Communication method, apparatus and system
By performing overall performance monitoring of multiple AI functions deployed in the wireless air interface, the problem of high signaling overhead was solved, efficient and flexible monitoring resource management was achieved, and monitoring overhead was reduced.
Patent Information
- Application Number
- PCT/CN2026/094364
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-09
- Filing Date
- 2026-04-30
- Publication Date
- 2026-11-12
Smart Images

Figure CN2026094364_12112026_PF_FP_ABST
Abstract
Description
Communication methods, devices and systems
[0001] This application claims priority to Chinese Patent Application No. 202510604456.7, filed on May 9, 2025, with the Chinese National Intellectual Property Administration, entitled “Communication Method, Apparatus and System”, the entire contents of which are incorporated herein by reference. Technical Field
[0002] This application relates to the field of communications, and more particularly to a communication method, apparatus and system. Background Technology
[0003] With the development of artificial intelligence (AI) and machine learning (ML), AI / ML technologies have been introduced into wireless air interfaces. For example, AI models can be applied to transmitters for data transmission processing, such as encoding, modulation, and precoding; they can also be applied to receivers for channel estimation and data reception processing, such as decoding, demodulation, and equalization; furthermore, AI models can be used at both the transmitting and receiving ends for channel compression and reconstruction; and still others, they can be used for beam management. To achieve better performance, performance monitoring of the AI models is necessary.
[0004] As AI / ML technologies become increasingly used in wireless air interfaces, the number of models (or functions) is also growing. Therefore, multiple models can be deployed in transmitters and / or receivers to implement different functionalities. When management nodes monitor the performance of these models, they can perform performance monitoring on a per-model basis, such as configuring resources for monitoring and reporting. This can potentially result in significant signaling overhead. Summary of the Invention
[0005] This application provides a communication method, apparatus, and system to reduce the overhead of model monitoring.
[0006] Firstly, a communication method is provided, which can be applied to a first device, which can be a terminal side. The terminal side can be a terminal-side device, or a chip or circuit of the terminal-side device. The terminal-side device can be replaced by a terminal device-side device, which can include at least one of a terminal device or an AI entity on the terminal device side. The AI entity on the terminal device side can be the terminal device itself, or an AI entity serving the terminal device, such as a server, like an over-the-top (OTT) server or a cloud server. Components deployed in the terminal device can be, for example, circuits or chips inside the terminal device (such as modem chips, also known as baseband chips, or system-on-chip (SoC) chips or system-in-package (SIP) chips containing modem cores, or chips for AI or SoC or SIP chips containing chips for AI, etc.); or, they can be applied to logic modules or software capable of implementing all or part of the functions of the first device, etc.
[0007] For example, the method includes: receiving first information, the first information being used to indicate a first monitoring resource configured for N AI functions, where N is a positive integer greater than 1; and sending or receiving a first signal based on the first monitoring resource, the first signal being used to obtain a first monitoring result for performance monitoring of the N AI functions as a whole.
[0008] Based on the above scheme, since N AI functions can be monitored as a whole, the first information can be used to configure the first monitoring resources for the performance monitoring of these N AI functions as a whole. If the performance monitoring of these N AI functions as a whole meets the performance requirements, then the subsequent configuration for performance monitoring of individual AI functions can be eliminated, and it is also unnecessary to configure monitoring resources for each AI function. This reduces signaling overhead and monitoring overhead, which is beneficial for monitoring the overall performance of N AI functions.
[0009] In conjunction with the first aspect, in some possible implementations of the first aspect, the first signal is used as input when N AI functions are treated as a whole, and the first monitoring result is determined based on the output when N AI functions are treated as a whole.
[0010] In conjunction with the first aspect, in some possible implementations of the first aspect, N AI functions are deployed on the terminal side, and the method also includes: determining the first monitoring result.
[0011] Optionally, the method further includes: sending the first monitoring result if the first monitoring result meets the first condition.
[0012] The first device can send the first monitoring result to the second device so that the second device can understand the performance monitoring of the N AI functions and make further reasonable decisions.
