the transmitting sequence of the DMRS signal). In this situation, the content of the RRC may include at least one of: an OFDM symbol offset and a subcarrier offset between the repeated DMRS and the DMRS signal, and the repeated DMRS using the same antenna port as the DMRS signal. For example, the content included in the RRC may be as shown in Table 1.

[0223] [Table 1]

[0224]

[0225] In the above RRC configuration, the OFDM symbol used by the repeated DMRS in the calibration signals is the next OFDM symbol of the OFDM symbol used by the DMRS signal, the serial number of the subcarrier used by the repeated DMRS is the same as that of the DMRS signal, and the repeated DMRS uses the same antenna port and transmitting sequence (including OCC, precoding, and the like) as the DMRS signal. In addition, the calibration signal may further contain a repetition of the additional DMRS signal, which is positioned in the next OFDM symbol of the OFDM symbol used by the additional DMRS signal and uses the same subcarrier as the additional DMRS signal. Moreover, the parameters such as the antenna port remain unchanged.

[0226] Example 2: The calibration signal may refer to a new signal. In this sitution, the content of the RRC may include at least one of: an associated neural network, a sequence ID for generating a transmitting sequence, an OFDM symbol position of the calibration signal, a subcarrier position of the calibration signal, and an antenna port of the calibration signal. For example, the content included in the RRC may be as shown in Table 2.

[0227] [Table 2]

[0228]

[0229] In the above RRC configuration, the new reference signal is used as a transmitting signal based on the transmitting sequence (e.g., a ZC sequence) generated through the ID given in the configuration information. This calibration signal will occur on the sixth OFDM symbol in each time slot and occur every four RBs, and the comb distribution of two calibration signals is implemented based on comb2 (the subcarrier interval is 2) in a corresponding RB. Only a corresponding calibration signal is transmitted on the above time-frequency resource, the antenna port of the calibration signal is determined by querying a table in according to the neural network associated with the calibration signal.

[0230] Example 3: The configuration for the calibration signal is based on the situation of a CSI-RS resource. The transmission of the calibration signal is based on the definition in an existing NR system. In this situation, the content of the RRC may include at least one of: the CSI-RS resource, and a selected CSI-RS antenna port. For example, the content included in the RRC may be as shown in Table 3.

[0231] [Table 3]

[0232]

[0233] In the above RRC configuration, the calibration signal is an CSI-RS signal having antenna ports 0 and 1 in the CSI-RS signals generated by the configuration information numbered #1 in the CSI-RS resource configuration in the system. The calibration signal will be used to perform step S820 on the neural network corresponding to the DMRS of the antenna port 0.

[0234] Example 4: The calibration signal refers to an SRS signal on two OFDM symbols. The transmission of the calibration signal is based on the definition in an existing NR system. In this situation, the content of the RRC may include at least one of: a generation ID of a transmitting sequence (e.g., a ZC sequence) used by the SRS, a number of OFDM symbols occupied by a time domain, a frequency domain position related parameter, and / or a frequency hopping related parameter. For example, the content included in the RRC may be as shown in Table 4.

[0235] [Table 4]

[0236]

[0237] In the above RRC configuration, the calibration signal is a special SRS signal positioned on two OFDM symbols. The calibration signal is a transmitting sequence (e.g., a ZC sequence) obtained based on 8192, and comb-shaped in a frequency domain, and occupies the last two OFDM symbols in a time slot. The calibration signal will be used to perform step S820 on the neural network corresponding to the DMRS of the antenna port 0.

[0238] In some examples, the content of the RRC may further include a frequency hopping related parameter such as a frequency hopping mode and / or a repetition factor. Here, the frequency hopping mode may refer to no frequency hopping, inter-slot frequency hopping, or intra-slot frequency hopping. If the frequency hopping mode refers to the inter-slot frequency hopping, a plurality of calibration signals use fixed time-frequency resources in one transmission time slot, but use different resources in different time slots, and the resources are replaced at intervals of R time slots, where R is a repetition factor. If the frequency hopping mode refers to the intra-slot frequency hopping, a plurality of calibration signals use different resources in one transmission time slot, and the resources are replaced once at intervals of R OFDM symbols, where R is a repetition factor. The configuration of the repetition factor should ensure that adjacent reference signals can be transmitted on at least two sets of REs.

[0239] As an example, in a situation where a DMRS in a PUSCH and a repetition of the DMRS are used as calibration signals, the repeated DMRS signal may be configured in an inter-slot frequency hopping mode. In this mode, the repeated DMRS signal will be transmitted on only some subcarriers in each time slot, and the subcarriers is replaced by different subcarriers each time the repeated DMRS is transmitted. The traversal for all frequency domain resources is achieved through a plurality of PUSCH transmissions, thereby reducing the overhead of the reference signal in each transmission.

[0240] As another example, in a situation where a special SRS signal occupying 4 OFDM symbols is used as calibration signals, the SRS signal may be configured in an intra-slot frequency hopping mode in which the repetition factor is 2. Through this configuration, it may be implemented that the first two calibration signals and the last two calibration signals are respectively transmitted on the same subcarrier in the upper half part and the lower half part of the frequency domain. If the opposite frequency hopping mode of an other terminal is used together, it may be implemented that the two terminals multiplex the same calibration signal transmission time slot.

[0241] The downlink control information (DCI) may contain reception indication information for the calibration signal, for indicating the occurrence of the calibration signal. For example, the DCI for scheduling a data transmission may include one bit for indicating whether a calibration signal will occur during this transmission. If the bit is 1, this transmission contains the calibration signal. Conversely, this transmission does not contain the calibration signal.

[0242] In some examples, the DCI for scheduling the data transmission may further include indication information for a calibration signal configuration, and the information will indicate the calibration signal configuration used in this transmission. That is, the RRC signaling may transmit various different calibration signal configuration information, and the terminal determines specific configuration information of the calibration signal to be received, according to the indication in the DCI signaling.

[0243] In some examples, the DCI for scheduling the data transmission may further include indication information for an antenna port used by the calibration signal, i.e., the antenna port used by the calibration signal contained in the PDSCH and / or the PUSCH scheduled by the DCI.

[0244] Alternatively, if this structure is positioned in a terminal-side communication device, this step may further include: transmitting (or reporting) information on a calibration related capability supported by the terminal; and receiving calibration process related configuration information.

[0245] In some examples, the calibration related capability supported by the terminal may include at least one of: a capability to receive and process the foregoing calibration signals, a physical resource configuration that the terminal can support, a terminal computing capability, a time required to complete a calibration process, and the like.

[0246] Here, an exemplary embodiment is given. The capability information transmitted (or reported) by the communication device may include, for example, at least one of: a capability level that the device can support. The capability level is used to describe the capability of the terminal to perform the neural network-based channel estimation process in the above step S420 and the neural network online update process in step S840. Here,

[0247] a Level-1 terminal does not support the use of the neural network-based channel estimation process;

[0248] a Level-2 terminal supports the use of the neural network-based channel estimation process; and

[0249] a Level-3 terminal supports not only the neural network-based channel estimation process but also the neural network online update process.

[0250] Alternatively, the Level-2 terminal may be further subdivided into:

[0251] a Level 2-1 terminal that supports the neural network-based channel estimation process performed based on the DMRS signal;

[0252] a Level 2-2 terminal that supports the neural network-based channel estimation process performed based on the DMRS signal and the PT-RS signal; and

[0253] a Level 2-2 terminal that supports the neural network-based channel estimation process performed based on the DMRS signal, the PT-RS signal and the CSI-RS signal.

[0254] Alternatively, the Level-3 terminal may be further subdivided into:

[0255] a Level 3-1 terminal that only supports the neural network online update performed using the calibration signal group composed of a single calibration signal;

[0256] a Level 3-2 terminal that supports the neural network online update performed using the calibration signal group composed of the DMRS signal;

[0257] a Level 3-3 terminal that supports the neural network online update performed using the calibration signal group composed of a new calibration signal; and

[0258] a Level 3-4 terminal that supports the neural network online update performed using the calibration signal group composed of the DMRS signal and / or the new calibration signal.

[0259] In some examples, the capability information transmitted (or reported) by the communication device may further include a time-frequency resource mode used by the calibration signal supported by the terminal. Specifically, the time-frequency resource mode may include at least one of: a position of a subcarrier where the supported calibration signal occurs, an interval between subcarriers where the supported calibration signal occurs, a position of an OFDM symbol where the supported calibration signal occurs, an interval between OFDM symbols where the supported calibration signal occurs, and the like.

[0260] In some examples, the capability information transmitted (or reported) by the communication device may further include a neural network update opportunity supported by the terminal. Specifically, the neural network update opportunity supported by the terminal may be, for example, once per time slot, once every two time slots, and once per data transmission.

[0261] FIG. 11 illustrates a flowchart of a method performed by a communication device according to an exemplary embodiment of the present disclosure.

[0262] According to the embodiment, the communication device can be any device capable of acquiring signals and performing signal detection. For example, the communication device can be a base station or a terminal, but is not limited thereto. For instance, it can even be a user equipment with a large number of antenna ports, etc.

[0263] As shown in FIG. 11, at step S1110, configuration information related to the reference signals is received. The received configuration information may be configuration information related to a second reference signal (or a second reference signal port) in multiplexing association with the first reference signal. Herein, the expression "in multiplexing association" means that both the first reference signal (or the first reference signal port) and the second reference signal (or the second reference signal port) are mapped on the same time-domain and / or frequency-domain resource. At step S1120, it is determined whether the first reference signal (or the first reference signal port) is multiplexed (or whether it is non-exclusive / non-independent). "The first reference signal is multiplexed" means that there is a second reference signal transmitted by being superimposed on the time-domain and / or frequency-domain resources used by the first reference signal (for example, superimposed based on Orthogonal Cover Code (OCC)). "The first reference signal is not multiplexed" means that only the first reference signal is carried on the time-domain resources and / or frequency-domain resources used by the first reference signal.

[0264] The first reference signal and the second reference signal may be reference signals of the same type using different antenna ports, such as, DMRS or CSI-RS.

[0265] According to a method of the present disclosure, a received first reference signal (or first reference signal port) can be divided according to time-frequency resources. For example, the first reference signal (or the first reference signal port) and the second reference signal (or the second reference signal port) are multiplexed in the same time-domain and / or frequency-domain resources according to OCC multiplexing. As an example, the first reference signal port uses the OCC code [1,1] and the second reference signal port uses the OCC code [1, -1]. At this time, each symbol in the two reference signals can be extended into two new symbols and mapped to two adjacent sub-carriers (frequency-domain resources) and / or two consecutive OFDM symbols (time-domain resources). If the first reference signal is not multiplexed, the time-domain and / or frequency-domain resource carrying the two extended new symbols can be considered as being used for carrying two calibration signals. For example, the time-frequency resource on the previous OFDM symbol carries the first calibration signal, and the time-frequency resource on the latter OFDM symbol carries the second calibration signal. Alternatively, the time-frequency resource with a lower sub-carrier number carries the first calibration signal, and the time-frequency resource with a higher sub-carrier number carries the second calibration signal. In this way, the communication device can obtain two calibration signals based on the received reference signals with lower downlink signaling overhead, and achieve the online update of its internal neural network. For the specific process, refer to steps S810-S840 and other contents.

[0266] Next, the contents involved in the above-mentioned steps S1110 to S1120 will be described in detail respectively.

[0267] Referring to FIG.11, at step S1110, configuration information related to the reference signals is received. The received configuration information may be configuration information related to a second reference signal (or a second reference signal port) in multiplexing association with the first reference signal. Herein, the expression "in multiplexing association" means that both the first reference signal (or the first reference signal port) and the second reference signal (or the second reference signal port) are mapped on the same time-domain and / or frequency-domain resource. The configuration information can be obtained through Radio Resource Control (RRC) messages and / or Medium Access Control (MAC) messages and / or Downlink Control Information (DCI). For example, the configuration information related to the second reference signal that has a multiplexing association with the first reference signal may include at least one of the following: information on whether a second reference signal is transmitted or used, information on whether a second reference signal is mapped to the same time-domain and / or frequency-domain resource as the first reference signal, and / or information on the multiplexing manner of a second reference signal in the time-domain and / or frequency-domain resources.

[0268] At step S1120, it is determined whether the first reference signal (or the first reference signal port) is multiplexed (or whether it is non-exclusive / non-independent). According to an embodiment, if reference signals from a plurality of antenna ports are superimposed on the same time-frequency resources for transmission (for example, superimposed based on Orthogonal Cover Code (OCC)), these reference signals are considered to be in a multiplexing state, or the antenna ports of the reference signals are in a multiplexing state, or the time-frequency resource of the reference signals is in a multiplexing state. To the contrary, if only a reference signal from one antenna port is carried on the time-frequency resource used by the reference signal, the reference signal is in a non-multiplexing state, or the antenna port of the reference signal is in a non-multiplexing state, or the time-frequency resource of the reference signal is in a non-multiplexing state.

[0269] In an embodiment, the received configuration information includes information indicating transmission / usage of at least one second reference signal.

[0270] In some examples of the embodiment, DCI signaling may include indication information indicating, for all DMRS scheduled by the current DCI, whether a second reference signal in multiplexing association therewith is transmitted / used. According to the indication information, the UE can determine whether the DMRS scheduled by the current DCI is multiplexed. For example, the DCI signaling include information of 1 bit. If the bit is “1”, it is indicated that, for at least one DMRS in all the DMRS scheduled by the current DCI (i.e., a first reference signal), at least one second reference signal in multiplexing association therewith is transmitted. Then, the UE considers that all DMRS scheduled by the current DCI are in the multiplexing state. If the bit is “0”, it is indicated that for each DMRS in all the DMRS scheduled by the current DCI (i.e., a first reference signal), no second reference signal in multiplexing association therewith is transmitted. Then, the UE considers that all DMRS scheduled by the current DCI are in the non-multiplexing state. Vice versa. Details are not elaborated here.

[0271] In some examples of the embodiment, DCI signaling may include indication information indicating, for each reference signal scheduled by the current DCI, whether a second reference signal in multiplexing association therewith is transmitted / used. According to the indication information, the UE can determine whether each DMRS scheduled by the current DCI is multiplexed. For example, the DCI includes a bit map, each bit in which is corresponding to a respective reference signal (or the corresponding antenna port) scheduled by the current DCI. If a bit is “1”, it is indicated that, for the DMRS corresponding to the bit (i.e., a first reference signal), at least one second reference signal in multiplexing association therewith is transmitted. Then, the DMRS corresponding to the bit is in the multiplexing state. If the bit is “0”, it is indicated that for the DMRS corresponding to the bit, no second reference signal in multiplexing association therewith is transmitted. Then, the DMRS corresponding to the bit is in the non-multiplexing state. Vice versa. Details are not elaborated here.

[0272] In some examples of the embodiment, DCI signaling may include indication information indicating, for some reference signals scheduled by the current DCI, whether a second reference signal in multiplexing association therewith is transmitted / used. Here, said some reference signals are the reference signals among all the reference signals (or the corresponding antenna ports) scheduled by the current DCI, the multiplexing or non-multiplexing state of which cannot be determined, for example, a certain scheduled DMRS belongs to a CDM group different from other scheduled DMRS. For all the remaining reference signals, the UE can determine, for each of them, whether it is multiplexed according to other information (such as, two reference signals belong to a same CDM group). For example, the DCI includes a bit map, each bit in which is corresponding to a respective reference signal among said some reference signals (or the corresponding antenna ports) scheduled by the current DCI. If a bit is “1”, it is indicated that, for the DMRS corresponding to the bit (i.e., a first reference signal), at least one second reference signal in multiplexing association therewith is transmitted. Then, the DMRS corresponding to the bit is in the multiplexing state. If the bit is “0”, it is indicated that for the DMRS corresponding to the bit, no second reference signal in multiplexing association therewith is transmitted. Then, the DMRS corresponding to the bit is in the non-multiplexing state. All the remaining DMRS scheduled by the current DCI are in the multiplexing state (these DMRS belong to same CDM group and can be considered as the second reference signals for each other). Vice versa. Details are not elaborated here.