[0013] In conjunction with the first aspect, in some possible implementations of the first aspect, the method further includes: receiving second information, the second information being used to indicate one or more of the following: a first performance metric, a first condition, or a first resource; the first performance metric being a performance metric on which performance monitoring of N AI functions as a whole is based, and the first resource being a resource used to transmit the first monitoring result.
[0014] By indicating a first performance indicator to the first device, the first device can determine a first monitoring result based on this indicator. By indicating a first condition to the first device, the first device can determine whether to send the first monitoring result based on this condition. By indicating a first resource to the first device, the first device can send the first monitoring result on the corresponding resource, thereby facilitating the second device to receive it on the same resource.
[0015] In another implementation, the first performance metric may be determined by the first device or may be predefined.
[0016] In another implementation, the first condition may be determined by the first device or it may be predefined.
[0017] Therefore, the second device does not need to indicate the first performance index and / or the first condition through the second information, which can reduce the overhead of signaling indication.
[0018] In conjunction with the first aspect, in some possible implementations of the first aspect, the method further includes: receiving third information, which is used to indicate N AI functions.
[0019] By using third-party information to indicate the N AI functions to be monitored, the first device can determine which AI functions to monitor for performance. The second device can flexibly adjust the N AI functions to be monitored, making the monitoring of AI functions more flexible.
[0020] In another implementation, the N AI functions can be determined by the first device or they can be predefined.
[0021] Therefore, the second device does not need to use third information to instruct N AI functions, which can reduce the overhead of signaling instructions.
[0022] In conjunction with the first aspect, in some possible implementations of the first aspect, the method further includes: receiving fourth information, the fourth information being used to indicate M AI functions out of N AI functions, and a second monitoring resource configured for the M AI functions, where M is a positive integer less than N; and sending or receiving a second signal based on the second monitoring resource, the second signal being used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole.
[0023] Based on the above method, the first device can progressively monitor the performance of N AI functions and each subset of the N AI functions. If the performance monitoring of the N AI functions or one or more subsets meets the performance requirements, it is not necessary to monitor the performance of each AI function individually. Compared with monitoring the performance of each AI function individually among the N AI functions, the monitoring overhead can be reduced.
[0024] In conjunction with the first aspect, in some possible implementations of the first aspect, the second signal is used as the input when the M AI functions are treated as a whole, and the second monitoring result is determined based on the output when the M AI functions are treated as a whole.
[0025] In conjunction with the first aspect, in some possible implementations of the first aspect, the method further includes: receiving fifth information, the fifth information indicating one or more of the following: a second performance indicator, a second condition, a second resource, or a triggering condition; the second performance indicator is the performance indicator on which performance monitoring of the M AI functions as a whole is based; the second condition is the condition satisfied for sending the second monitoring result; the second resource is the resource used to transmit the second monitoring result; and the triggering condition is the condition for triggering performance monitoring of the M AI functions as a whole.
[0026] By indicating a trigger condition to the first device, the first device can determine whether to perform performance monitoring on the M AI functions. By indicating a second performance indicator to the first device, the first device can determine the performance indicator on which the performance monitoring of the M AI functions is based. By indicating a second condition to the first device, the first device can determine whether to send a second monitoring result based on this condition. By indicating a second resource to the first device, the first device can send the second monitoring result on the corresponding resource, thereby facilitating the second device to receive it on the same resource.
[0027] Secondly, a communication method is provided, which can be applied to a second device, which can be a network side. The network side can be a network-side device, or a chip or circuit of a network-side device. The network-side device can be replaced by a device on the network device side, which can include at least one of a network device or an AI entity on the network device side. The AI entity on the network device side can be the network device itself or an AI entity serving the network device, such as an intelligent network element on the network device side. Components deployed in the network device can be, for example, circuits or chips inside the network device (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core, or a chip for AI or a SoC or SIP chip containing a chip for AI, etc.); or, they can be applied to logic modules or software capable of implementing all or part of the functions of the second device, etc.