[0273] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the scheduled reference signals that can be scheduled. The indication information for each reference signal is used for indicating whether a second reference signal in multiplexing association with the reference signal (or the corresponding antenna port) is to be transmitted. According to the reference signal (or the corresponding antenna port) scheduled by the current DCI, the UE may search the corresponding indication information, and determine the multiplexing or non-multiplexing state of each reference signal scheduled by the current DCI. For example, the configuration information may include indication information indicating that the second reference signal in multiplexing association with port #0 will be transmitted, and the second reference signal in multiplexing association with port #2 will not be transmitted. Thus, if the antenna port of the DMRS scheduled by DCI is port #0, it can be determined that the DMRS is in the multiplexing state by searching the configuration information corresponding to the port #0. If the antenna port of the DMRS scheduled by DCI is port #2, it can be determined that the DMRS is in the non-multiplexing state by searching the configuration information corresponding to the port #2. For another example, the configuration information may include indication information indicating that the second reference signal in multiplexing association with port #0 will be transmitted, and the second reference signal in multiplexing association with port #2 will not be transmitted. If the antenna ports of the DMRS scheduled by DCI are ports #0, #1 and #2, since ports #0 and #1 belong to the same CDM group, the UE can determine that the DMRS corresponding to ports #0 and #1 are in the multiplexing state. In addition, by searching the configuration information corresponding to the port #2, it can be determined that the DMRS corresponding to port #2 is in the non-multiplexing state.

[0274] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that may be scheduled. The indication information for each reference signal is used for indicating whether the reference signal is to be transmitted (whether to be transmitted on the antenna port of the reference signal). According to the reference signal (or the corresponding antenna port) scheduled by the current DCI, the UE may search the indication information corresponding to the reference signal in multiplexing association therewith, and determine the multiplexing or non-multiplexing state of each reference signal scheduled by the current DCI. For example, the configuration information may include indication information indicating that the reference signal at port #0 will be transmitted, and the reference signal at port #2 will not be transmitted. Thus, if the antenna port of the DMRS scheduled by DCI is port #1, it can be determined that the DMRS at port #1 scheduled by the current DCI is in the multiplexing state by searching the configuration information corresponding to the port #0 that is in multiplexing association with port #1. If the antenna port of the DMRS scheduled by DCI is port #3, it can be determined that the DMRS at port #3 scheduled by the current DCI is in the non-multiplexing state by searching the configuration information corresponding to the port #2 that is in multiplexing association with port #3.

[0275] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that may be scheduled. Based on indication information for the group (such as the CDM group) to which the reference signal belongs, it can be determined that whether the second reference signal in multiplexing association with the first reference signal is to be transmitted. For example, the configuration information indicates that the second reference signal in multiplexing association with reference signals in CDM group #0 will not be transmitted. If the antenna port of the DMRS scheduled by DCI is port #1, by searching the indication information of CDM #0 to which port #1 belongs, it is determined that the second reference signal in multiplexing association with the DMRS at port #1 will not be transmitted. Thus, the DMRS at port #1 scheduled by the current DCI is in the non-multiplexing state.

[0276] In some embodiments, the received configuration information includes information on whether the second reference signal is mapped on the same time-domain and / or frequency-domain resources as the first reference signal, and / or information on the multiplexing manner of at least one second reference signal in the time-domain and / or frequency-domain resources.

[0277] In some examples in an embodiment, DCI includes indication information for indicating whether a second reference signal is mapped on the same time-domain and / or frequency-domain resources as the first reference signal (all the second reference signals use the same manner for resource multiplexing), wherein all the reference signals scheduled by the current DCI can be used as the first reference signal, and all the reference signals in multiplexing association with the first reference signal can be used as the second reference signals. For example, if the single-symbol DMRS is used, the DCI includes information of 1 bit. If the bit is “1”, the second reference signal is transmitted using a frequency-domain coding different from that of the first reference signal. If the bit is “0”, no second reference signal is transmitted. Vice versa. Details are not elaborated here. For another example, if the double-symbol DMRS is used, the DCI includes information of 1 bit. If the bit is “1”, the second reference signal is transmitted using a time-domain coding different from that of the first reference signal. If the bit is “0”, no second reference signal is transmitted. Vice versa. Details are not elaborated here. For another example, if the double-symbol DMRS is used, the DCI includes information of 2 bits, the value of which may be “00”, “01”, “10” and “11” for indicating four different situations, including no second reference signal is transmitted, the second reference signal is transmitted using a time-domain coding different from that of the first reference signal, the second reference signal is transmitted using a frequency-domain coding different from that of the first reference signal, and the second reference signal is transmitted using a time-domain coding and a frequency-domain coding different from those of the first reference signal.

[0278] In some examples in an embodiment, DCI includes indication information for indicating whether each second reference signal in multiplexing association with all or some of the reference signals scheduled by the current DCI is mapped on the time-domain and / or frequency-domain resources as same as the first reference signal. According to the indication information, the UE may determine whether each DMRS scheduled by the current DCI is multiplexed. For example, the DCI signaling may include a plurality of bit maps. Each bit map is corresponding to a reference signal (the corresponding antenna port) scheduled by the current DCI, in which each bit is corresponding to a second reference signal in multiplexing association with the reference signal, e.g., other DMRS belonging to a same CDM group with the DMRS. If a bit in a bit map is “1”, it is indicated that a second reference signal corresponding to this bit is transmitted and the DMRS corresponding to the bit map is in the multiplexing state. The UE may determine that the second reference signal is in time-domain and / or frequency-domain multiplexing state with the first reference signal (the manner of multiplexing association between the second reference signal and the first reference signal is preset.) If all bits in a bit map corresponding to the DMRS are “0”, it is indicated that no second reference signal in multiplexing association with the DMRS corresponding to the bit map is transmitted, and the DMRS corresponding to the bit map is in the non-multiplexing state. Vice versa. Details are not elaborated here.

[0279] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the possible reference signals. The indication information for each reference signal is used for indicating the multiplexing manner of the second reference signal (or the corresponding antenna portion) in multiplexing association with the reference signal in time-domain and / or frequency-domain resources. The UE may search the corresponding indication information according to the reference signal (or the corresponding antenna portion) scheduled by the current DCI, and determine the multiplexing or non-multiplexing state of each reference signal scheduled by the current DCI. For example, the configuration information may include indication information indicating that the second reference signal in multiplexing association with port #0 will not multiplex the time-frequency resource, and the second reference signal in multiplexing association with port #2 will multiplex the time-domain resource. Thus, if the antenna port of the DMRS scheduled by DCI is port #0, it can be determined the DMRS at port #0 is in the non-multiplexing state by searching the configuration information corresponding to the port #0. If the antenna port of the DMRS scheduled by DCI is port #2, it can be determined the DMRS at port #2 is in the multiplexing state for time-domain resources by searching the configuration information corresponding to the port #2. For another example, the configuration information may include indication information indicating that the second reference signal in multiplexing association with port #0 will not multiplex time-frequency resources, the second reference signal in multiplexing association with port #1 will multiplex frequency-domain resources, and the second reference signal in multiplexing association with port #2 will not multiplex time-frequency resources. If the antenna ports of the DMRS scheduled by DCI are ports #0, #1 and #2, since ports #0 and #1 belong to a same CDM group, the UE can determine that the DMRS corresponding to ports #0 and #1 are in the multiplexing state, and that the DMRS corresponding to port #2 is in the non-multiplexing state by searching the configuration information corresponding to the port #2.

[0280] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that can be scheduled. The indication information for each reference signal is used for indicating the multiplexing manner of the reference signal (or the corresponding antenna portion) as the second reference signal. According to the reference signal (or the corresponding antenna portion) scheduled by the current DCI, the UE may search the indication information corresponding to the reference signal in multiplexing association therewith, and determine the multiplexing or non-multiplexing state of each reference signal scheduled by the current DCI. For example, the configuration information may include indication information indicating that port #0 is multiplexed in frequency domain, port #2 is not multiplexed. Thus, if the antenna port of the DMRS scheduled by DCI is port #1, it can be determined that the DMRS at port #1 scheduled by the current DCI is in the multiplexing state by searching the configuration information corresponding to the port #0 that is in multiplexing association with port #1. If the antenna port of the DMRS scheduled by DCI is port #3, it can be determined that the DMRS at port #3 scheduled by the current DCI is in the non-multiplexing state by searching the configuration information corresponding to the port #2 that is in multiplexing association with port #3.