[0028] It should be understood that the technical solution of the second aspect corresponds to the technical solution of the first aspect. Therefore, the relevant descriptions and technical effects of the first aspect are equally applicable to the second aspect. Thus, the content that is the same as or corresponding to the first aspect can be referred to the relevant descriptions of the first aspect, and will not be repeated here.
[0029] For example, the method includes: sending first information, the first information being used to indicate a first monitoring resource configured for N AI functions, where N is a positive integer greater than 1; and sending a first signal on the first monitoring resource, the first signal being used to obtain a first monitoring result for performance monitoring of the N AI functions as a whole.
[0030] In conjunction with the second aspect, in some possible implementations of the second aspect, the first signal is used as the input when the N AI functions are treated as a whole, and the first monitoring result is determined based on the output when the N AI functions are treated as a whole.
[0031] In conjunction with the second aspect, in some possible implementations of the second aspect, the method further includes: receiving a first monitoring result.
[0032] In conjunction with the second aspect, in some possible implementations of the second aspect, the method further includes: sending second information, the second information being used to indicate one or more of the following: a first performance metric, a first condition, or a first resource; the first performance metric being a performance metric on which performance monitoring of N AI functions as a whole is based, the first condition being a condition satisfied for transmitting the first monitoring result, and the first resource being a resource used for transmitting the first monitoring result.
[0033] In conjunction with the second aspect, in some possible implementations of the second aspect, the method further includes: sending a third message, which is used to indicate N AI functions.
[0034] In conjunction with the second aspect, in some possible implementations of the second aspect, some of the N AI functions are deployed on the network side, and the method also includes: determining the first monitoring result.
[0035] In conjunction with the second aspect, in some possible implementations of the second aspect, the method further includes: sending fourth information, the fourth information being used to indicate M AI functions out of N AI functions, and a second monitoring resource configured for the M AI functions, the second monitoring resource being used to carry a second signal; sending or receiving the second signal based on the second monitoring resource, the second signal being used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole, where M is a positive integer less than N.
[0036] In conjunction with the second aspect, in some possible implementations of the second aspect, the second signal is used as the input when the M AI functions are treated as a whole, and the second monitoring result is determined based on the output when the M AI functions are treated as a whole, where M is a positive integer less than N.
[0037] In conjunction with the second aspect, in some possible implementations of the second aspect, before sending the fourth information, the method further includes: if the first monitoring result does not meet the performance requirements, determining to perform performance monitoring on M of the N AI functions.
[0038] In conjunction with the second aspect, in some possible implementations of the second aspect, the method further includes: sending fifth information, the fifth information indicating one or more of the following: a second performance metric, a second condition, a second resource, or a triggering condition; the second performance metric is the performance metric on which performance monitoring of the M AI functions as a whole is based; the second condition is the condition satisfied for transmitting the second monitoring result; the second resource is the resource used for transmitting the second monitoring result; and the triggering condition is the condition satisfied for triggering performance monitoring of the M AI functions as a whole.
[0039] The following example illustrates performance monitoring of N AI functions.
[0040] Optionally, the N AI functions include: a first AI function for channel estimation and a second AI function for channel prediction; the first monitoring resource is used to transmit a first reference signal and a second reference signal, wherein the density of the first reference signal on the resource is lower than the density of the second reference signal on the resource; the first reference signal is the input when the N AI functions are used as a whole, and the output when the N AI functions are used as a whole is the predicted channel information; the first monitoring result is determined based on the predicted channel information and the channel information determined by the second reference signal, or the first monitoring result is determined based on the predicted received signal of the second reference signal and the received signal of the second reference signal, wherein the predicted received signal of the second reference signal is determined based on the predicted channel information.
[0041] Optionally, the N AI functions include: a first AI function for channel prediction and a second AI function for channel compression and reconstruction; the first monitoring resource is used to transmit a first reference signal and a second reference signal; the first reference signal is the input when the N AI functions are used as a whole, and the output when the N AI functions are used as a whole is the reconstructed channel information; the first monitoring result is determined based on the reconstructed channel information and the channel information determined by the second reference signal.
[0042] Optionally, the N AI functions include: a first AI function for channel estimation and a second AI function for demodulation; the first monitoring resource is used to transmit a first reference signal, the first reference signal being the input of the N AI functions as a whole; the output of the N AI functions as a whole is at least one estimated demodulated bit or at least one likelihood log-likelihood ratio corresponding to the at least one demodulated bit, the first monitoring result being determined based on the estimated bit and tag data, the estimated bit being obtained by decoding the at least one demodulated bit or determined according to the at least one log-likelihood ratio, and the tag data being known bits.
[0043] Optionally, the N AI functions may also include a third AI function for balancing.
[0044] Optionally, the N AI functions include: a first AI function for channel estimation and a second AI function for decoding; the first monitoring resource is used to transmit a first reference signal, the first reference signal being the input of the N AI functions as a whole; the first monitoring result is determined based on estimated decoded bits and tag data, the tag data being a known bit sequence, the estimated bits being the output of the N AI functions as a whole, or the estimated bits being determined based on the probability of the bit values output by the N AI functions as a whole.
[0045] Optionally, the N AI functions may further include a third AI function for equalization and / or a fourth AI function for demodulation.
[0046] Thirdly, an apparatus is provided. This apparatus may include functional modules corresponding to each method / operation / step / action described in the first aspect and any possible implementation thereof, or may include functional modules corresponding to each method / operation / step / action described in the second aspect and any possible implementation thereof. The functional modules may be hardware circuits, software, or a combination of hardware circuits and software.
[0047] In one design, the device may include a processing module and a communication module. The communication module is used to perform the sending and receiving actions performed by the first device as described in the first aspect and any possible implementation thereof, while the processing module is used to perform the processing-related actions performed by the first device in the methods described in the first aspect and any possible implementation thereof.
[0048] In one design, the device can be a terminal device, or a device, module, circuit, or chip configured in the terminal device.
[0049] In one design, the device may include a processing module and a communication module. The communication module is used to perform the sending and receiving actions performed by the second device as described in the second aspect and any possible implementation thereof, while the processing module is used to perform the processing-related actions performed by the second device in the methods described in the second aspect and any possible implementation thereof.
[0050] In one design, the device can be a network device, or a device, module, circuit, or chip configured in the network device.
[0051] Fourthly, an apparatus is provided, including a processor and a storage medium storing instructions that, when executed by the processor, cause the method as described in the first aspect and any possible implementation thereof to be implemented, or cause the method as described in the second aspect and any possible implementation thereof to be implemented.
[0052] Fifthly, an apparatus is provided, including processing circuitry for processing data and / or information to enable the implementation of methods as described in the first aspect and any possible implementation thereof, or to enable the implementation of methods as described in the second aspect and any possible implementation thereof.
[0053] The processing circuit may include one or more processors, or all or part of the circuitry in one or more processors used for control or processing functions.
[0054] Optionally, the apparatus may further include a memory for storing programs or instructions, and a processor for running the programs or instructions such that the methods of any of the first to fourth aspects, and any possible implementation of any of the aspects, are implemented.
[0055] Optionally, the device may also include transceiver circuitry, or input / output interfaces.
[0056] In a sixth aspect, a chip is provided, including processing circuitry for running a program or instructions to implement a method as described in the first aspect and any possible implementation thereof, or to implement a method as described in the second aspect and any possible implementation thereof.
[0057] Optionally, the chip may also include a memory for storing programs or instructions.
[0058] Optionally, the chip may also include transceiver circuitry, or input / output interfaces.
[0059] A seventh aspect provides a computer-readable storage medium comprising instructions that, when executed by a processor, cause the method as described in the first aspect and any possible implementation thereof to be implemented, or cause the method as described in the second aspect and any possible implementation thereof to be implemented.
[0060] Eighthly, a computer program product is provided, the computer program product including computer program code or instructions, which, when executed, cause the method as described in the first aspect and any possible implementation thereof to be implemented, or cause the method as described in the second aspect and any possible implementation thereof to be implemented.