[0281] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that may be scheduled. Based on indication information for the group (such as the CDM group) to which the reference signal belongs, it can be determined that whether the reference signals in the group are mapped on the same time-domain and / or frequency-domain resources. For example, it is indicated in the configuration information that the DMRS belonging to CDM group #1 do not use the same time-domain and / or frequency-domain resources. If the antenna port of the DMRS scheduled by the DCI is port #1, by searching the indication information of CDM group #0 to which port #1 belongs, it can be determined that the second reference signals in multiplexing association with port #1 do not use the same time-domain and / or frequency-domain resources, and that the DMRS at port #1 scheduled by the current DCI is in non-multiplexing state. For another example, it is indicated in the configuration information that the DMRS belonging to CDM group #0 use the same frequency-domain resources. If the antenna port of the DMRS scheduled by the DCI is port #1, by searching the indication information of CDM group #0 to which port #1 belongs, it can be determined that the second reference signals in multiplexing association with port #1 use the same frequency-domain resources, and that the DMRS at port #1 scheduled by the current DCI is in frequency-domain multiplexing state.

[0282] FIG. 12 shows a flowchart of a method executed by a communication device according to an embodiment of the present disclosure.

[0283] According to an embodiment, the communication device can be any device capable of acquiring signals and performing signal detection. For example, the communication device can be a base station or a terminal, but is not limited thereto. For instance, it can even be a user equipment with a large number of antenna ports, etc.

[0284] As shown in FIG. 12, at step S1210, configuration information related to a reference signal is received. The reference signal can be, for example, DMRS, CSI-RS, etc. At step S1220, it is determined whether the time-frequency resources mapped by the reference signal (or the reference signal port) are exclusive to the reference signal (or the reference signal port). Here, “exclusive to the reference signal (or the reference signal port)” means that all the time-frequency resources used by this reference signal are not used by other downlink signals (such as different antenna ports of the same reference signal, different reference signals, other physical channels like PDSCH).

[0285] According to the method of the present disclosure, if the UE determines that the time-frequency resources mapped by the reference signal are exclusive to the reference signal, the time-domain and / or frequency-domain resources used to carry each symbol of this reference signal can be regarded as being used to carry two calibration signals. For example, the time-frequency resources on the previous OFDM symbol carry the first calibration signal, and the time-frequency resources on the subsequent OFDM symbol carry the second calibration signal. In this way, the communication device can obtain two calibration signals based on the received reference signals with lower downlink signaling overhead, and achieve the online update of its internal neural network. For the specific process, refer to steps S810-S840 and other contents.

[0286] Next, the contents involved in the above-mentioned steps S1210 to S1220 will be described in detail respectively.

[0287] Referring to FIG. 12, at step S1210, configuration information related to a reference signal is received. The reference signal can be, for example, DMRS or CSI-RS. The configuration information can be obtained through Radio Resource Control (RRC) messages and / or Medium Access Control (MAC) messages and / or Downlink Control Information (DCI). The configuration information includes at least one of the following: information indicating whether the reference signal exclusively occupies time-frequency resources; information indicating the multiplexing method of the reference signal in time-domain and / or frequency-domain resources, etc.

[0288] At step S1220, it is determined whether the time-frequency resources mapped by the reference signal (or the reference signal port) are exclusive to the reference signal (or the reference signal port). Here, “exclusive to the reference signal (or the reference signal port)” means that all the time-frequency resources used by this reference signal are not used by other downlink signals (such as different antenna ports of the same reference signal, different reference signals, other physical channels like PDSCH). According to the embodiment, reference signals of multiple different antenna ports can be transmitted using the same time-frequency resources (for example: based on Orthogonal Cover Code (OCC)). The communication device can determine whether there are reference signals of other antenna ports on the time-frequency resources based on the configuration information related to the reference signal.

[0289] In some examples, the received configuration information contains corresponding indication information, which is used to indicate whether the currently scheduled reference signal exclusively occupies the time-frequency resources it uses, that is, whether there are reference signals of other antenna ports on the time-frequency resources used by the currently scheduled reference signal.

[0290] In some examples in the embodiment, the DCI signaling may include indication information for indicating whether all the DMRS scheduled by the current DCI exclusively occupy the time-frequency resources they use. For example, the DCI signaling includes information of 1 bit. If the bit is “1”, it is indicated that at least one DMRS scheduled by the current DCI does not exclusively occupy the time-frequency resources it uses. If the bit is “0”, it is indicated that all DMRS scheduled by the current DCI exclusively occupy the time-frequency resources they use. Vice versa. Details are not elaborated here.

[0291] In some examples in the embodiment, the DCI signaling may include indication information for indicating whether each reference signal scheduled by the current DCI exclusively occupies the time-frequency resources it uses. For example, the DCI signaling includes a bit map, each bit in which corresponds to a respective reference signal (the corresponding antenna port) scheduled by the current DCI. If a bit is “1”, it is indicated that the corresponding DMRS exclusively occupies the time-frequency resources it uses. If a bit is “0”, it is indicated that the corresponding DMRS does not exclusively occupy the time-frequency resources it uses. Vice versa. Details are not elaborated here.

[0292] In some examples in the embodiment, the DCI signaling may include indication information for indicating whether some reference signals among all the reference signals scheduled by the current DCI exclusively occupy the time-frequency resources they use, wherein said some reference signals refer to the reference signals among all the reference signals (the corresponding antenna ports) scheduled by the current DCI for which it cannot be determined whether they exclusively occupy the time-frequency resources they use (for example, a certain scheduled DMRS belongs to a CDM group different from other scheduled DMRS). Whether the remaining reference signals exclusively occupy the time-frequency resources they use can be determined based on other information (such as the CDM group they belong to). For example, the DCI signaling schedules three antenna ports #0, #1 and #2. Since port #0 and port #1 belong to the same CDM group, the UE may determine that their time-frequency resources are not exclusively occupied. The state of port #2 may be determined based on the 1-bit information included in the DCI signaling. For example, if the bit is “1”, it is indicated that the DMRS at port #2 exclusively occupies the time-frequency resource. If the bit is “0”, it is indicated that the DMRS at port #2 does not exclusively occupy the time-frequency resource. Vice versa. Details are not elaborated here.

[0293] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the scheduled reference signals that can be scheduled. The indication information for each reference signal is used for indicating whether the reference signal (or the corresponding antenna port) (if being scheduled) exclusively occupies the time-frequency resource. According to the reference signal (or the corresponding antenna port) scheduled by the current DCI, the UE may search the corresponding indication information, and determine whether each reference signal scheduled by the current DCI exclusively occupies the time-frequency resource it uses. For example, the configuration information may include indication information indicating that port #0 (if being scheduled) exclusively occupies the time-frequency resource it uses, and port #2 (if being scheduled) does not exclusively occupy the time-frequency resource it uses. Thus, if the antenna port of the DMRS scheduled by DCI is port #0, it can be determined, by searching the configuration information corresponding to the port #0, that the DMRS at port #0 exclusively occupies the time-frequency resource it uses. If the antenna port of the DMRS scheduled by DCI is port #2, it can be determined, by searching the configuration information corresponding to the port #2, that the DMRS at port #2 does not exclusively occupy the time-frequency resource it uses. For another example, the configuration information may include indication information indicating that port #0 (if being scheduled) exclusively occupies the time-frequency resource it uses, and port #2 (if being scheduled) does not exclusively occupy the time-frequency resource it uses. If the antenna ports of the DMRS scheduled by DCI are ports #0, #1 and #2, since ports #0 and #1 belong to the same CDM group, the UE can determine that the DMRS corresponding to ports #0 and #1 do not exclusively occupy the time-frequency resource they use, without searching the configuration information. In addition, by searching the configuration information corresponding to the port #2, it can be determined that the DMRS corresponding to port #2 does not exclusively occupy the time-frequency resource it uses.

[0294] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that may be scheduled. Based on indication information for the group (such as the CDM group) to which the reference signal belongs, it can be determined that whether the reference signals in the group exclusively occupy the time-frequency resource they use. For example, the configuration information may indicate that the reference signals belonging to CDM group #1 exclusively occupy the time-frequency resource they use, If the antenna port of the DMRS scheduled by DCI is port #1, by searching the indication information of CDM #0 to which port #1 belongs, it is determined that the DMRS at port #1 exclusively occupies the time-frequency resource it uses.

[0295] In some embodiments, the received configuration information contains information indicating the multiplexing mode of the reference signal in time-domain and / or frequency-domain resources.