[0061] Ninth aspect, a communication system is provided, the communication system comprising a first device and a second device.
[0062] Optionally, the first device is used to perform the method in the first aspect and any possible implementation of the first aspect, and the second device is used to perform the method in the second aspect and any possible implementation of the second aspect.
[0063] It should be understood that the third to ninth aspects of this application correspond to the technical solutions of the first to second aspects of this application, and the beneficial effects obtained by each aspect and the corresponding feasible implementation are similar, and will not be repeated here. Attached Figure Description
[0064] Figure 1 is a schematic diagram of a communication system applicable to the communication method of this application embodiment;
[0065] Figure 2 is a schematic diagram of another communication system applicable to the communication method of this application embodiment;
[0066] Figure 3 is a schematic diagram of a possible application framework in a communication system;
[0067] Figure 4 is a schematic diagram of another possible application framework in a communication system;
[0068] Figure 5 is a schematic diagram of the AI / ML functional framework for wireless air interface;
[0069] Figure 6 is a schematic diagram of the AI model used to implement channel prediction function;
[0070] Figure 7 is a schematic diagram of the AI model used to implement channel compression and reconstruction (or reconstruction) functions;
[0071] Figure 8 is a schematic diagram of the AI model used to implement the channel estimation function;
[0072] Figure 9 is a schematic diagram of the AI model used to implement the multi-input multi-output (MIMO) equalization function;
[0073] Figure 10 is a schematic diagram of the AI model used to implement the demodulation function;
[0074] Figure 11 is a schematic diagram of the AI model used to implement the decoding function;
[0075] Figure 12 is a wireless communication system including a transceiver based on a neural network provided in an embodiment of this application;
[0076] Figure 13 is an example of a transceiver based on a neural network provided in an embodiment of this application;
[0077] Figure 14 is another example of a transceiver based on a neural network provided in the embodiments of this application;
[0078] Figure 15 is yet another example of a transceiver based on a neural network provided in the embodiments of this application;
[0079] Figure 16 is a schematic diagram of the training process of the intelligent air interface provided in the embodiment of this application;
[0080] Figure 17 is a schematic diagram of performance monitoring of the model;
[0081] Figure 18 is a schematic diagram of the communication method provided in an embodiment of this application;
[0082] Figure 19 is a schematic diagram of sequential monitoring provided in an embodiment of this application;
[0083] Figure 20 is a schematic diagram of performance monitoring of AI functions deployed on the terminal side;
[0084] Figure 21 is another schematic diagram of performance monitoring of AI functions deployed on the terminal side;
[0085] Figure 22 is another schematic diagram of performance monitoring of AI functions deployed on the terminal side;
[0086] Figure 23 is a schematic diagram of the communication device provided in an embodiment of this application;
[0087] Figure 24 is a schematic diagram of the structure of a communication device provided in another embodiment of this application;
[0088] Figure 25 is a schematic diagram of the structure of the AI processor provided in an embodiment of this application. Detailed Implementation
[0089] The technical solutions in this application will now be described with reference to the accompanying drawings.
[0090] To facilitate understanding of the embodiments of this application, the following points will be explained first:
[0091] First, in this application, the indication includes explicit indication (also known as direct indication) and implicit indication (also known as indirect indication). Explicit indication information A means including information A; implicit indication information A means indicating information A through the correspondence between information A and information B, and direct indication information B. The correspondence between information A and information B can be predefined, pre-stored, pre-burned, or pre-configured; or it can refer to indicating information A through information B and preset rules.
[0092] Second, in this application, information C is used to determine information D, which includes both determining information D based solely on information C and determining it based on information C and other information. Furthermore, information C can also be used to determine information D indirectly, for example, in the case where information D is determined based on information E, and information E is determined based on information C.
[0093] Third, in this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the preceding and following related objects, but it does not exclude the possibility of indicating an "and" relationship; the specific meaning can be understood in context. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c; a and b; a and c; b and c; or a and b and c. Here, a, b, and c can be single or multiple.
[0094] Fourth, in this application, the use of prefixes such as "first" and "second" is merely for the purpose of distinguishing and describing different things belonging to the same category, and does not constrain the order, size, or quantity of things. For example, "first information" and "second information" are simply different pieces of information, and there is no temporal sequence, size, or priority relationship between them.
[0095] Fifth, in this application, "send" and "receive" indicate the direction of signal transmission. For example, "send information to the first device" can be understood as the destination of the information being the first device, which may include direct transmission via the air interface or indirect transmission via the air interface by other units or modules. "Receive information from the second device" can be understood as the source of the information being the second device, which may include direct reception from the second device via the air interface or indirect reception from the second device via the air interface by other units or modules. "Send" can also be understood as the "output" of the chip interface, and "receive" can also be understood as the "input" of the chip interface.
[0096] In other words, sending and receiving can occur between devices, such as between terminal devices and network devices; or they can occur within a device, such as between components, modules, chips, software modules, or hardware modules within a device via a bus, wiring, or interface.
[0097] Sixth, in the embodiments of this application, "when," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a time, nor do they require the device to make a judgment action when it is implemented, nor do they mean that there are other limitations.
[0098] Seventh, in this application, the words "example," "exemplarily," "for example," or "such as" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "example," "exemplarily," "for example," or "such as" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "example," "exemplarily," "for example," or "such as" is intended to present the relevant concepts in a specific manner.
[0099] The technical solutions provided in this application can be applied to various communication systems, such as 5th generation (5G) or new radio (NR) systems, frequency division duplex (FDD) systems, time division duplex (TDD) systems, wireless local area network (WLAN) systems, satellite communication systems, future communication systems, or integrated systems of multiple systems. The technical solutions provided in this application can also be applied to device-to-device (D2D) communication, vehicle-to-everything (V2X) communication, machine-to-machine (M2M) communication, machine-type communication (MTC), and Internet of Things (IoT) communication systems or other communication systems.
[0100] In a communication system, a device can send signals to or receive signals from another device. These signals can include information, signaling, or data. The device can also be replaced by an entity, network entity, communication equipment, communication module, node, communication node, etc.; this disclosure uses a device as an example. For instance, a communication system can include at least one terminal device and at least one network device. The network device can send downlink signals to the terminal device, and / or the terminal device can send uplink signals to the network device. It is understood that the terminal device in this disclosure can be replaced by a first communication device, and the network device can be replaced by a second communication device, both performing the corresponding communication methods described in this disclosure.
[0101] In wireless communication networks, such as mobile communication networks, the services supported by the networks are becoming increasingly diverse, thus requiring increasingly diverse demands. For example, networks need to support ultra-high speeds, ultra-low latency, and / or massive connectivity. This characteristic makes network planning, network configuration, and / or resource scheduling increasingly complex. Furthermore, as network functions become more powerful, such as supporting higher spectrum levels, supporting higher-order multiple-input multiple-output (MIMO) technologies, supporting beamforming (BF), and / or supporting beam management, network energy efficiency has become a hot research topic. These new demands, new scenarios, and new characteristics bring unprecedented challenges to network planning, operation, and efficient operation. To meet these challenges, artificial intelligence (AI) technology can be introduced into wireless communication networks to achieve network intelligence. To support AI technology in wireless networks, AI nodes may also be introduced. AI nodes can be the AI network elements shown in Figure 2 below or the AI modules shown in Figure 3.
[0102] Figure 1 is a schematic diagram of a communication system applicable to the communication method of this application embodiment. As shown in Figure 1, the communication system 100A may include at least one access network device, such as access network device 110 shown in Figure 1; the communication system 100A may also include at least one terminal device, such as terminal device 120 and terminal device 130 shown in Figure 1. Access network device 110 and terminal devices (such as terminal device 120 and terminal device 130) can communicate via a wireless link. The communication devices in this communication system, for example, access network device 110 and terminal device 120, can communicate via multi-antenna technology.