[0296] In some examples in an embodiment, DCI includes indication information for indicating the multiplexing manner of all reference signals scheduled by the current DCI in time-domain and / or frequency-domain resources. For example, if the single-symbol DMRS is used, the DCI includes information of 1 bit. If the bit is “1”, the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of frequency-domain coding. If the bit is “0”, the DMRS scheduled by the current DCI exclusively occupies the time-frequency resource. Vice versa. Details are not elaborated here. For another example, if the double-symbol DMRS is used, the DCI includes information of 1 bit. If the bit is “1”, the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of time-domain coding. If the bit is “0”, the DMRS scheduled by the current DCI exclusively occupies the time-frequency resource. Vice versa. Details are not elaborated here. For another example, if the double-symbol DMRS is used, the DCI includes information of 2 bits, the value of which may be “00”, “01”, “10” and “11” for indicating four different situations, including the DMRS scheduled by the current DCI exclusively occupies time-frequency resources, the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of frequency-domain coding, the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of time-domain coding, and the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of both time-domain coding and frequency-domain coding.

[0297] In some examples in an embodiment, DCI includes indication information for indicating the multiplexing manner of all or some of the reference signals scheduled by the current DCI in time-domain and / or frequency-domain resources. For example, the DCI signaling includes a bit map, each two bits in which correspond to the multiplexing manner of a respective reference signal scheduled by the current DCI. The value may be “00”, “01”, “10” and “11” for indicating four different situations, including the DMRS scheduled by the current DCI exclusively occupies time-frequency resources, the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of frequency-domain coding, the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of time-domain coding, and the DMRS scheduled by the current DCI multiplexes time-frequency resources by means of both time-domain coding and frequency-domain coding. For another example, the DCI signaling includes a bit map, each two bits in which correspond to the multiplexing manner of a respective reference signal among some reference signals (DMRS belonging to a CDM group different from other DMRS) scheduled by the current DCI, with the meaning similar to the above example. Details are not elaborated here. Specifically, if the current DCI schedules DMRS at ports #0, #1 and #2, since ports #0 and #1 belong to the same CDM group, it can be determined that the time-frequency resources are multiplexed by means of frequency-domain coding; the DCI includes indication information of 2 bits for indicating the multiplexing manner in time-domain and / or frequency-domain resources for the DMRS at port #2.

[0298] In some examples in the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that can be scheduled. The indication information for each reference signal is used for indicating the multiplexing manner of the reference signal (or the corresponding antenna portion) in time-domain and / or frequency-domain resources. The UE may search the corresponding indication information according to the reference signal (or the corresponding antenna portion) scheduled by the current DCI, and determine whether each reference signal scheduled by the current DCI exclusively occupies the time-frequency resource. For example, the configuration information may include indication information indicating that port #0 exclusively occupies the time-frequency resource, and port #2 is multiplexed in frequency-domain resources. Thus, if the antenna port of the DMRS scheduled by DCI is port #0, it can be determined the DMRS at port #0 exclusively occupies the time-frequency resource by searching the configuration information corresponding to the port #0. If the antenna port of the DMRS scheduled by DCI is port #2, it can be determined the DMRS at port #2 is multiplexed in frequency-domain resources by searching the configuration information corresponding to the port #2. For another example, the configuration information may include indication information indicating that port #0 exclusively occupies the time-frequency resource, port #1 is multiplexed in frequency-domain resources, and port #2 is multiplexed in time-frequency resources. If the antenna ports of the DMRS scheduled by DCI are ports #0, #1 and #2, since ports #0 and #1 belong to a same CDM group, the UE can determine that the DMRS corresponding to ports #0 and #1 are multiplexed in frequency-domain resources (based on a preset for the CDM group), and that the DMRS corresponding to port #2 exclusively occupies the time-frequency resource by searching the configuration information corresponding to the port #2.

[0299] In some examples of the embodiment, the configuration information obtained from RRC signaling may include indication information for all the reference signals that may be scheduled. Based on indication information for the group (such as the CDM group) to which the reference signal belongs, it can be determined whether the reference signals in the group multiplex time-domain and / or frequency-domain resources. If a CDM group is indicated as exclusive, the communication device can consider that the reference signals using any antenna port belonging to the CDM group exclusively occupy the time-frequency resources they use. If the CDM group is indicated as being multiplexed in time domain (exclusive in frequency domain), the communication device can consider that, on the time-frequency resources used by the reference signals of any antenna port belonging to the CDM group, there are only reference signals of different antenna ports using different time-domain coding. If the CDM group is indicated as being multiplexed in frequency domain (exclusive in time domain), the communication device can consider that, on the time-frequency resources used by the reference signals of any antenna port belonging to the CDM group, there are only reference signals of different antenna ports using different frequency-domain coding. If the CDM group is indicated as being multiplexed in time domain and frequency domain (non-exclusive), the communication device can consider that, on the time-frequency resources used by the reference signals of any antenna port belonging to the CDM group, there are reference signals of different antenna ports using different time-domain coding and / or frequency-domain coding. For example, the configuration information may indicate that the DMRS belonging to CDM group #0 do not use the same time-domain and / or frequency-domain resources, if the antenna port of the DMRS scheduled by the current DCI is port #1, by searching the indication information of the CDM group #0 to which port #1 belongs, it is determined that the DMRS corresponding to port #1 exclusively occupies the time-frequency resource. For another example, the configuration information may indicate that the DMRS belonging to CDM group #0 multiplex the time-frequency resources by frequency-domain coding, if the antenna port of the DMRS scheduled by the current DCI is port #1, by searching the indication information of the CDM group #0 to which port #1 belongs, it is determined that the DMRS corresponding to port #1 multiplex the time-frequency resources by frequency-domain coding.

[0300] FIG. 13 is a flowchart of a method performed by a communication device according to an embodiment of the present disclosure.

[0301] According to the embodiment, the communication device may be any device capable of acquiring a signal and performing a signal detection. For example, the communication device may be a base station or a terminal, but is not limited thereto. Alternatively, for example, the communication device may even be a user equipment having a large scale of antenna ports, etc.

[0302] As shown in FIG. 13, in step S1310, a first message for transmitting a calibration signal for a neural network calibration is obtained. In step S1320, at least two calibration signals are transmitted based on the received first message.

[0303] In step S1310, the first message for transmitting the calibration signal for the neural network calibration is obtained. Here, the first message is consistent with that in step S710.

[0304] In some examples, the first message may be request information for the calibration signal for the neural network calibration. For example, the first message may be directly from received signaling. For example, the first message may be contained in DCI and / or uplink control information (UCI).

[0305] In some examples, the first message may be information on a relevant measurement result of a neural network. For example, the calibration signal required to be used for the neural network may be determined according to the received report information for the measurement result. For example, if the report information contains the numerical value of a noise reduction gain of the neural network, if the numerical value is lower than a threshold, it may be considered that the indication information for a requirement for the calibration signal is received.

[0306] For another situation of whether a calibration signal is required, reference may be made to the description of step S710.

[0307] In step S1320, the at least two calibration signals are transmitted based on the received first message. Here, for the association relationship between a calibration signal and a reference signal and / or the neural network, reference may be made to the description in step S720.

[0308] Alternatively, the step of transmitting / receiving configuration information on the calibration signal may further be included. For the transmitting / reception of the configuration information, reference may be made to the above description for the configuration information.

[0309] FIG. 14 is a flowchart of a method performed by a communication device according to an embodiment of the present disclosure.

[0310] According to the embodiment, the communication device may be any device capable of acquiring a signal and performing a signal detection. For example, the communication device may be a base station or a terminal, but is not limited thereto. Alternatively, for example, the communication device may even be a user equipment having a large scale of antenna ports, etc.

[0311] As shown in FIG. 14, in step S1410, a first measurement value related to first channel estimation information and / or a second measurement value related to second channel estimation information are obtained based on a received signal. The first channel estimation information is, for example, the first information obtained according to step S410 in FIG. 4. The second channel estimation information is, for example, the second information obtained according to step S420 in FIG. 4. In step S1420, a measurement value related to performance of a neural network is obtained based on the first measurement value and / or the second measurement value. Alternatively, in the situation where the communication device is a terminal, the measurement result of the performance of the neural network may further be reported based on the obtained measurement value in step S1430.

[0312] According to the above method, the communication system can monitor the operation state of the neural network in real time. If the performance of the neural network deviates from expectations, the system can switch to a traditional signal processing scheme and / or perform the foregoing neural network online training process in time. This scheme effectively avoids communication interruption of the system caused by the performance degradation of the neural network, and achieves the communication robustness in changing channel environments.