[0103] Figure 2 is a schematic diagram of another communication system applicable to the communication method of this application embodiment. Compared with the communication system 100A shown in Figure 1, the communication system 100B shown in Figure 2 further includes an AI network element 140. The AI network element 140 is used to perform AI-related operations, such as building datasets or AI models. The AI network element can also be simply referred to as an intelligent network element. In this disclosure, the AI model can be simply referred to as a model.
[0104] In one possible implementation, access network device 110 can send data related to the training of the AI model to AI network element 140, whereby AI network element 140 constructs a dataset and trains the AI model. For example, the data related to the training of the AI model may include data reported by terminal devices. AI network element 140 can send the results of operations related to the AI model to access network device 110, and then forward them to terminal devices via access network device 110. For example, the results of operations related to the AI model may include at least one of the following: a trained AI model, model evaluation results, or test results, etc. Exemplarily, a portion of the trained AI model may be deployed on access network device 110, and another portion on terminal devices 120 and / or 130. Alternatively, the trained AI model may be deployed on access network device 110. Or, the trained AI model may be deployed on terminal devices 120 and / or 130.
[0105] It should be understood that Figure 2 is only used as an example of the AI network element 140 being directly connected to the access network device 110. In other scenarios, the AI network element 140 can also be connected to a terminal device (such as terminal device 120 and / or terminal device 130); or, the AI network element 140 can be connected to both the access network device 110 and the terminal device (such as terminal device 120 and / or terminal device 130) simultaneously; or, the AI network element 140 can also be connected to the access network device 110 through a third-party network element; the AI network element 140 can also be set as a module in the access network device 110 or the terminal device (such as terminal device 120 and / or terminal device 130) shown in Figure 1. This application embodiment does not limit the connection relationship between the AI network element and other network elements.
[0106] It should be noted that Figures 1 and 2 are simplified schematic diagrams for ease of understanding. For example, the communication system may also include other devices, such as wireless relay devices and / or wireless backhaul devices, which are not shown in Figures 1 and 2. In practical applications, the communication system may include multiple access network devices and multiple terminal devices. The embodiments of this application do not limit the number of access network devices and terminal devices included in the communication system.
[0107] In the embodiments of this application, the terminal device may also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent, or user apparatus.
[0108] Terminal devices can be devices that provide voice / data, such as handheld devices with wireless connectivity, in-vehicle devices, etc. Currently, examples of terminals include: mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to a wireless modem, wearable devices, terminal devices in 5G networks, or future public land mobile communication networks. Terminal devices in a network (PLMN), etc., are not limited to this in the embodiments of this application.
[0109] By way of example and not limitation, in this embodiment, the terminal device can also be a wearable device. Wearable devices, also known as wearable smart devices, are a general term for devices that utilize wearable technology to intelligently design and develop everyday wearables, such as glasses, gloves, watches, clothing, and shoes. Wearable devices are portable devices that are worn directly on the body or integrated into the user's clothing or accessories. Wearable devices are not merely hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are feature-rich, large in size, and can achieve complete or partial functions without relying on a smartphone, such as smartwatches or smart glasses, as well as those that focus on a specific type of application function and require the use of other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.
[0110] In this embodiment, the device for implementing the functions of the terminal device can be the terminal device itself, or it can be any device capable of supporting the terminal device in implementing those functions, such as a chip system. This device can be installed in or used in conjunction with the terminal device. In this embodiment, the chip system can consist of chips or include chips and other discrete components. This embodiment only uses the terminal device as an example to illustrate the device for implementing the functions of the terminal device, and does not constitute a limitation on the solutions of this embodiment.