[0313] The contents involved in the above steps S1410-S1420 will be respectively described below in detail.

[0314] Referring to FIG. 14, in step S1410, the first measurement value related to the first channel estimation information and / or the second measurement value related to the second channel estimation information are obtained based on the received signal. According to an embodiment, this step may contain: extracting a reference signal from the received signal; obtaining first information based on the reference signal; obtaining second information through the neural network; and obtaining the first measurement value based on the first information, and / or obtaining the second measurement value based on the second information. In step S1420, a final measurement value or measurement result related to the performance of the neural network is obtained based on the first measurement value and / or the second measurement value.

[0315] Example 1: The first measurement value and the second measurement value are respectively defined as a signal-to-noise ratio obtained based on an input signal of the neural network and a signal-to-noise ratio obtained based on an output signal of the neural network. Here, the signal-to-noise ratio is defined as a division of the linear average value of the square of the channel amplitude (modulus) of each RE in the signal by the linear average value of the power contribution of the noise and / or interference. The signal-to-noise ratios obtained through the two measurements are combined to be used as the final measurement value.

[0316] Example 2: The first measurement value and the second measurement value are respectively defined as a noise intensity obtained based on an input signal of the neural network and a noise intensity obtained based on an output signal of the neural network. Here, the noise intensity is defined as the linear average value of the power contribution of the noise and / or interference of the signal. The difference between the noise intensities obtained through the two measurements is used as the final measurement value. Here, the noise intensity can be obtained by performing a differential calculation on adjacent REs.

[0317] In addition, it should be noted that the above operations are exemplary only, and the present disclosure is not limited thereto.

[0318] In some examples, the measurement value of the performance of the neural network may be obtained by measuring the noise reduction gain σ of the neural network.

[0319] In some examples, the measurement value of the performance of the neural network may be obtained by measuring a noise power estimation in the second information outputted by the neural network, e.g., a noise variance σ_2 in the second information.

[0320] In some examples, the measurement value of the performance of the neural network may be obtained by measuring the signal-to-noise ratios in the signals inputted and outputted by the neural network (i.e., the noise variances σ_1and σ_2 in the first information and the second information) and by calculating the quotient σ_1 / σ_2 of the noise variances.

[0321] In some examples, the above neural network performance measurement step is dependent on the reference signal periodically occurring in the received signal. For example, the above measurement process is dependent on a CSI-RS signal. Each time the received signal contains the CSI-RS, the terminal will perform the above measurement step once to obtain a corresponding measurement value.

[0322] In some examples, the above neural network performance measurement step is dependent on the reference signal occurring only at a specific transmission opportunity (e.g., a time slot, a subframe, and a frame). For example, the above measurement process is dependent on a DMRS signal. If a data transmission is performed, the received signal contains the DMRS. In this situation, the terminal will perform the above measurement step based on one or more DMRS signals, to obtain a corresponding measurement value.

[0323] In some examples, the measurement result related to the performance of the neural network may be further determined based on the obtained measurement value. For example, whether the obtained measurement value exceeds a threshold may be determined, or whether an update is required to be performed on the neural network may be determined based on the obtained measurement value.

[0324] In some examples, the execution of the above neural network performance measurement step is controlled based on certain special indications. As an example, only if the DCI signaling for scheduling this transmission indicates that the measurement for the performance of the neural network is required, the terminal performs the above measurement process. As another example, the DCI signaling may indicate that the terminal performs the above neural network performance measurement process at the transmission time of the last DMRS during this data transmission.

[0325] Alternatively, the above method may further include: reporting the measurement result of the performance of the neural network. Specifically, based on the measurement value obtained in step S1420, the terminal may transmit the report content on a preset time-frequency resource. The report content may include at least one of: a noise reduction gain of the neural network, a difference between the noise reduction gain of the neural network and a preset threshold, whether the noise reduction gain of the neural network exceeds the preset threshold, a signal-noise intensity and / or a signal-to-noise ratio of an output signal of the neural network, a signal-noise intensity and / or a signal-to-noise ratio of an input signal of the neural network, a difference between the signal-noise intensity and / or the signal-to-noise ratio of the output signal of the neural network and a preset threshold, whether the signal-noise intensity and / or the signal-to-noise ratio of the output signal of the neural network exceed the preset threshold, whether the neural network needs to be updated, and the like.

[0326] In some examples, the above step may be performed periodically. Specifically, according to this step, the performance of the neural network will be periodically measured, and the transmission is performed through a specific time-frequency resource. For example, the above measurement process is dependent on a periodic CSI-RS signal, and the measurement value is transmitted in a time slot following the above measurement process.

[0327] In some examples, the determination for the opportunity in the above step is based on certain specific indications. As an example, only if the DCI signaling for scheduling this transmission indicates that the measurement for the performance of the neural network needs to be reported, the terminal performs the complete measurement step and the reporting step. As another example, the terminal measures the performance of the neural network based on the DMRS in the received signal in each data transmission process, but performs the performance reporting process only after receiving the corresponding DCI signaling indication.

[0328] In some examples, the report content may be combined with other report content to be transmitted. The specific implementation may include: reporting the above report content as a special access request, using the above report content as a part of a CSI report, and inserting the above report content into a hybrid automatic repeat request (HARQ) message for data transmission. Several specific examples are given below.

[0329] Example 1: If the terminal is to transmit a positive update request, the terminal determines a sequence required to be transmitted, according to a hybrid automatic repeat request acknowledgment (HARQ-ACK) information bit, and performs a cyclic shift according to the sequence, where a cyclic shift value is a value corresponding to a positive update request configured by a higher layer parameter. Conversely, an other cyclic shift value is used. Here, the positive update request corresponds to a situation where the neural network of the terminal needs to be updated, the performance of the neural network does not meet a requirement, or the like.

[0330] Example 2: If the terminal is to transmit a positive update request, the terminal will transmit an HARQ-ACK information bit in a first format. Conversely, the terminal will transmit the HARQ-ACK information bit in a second format.

[0331] Example 3: If the terminal is to transmit a positive update request, the terminal will transmit an HARQ-ACK information bit using a first resource. Conversely, the terminal will transmit the HARQ-ACK information bit using a second resource.

[0332] Example 4: The terminal appends one or more bits representing a negative update request or a positive update request before and / or after a to-be-sent HARQ-ACK information bit (for example, the HARQ-ACK information bit refers to 1110, and one bit 1 (indicating a positive update request (i.e., an update is required)) may be added after the HARQ-ACK information bit, and accordingly, the complete feedback information is 11101). Then, the terminal will combine and transmit all bits. Here, each bit for transmitting an update request corresponds to one neural network or antenna port. If all bits are zero, it indicates that all update requests are negative.

[0333] Example 5: The terminal appends one or more bits representing a negative update request or a positive update request before and / or after a to-be-sent CSI report bit (for example, 4-bit information 0011 is appended after the CSI report, indicating that the neural networks corresponding to the antenna ports 0 and 1 do not need to be updated, and the neural networks corresponding to the antenna ports 1 and 2 need to be updated). Then, the terminal will combine and transmit all bits. Here, each bit for transmitting an update request corresponds to one neural network or antenna port. If all bits are zero, it indicates that all update requests are negative.

[0334] Example 6: If transmitting a CSI report, the terminal will respectively obtain corresponding CQI according to the input signal of the neural network and the output signal of the neural network, and combines the CQI as the content of the CSI report for transmission to the base station. Here, the CQI refers to channel quality information (CQI) in the CSI report in an NR system.

[0335] Alternatively, in some examples, the method performed by the communication device may further include acquiring relevant configuration information of the measurement reporting process. Specifically, the obtained relevant configuration information may include radio resource control signaling (RRC) and / or downlink control information (DCI). The RRC signaling received by the terminal may include at least one of: a reference signal used for the measurement, a measurement opportunity, report content, a time-frequency resource used, a transmission opportunity, a transmission mode, and the like. The DCI signaling received by the terminal may include a relevant indication for the execution of the performance measurement process and / or the reporting process by the terminal, i.e., a decision on whether to perform the performance measurement process and / or the reporting process.

[0336] In some examples, the reference signal used for the measurement may be, for example, a signal such as a CSI-RS and a DMRS, or may be a calibration signal or an other reference signal additionally transmitted.

[0337] In some examples, in the configuration, the physical resource used in the above reporting process are directly and / or indirectly indicated by parameters such as a frequency domain resource range, a time domain resource range and an antenna port. For example, in the configuration, the physical resource used in the reporting process is determined by an index provided in the RRC signaling or DCI signaling. A corresponding reporting resource set is searched through the index, thus obtaining the position of the physical resource.

[0338] In some examples, the terminal may be configured with a plurality of different measurement reporting processes, and the processes are performed independently of the other. For example, the terminal may be configured with two independent measurement reporting processes. In the first process, the neural network is measured based on a periodic CSI-RS signal, and the obtained noise reduction gain value is added to a periodic CSI report for transmission. In the second process, the neural network is measured based on a DMRS signal in a PDSCH, the output noise intensity of the neural network is measured a plurality of times by using the DMRS transmitted by the PDSCH, whether the neural network needs to perform a parameter update is determined based on the results of the plurality of measurements, and the determination is added to the HARQ feedback in this data transmission.

[0339] Alternatively, in some examples, the method performed by the communication device may further include reporting a measurement reporting related capability of the communication device with respect to the neural network. Specifically, the related capability includes at least one of the following information: a capability to measure the neural network with respect to the above reference signals, an uplink physical resource configuration capable of being supported, a measurement reporting approach capable of being supported, and the like.

[0340] Here, an exemplary embodiment is given. The capability information reported by the communication device may be, for example, one of: supporting only the use of the DMRS for the neural network performance measurement, not supporting a dual CQI feedback, and supporting HARQ feedback-based measurement reporting.

[0341] Fig. 15 shows a flowchart of a method executed by a communication device according to an embodiment of the present disclosure.

[0342] According to the embodiment, the communication device can be any device capable of acquiring signals and performing signal detection. For example, the communication device can be a base station or a terminal, but is not limited thereto. For instance, it can even be a user equipment with a large number of antenna ports, etc.

[0343] As shown in Fig. 15, at step S1510, at least one configuration parameter related to the neural network is obtained based on the configuration information and / or preset information. The configuration parameters related to the neural network include the structure of the neural network and the calculation coefficients required in each of its structures. At step S1520, the performance of at least one neural network is measured based on the received signal. At step S1530, based on the measurement results, the neural network to be used in subsequent processes is determined. The subsequent processes may include, but are not limited to, the aforementioned channel estimation process (Fig. 4), the neural network update process (Fig. 8), and the neural network performance measurement process (Fig. 14). Optionally, step S1530 may also include reporting the measurement results; determining the neural network to be used in subsequent processes based on the received indication information.

[0344] According to the method of the present disclosure, the communication system can relatively quickly switch the neural network used in subsequent signal processing processes, providing an initial neural network for the aforementioned processes, thus effectively avoiding the large amount of time consumed in training the neural network from scratch. For example, based on the received reference signal, the terminal first determines, from multiple pre-set neural networks in itself, a neural network for performing the aforementioned channel estimation process (Fig. 4) according to steps S1510-S1530. At the same time, it performs the neural network performance measurement process (Fig. 14) and reports the measurement results. Then, according to the subsequently received reference signal and / or calibration signal, it executes the neural network update process (Fig. 8) for this neural network.

[0345] Next, the contents involved in the above-mentioned steps S1510 to S1530 will be described in detail respectively.

[0346] Referring to Fig. 15, at step S1510, at least one configuration parameter related to the neural network is obtained based on the configuration information and / or preset information. The configuration parameters related to the neural network include the structure of the neural network and the calculation coefficients required in each of its structures. According to the embodiment, the neural network described in this step can be the neural network shown in Fig. 5, and at least one neural network can have the same and / or different structures and / or calculation coefficients.

[0347] In some examples, the terminal can obtain the structures of one or more neural networks and their corresponding calculation parameters based on the configuration information obtained through the RRC process. For example, the information in the configuration information can indicate an 8-layer convolutional neural network, including giving: the size of the core function of each convolutional layer and the corresponding calculation coefficients, the activation function, normalization parameters, pooling function, etc. between each layer.

[0348] In some examples, the terminal can obtain the structures of one or more neural networks and their corresponding calculation parameters based on the preset information. The preset information can be the information pre-stored if the device leaves the factory or the information preset by other means. For example, a 3-layer neural network is pre-set if the device leaves the factory, and the calculation coefficients and activation functions used by the neurons in each layer are pre-set.

[0349] In some examples, the terminal can obtain the structures of one or more neural networks and their corresponding calculation parameters from the preset information based on the indication information given in the configuration information. For example, if the device leaves the factory, the configuration parameters of multiple neural networks are pre-set, and according to the indication information in the configuration information, the corresponding neural network is selected.

[0350] At step S1520, based on the received signal, the performance of at least one neural network is measured. Specifically, the process of measuring the performance of each neural network is as described in steps S1410-1420.

[0351] At step S1530, based on the measurement results, the neural network to be used in subsequent processes is determined. The subsequent processes may include, but are not limited to, the aforementioned channel estimation process (Fig. 4), the neural network update process (Fig. 8), and the neural network performance measurement process (Fig. 14).

[0352] In some examples, the communication device determines on its own the neural network to be used in subsequent processes based on the obtained measurement results. For example, the communication device selects the neural network with the highest output signal-to-noise ratio (lowest noise power), or the communication device selects the neural network with the highest noise reduction gain, etc.

[0353] Optionally, step S1530 may also include reporting the measurement results and determining the neural network to be used in subsequent processes based on the received indication information. Specifically, the process of reporting the performance of each neural network is as described in step S1430.

[0354] In some examples, the performance measurement reports of multiple neural networks can be carried out together, or the performance measurement and / or performance reporting process can be carried out separately for each neural network.

[0355] In some examples, the received indication information may contain the index information of the neural network. The communication device can select the corresponding neural network for use according to this index information. For example, the index information corresponds one-to-one with each neural network used for performance measurement and reporting in step S1520, and the communication device selects the corresponding neural network based on the index information for subsequent processes.

[0356] In some examples, the received indication information may also contain the relevant configuration information used for transmitting the reference signal and / or calibration signal, and the neural network is determined based on this configuration information. For example, the received indication information gives the antenna port information used by the calibration signal, and the communication device selects the associated neural network based on this antenna port to execute subsequent processes. For the specific process, refer to the relevant content of step S820.

[0357] FIG. 16 illustrates an exemplary structure of each node of a communication device applicable to the present disclosure. The exemplary node shown in FIG. 16 includes a transceiver 1610 and a processor 1620 coupled to the transceiver 1610. The transceiver 1610 is configured to transmit and receive a signal. The processor 1620 is configured to perform the method described in the present disclosure. The present disclosure may alternatively be implemented as a computer storage medium. The computer storage medium stores a computer executable instruction. If the stored computer executable instruction is executed by the processor, the processor performs the method described in the present disclosure.

[0358] The various illustrative logical blocks, modules, and circuits described in the present disclosure may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, a discrete gate or transistor logic, a discrete hardware component, or any combination thereof designed to perform the functions described herein. The general purpose processor may be a microprocessor, but in an alternative scheme, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may alternatively be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in collaboration with a DSP core, or any other such configuration.

[0359] The steps of the method or algorithm described in the present disclosure may be embodied directly in hardware, in a software module executed by the processor, or in a combination of the two. The software module may reside in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, to enable the processor to read information from / write information to the storage medium. In an alternative scheme, the storage medium may be integrated to the processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative scheme, the processor and the storage medium may reside as discrete components in a user terminal.

[0360] In one or more exemplary designs, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in the software, the functions may be stored on or transmitted over a computer readable medium as one or more instructions or codes. The computer readable medium includes both a computer storage medium and a communication medium, the communication medium including any medium that facilitates the transfer of a computer program from one place to another. The storage medium may be any available medium that can be accessed by a general purpose or special purpose computer.

[0361] Example methods and apparatuses are described in combination with the accompanying drawings in the description set forth herein, and do not represent all the examples that may be implemented or that are within the scope of the claims. The term “example” used herein means “serving as an example, instance, or illustration,” rather than “preferred” or “advantageous over other examples.” The detailed description includes specific details for the purpose of providing an understanding of the described techniques. These techniques, however, may be practiced without these specific details. In some cases, well-known structures and devices are shown in the form of a block diagram in order to avoid obscuring the concepts of the described examples.

[0362] This specification contains many specific implementation details, but the implementation details should not be construed as a limitation to the scope of any disclosure or the scope claimed, but rather as a description for specific features in a specific embodiment of the specific disclosure. Certain features described in the context of separate embodiments in this specification may alternatively be implemented in combination in a single embodiment. Rather, the various features described in the context of a single embodiment may be implemented separately in a plurality of embodiments or implemented in any suitable sub-combination. Furthermore, the features may be described as functioning in certain combinations in the context, and even initially so claimed, but in some cases one or more features in a claimed combination may be deleted from the combination, and the claimed combination may be directed to a sub-combination or the variation of the sub-combination.

[0363] It should be understood that the specific order or hierarchy of steps in the method in the present disclosure is an illustration for an exemplary process. Based on design preferences, it may be understood that the specific order or hierarchy of the steps in the method may be rearranged to achieve the functions and effects disclosed in the present disclosure. The accompanying method claims present the elements of various steps in example order, but are not intended to be limited to the specific order or hierarchy presented, unless specifically stated otherwise. Furthermore, although an element may be described or claimed in a singular form, the plural can also be expected unless the limitation to the singular is explicitly stated. Thus, the present disclosure is not limited to the examples shown, and any apparatus for performing the functions described herein is included in the aspects of the present disclosure.

[0364] It can be understood that “at least one” described in the present disclosure includes any and / or all possible combinations of the listed items, the various embodiments described in the present disclosure and the various examples in the embodiments may be varied and combined in any suitable form, and “ / ” described in the present disclosure represents “and / or.”

[0365] The text and drawings are provided as examples only to help readers understand the present disclosure. The text and drawings are not intended to limit and should not be interpreted as limiting the scope of the present disclosure in any way. Although certain embodiments and examples have been provided, based on the content disclosed herein, it is obvious to those skilled in the art that modifications can be made to the illustrated embodiments and examples without departing from the scope of the present disclosure.

Claims

A method performed by a first communication device in a wireless communication system, the method comprising:determining based on a neural network related measurement, whether to transmit a first message for requesting a signal for a neural network calibration; andif the first message is transmitted, receiving at least two signals for the neural network calibration, wherein the at least two signals for the neural network calibration are transmitted through one antenna port, with time domain resources and frequency domain resources that carry the at least two signals for the neural network calibration satisfying a first condition.The method of claim 1, wherein it is determined that the first message is to be transmitted, if the neural network related measurement satisfies at least one of:a measurement result of the neural network being greater than a first threshold;the measurement result of the neural network being less than a second threshold; anda neural network-based channel estimation or subsequent data reception by the first communication device being failed.The method of claim 1, wherein the first message comprises at least one of request information for the signal for the neural network calibration or information on a result of the neural network related measurement,wherein the at least two signals for the neural network calibration comprise a configured reference signal,wherein an antenna port for transmission of a remaining signal in the at least two signals for the neural network calibration is determined based on an antenna port of the configured reference signal, andwherein time domain resources and frequency domain resources of the remaining signal and the configured reference signal satisfy the first condition.The method of claims 1, wherein the first condition comprises at least one of:a difference value between a maximum value and a minimum value in indices of time domain resources of the at least two signals for the neural network calibration being less than a third threshold;a difference value between a maximum value and a minimum value in indices of frequency domain resources of the at least two signals for the neural network calibration being less than a fourth threshold;an index of a time domain resource of a first signal in the at least two signals for the neural network calibration being a first value, and an interval between a time domain resource of a second signal in the at least two signals for the neural network calibration and the time domain resource of the first signal being a predefined second value; andan index of a frequency domain resource of the first signal in the at least two signals for the neural network calibration being a predefined third value, and an interval between a frequency domain resource of the second signal in the at least two signals for the neural network calibration and the frequency domain resource of the first signal being a predefined fourth value.The method of claim 1, further comprising:receiving configuration information,wherein the configuration information comprises at least one of:mapping information between the signal for the neural network calibration and the neural network,information on the time domain resources and the frequency domain resources for the at least two signals for the neural network calibration,information for determining an antenna port for transmission of the signal for the neural network calibration, andinformation on a sequence for generating the signal for the neural network calibration.The method of claim 3, wherein the request information for the signal for the neural network calibration comprises:information on an antenna port number associated with the signal;an identifier (ID) of the neural network; orinformation on an antenna port number associated with the configured reference signal.The method according to claim 2,wherein the first message comprises information on a first measurement result and indication information for requesting the signal for the neural network calibration,wherein the first message comprises information on a second measurement result and the indication information for requesting the signal for the neural network calibration,wherein the first message comprises the information on the first measurement result and the information on the second measurement result, orwherein the first measurement result is obtained based on a reference signal for a channel estimation, and the second measurement result is a channel measurement result obtained based on the first measurement result and the neural network.A method performed by a second communication device in a wireless communication system, the method comprising:receiving a first message for requesting a signal for a neural network calibration, wherein a transmission of the first message is determined by a first communication device based on a neural network related measurement; andtransmitting at least two signals for the neural network calibration, wherein the at least two signals for the neural network calibration are transmitted through one antenna port, with time domain resources and frequency domain resources that carry the at least two signals for the neural network calibration satisfying a first condition.A method performed by a communication device in a wireless communication system, the method comprising:performing a channel estimation based on at least two signals for a neural network calibration, to obtain a first estimation result and a second estimation result of a first channel estimation;obtaining a first estimation result of a second channel estimation through a neural network based on the first estimation result of the first channel estimation;determining a first error based on the second estimation result of the first channel estimation and the first estimation result of the second channel estimation; andupdating the neural network based on the first error,wherein the signals for the neural network calibration are transmitted through one antenna port with time domain resources and frequency domain resources that carry the signals for the neural network calibration satisfying a first condition.The method of claim 9, further comprising:obtaining a second estimation result of the second channel estimation through the neural network based on the second estimation result of the first channel estimation;determining a second error based on the first estimation result of the first channel estimation and the second estimation result of the second channel estimation; andupdating the neural network based on the first error and the second error.The method of claim 9, wherein the updating the neural network based on the first error and the second error comprises:updating the neural network based on the first error, and then re-updating the updated neural network based on the second error;updating the neural network based on the second error, and then re-updating the updated neural network based on the first error; ordetermining a total error based on the first error and the second error, and updating the neural network based on the total error.The method of claim 9, further comprising:obtaining, based on a first received signal comprising at least one reference signal, first channel estimation information on a resource element (RE) occupied by the at least one reference signal;obtaining second channel estimation information corresponding to the first channel estimation information through the updated neural network based on the first channel estimation information; anddetermining channel state information on all REs occupied by the first received signal based on the second channel estimation information,wherein the signals for the neural network calibration comprise a configured reference signal,wherein an antenna port for transmission of a remaining signal in the at least two signals is determined based on the antenna port of the configured reference signal, andwherein the time domain resources and the frequency domain resources of the remaining signal and the configured reference signal satisfy the first condition.A first communication device in a wireless communication system, the first communication device comprising:a transceiver; andat least one processor coupled with the transceiver and configured to:determine based on a neural network related measurement, whether to transmit a first message for requesting a signal for a neural network calibration, andif the first message is transmitted, receive at least two signals for the neural network calibration, wherein the at least two signals for the neural network calibration are transmitted through one antenna port, with time domain resources and frequency domain resources that carry the at least two signals for the neural network calibration satisfying a first condition.A second communication device in a wireless communication system, the second communication device comprising:a transceiver; andat least one processor coupled with the transceiver and configured to:receive a first message for requesting a signal for a neural network calibration, wherein a transmission of the first message is determined by a first communication device based on a neural network related measurement, andtransmit at least two signals for the neural network calibration, wherein the at least two signals for the neural network calibration are transmitted through one antenna port, with time domain resources and frequency domain resources that carry the at least two signals for the neural network calibration satisfying a first condition.A communication device in a wireless communication system, the communication device comprising:a transceiver; andat least one processor coupled with the transceiver and configured to:perform a channel estimation based on at least two signals for a neural network calibration, to obtain a first estimation result and a second estimation result of a first channel estimation,obtain a first estimation result of a second channel estimation through a neural network based on the first estimation result of the first channel estimation,determine a first error based on the second estimation result of the first channel estimation and the first estimation result of the second channel estimation, andupdate the neural network based on the first error,wherein the signals for the neural network calibration are transmitted through one antenna port with time domain resources and frequency domain resources that carry the signals for the neural network calibration satisfying a first condition.

Images