[0111] The network device in this application embodiment can be a device for communicating with a terminal device. This network device can include an access network device (i.e., an access network node) or a radio access network device, such as a base station. In this application embodiment, the radio access network device can refer to a radio access network (RAN) node (or device) that connects the terminal device to the wireless network. A base station can broadly encompass, or be replaced by, various names including: NodeB, evolved NodeB (eNB), next-generation NodeB (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master station, auxiliary station, motor slide retainer (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), radio unit (RU), positioning node, etc. A base station can be a macro base station, micro base station, relay node, donor node, or similar entities, or combinations thereof. A base station can also refer to a communication module, modem, or chip installed within the aforementioned equipment or apparatus. A base station can also be a mobile switching center, equipment performing base station functions in D2D, V2X, and M2M communications, network-side equipment in future communication networks, or equipment performing base station functions in future communication networks. A base station can support networks using the same or different access technologies. Optionally, a RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). The embodiments of this application do not limit the specific technologies or equipment forms used in the network equipment.
[0112] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move depending on the location of the mobile base station. In other examples, a helicopter or drone can be configured as a device to communicate with another base station.
[0113] In some deployments, the access network equipment mentioned in the embodiments of this application may be a device including a CU, or a DU, or a device including both CU and DU, or a device with a control plane CU node (central unit-control plane (CU-CP)) and a user plane CU node (central unit-user plane (CU-UP)) and a DU node. For example, the access network equipment may include gNB-CU-CP, gNB-CU-UP, and gNB-DU.
[0114] In some deployments, multiple RAN nodes collaborate to assist terminals in achieving wireless access, with different RAN nodes each implementing some of the base station's functions. For example, RAN nodes can be CUs, DUs, CU-CPs, CU-UPs, or RUs. CUs and DUs can be configured separately or included in the same network element, such as a BBU. RUs can be included in radio frequency equipment or radio frequency units, such as RRUs, AAUs, or RRHs.
[0115] RAN nodes can support one or more types of fronthaul interfaces, each corresponding to a DU and RU with different functions. If the fronthaul interface between the DU and RU is a Common Public Radio Interface (CPRI), the DU is configured to implement one or more baseband functions, and the RU is configured to implement one or more radio frequency functions. If the fronthaul interface between the DU and RU is another type of interface, relative to CPRI, it moves some downlink and / or uplink baseband functions—for example, for downlink, precoding, digital beamforming, or one or more of inverse fast Fourier transform (IFFT) / adding a cyclic prefix (CP)—from the DU to the RU; and for uplink, digital beamforming, or one or more of fast Fourier transform (FFT) / removing CP—from the DU to the RU. In one possible implementation, this interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the splitting methods between DU and RU are different, corresponding to different types (category, Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, F.
[0116] Taking eCPRI Cat A as an example, for downlink transmission, layer mapping is used as the dividing line. The DU is configured to implement one or more functions preceding layer mapping (i.e., coding, rate matching, scrambling, modulation, and layer mapping itself), while other functions following layer mapping (e.g., resource element (RE) mapping, digital beamforming, or IFFT / CP removal) are implemented in the RU. For uplink transmission, de-RE mapping is used as the dividing line. The DU is configured to implement one or more functions preceding de-mapping (i.e., decoding, rate matching de-matching, descrambling, demodulation, inverse discrete Fourier transform (IDFT), channel equalization, and de-RE mapping itself), while other functions following de-mapping (e.g., digital beamforming or FFT / CP removal) are implemented in the RU. It is understood that descriptions of the functions of the DU and RU corresponding to various types of eCPRI can be found in the eCPRI protocol and will not be elaborated upon here.
[0117] In one possible design, the processing unit in the BBU used to implement baseband functions is called the baseband high (BBH) unit, and the processing unit in the RRU / AAU / RRH used to implement baseband functions is called the baseband low (BBL) unit.
[0118] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an open RAN (ORAN) architecture, CU can also be called open CU (open-CU, O-CU), DU can also be called open DU (open-DU, O-DU), CU-CP can also be called open CU-CP (open-CU-CP) O-CU-CP, CU-UP can also be called open CU-UP (open-CU-UP, O-CU-UP), and RU can also be called open RU (open-RU, O-RU). Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software modules and hardware modules.
[0119] In this embodiment, the apparatus for implementing the functions of a network device can be a network device itself; it can also be an apparatus capable of supporting the network device in implementing those functions, such as a chip system, hardware circuit, software module, or a hardware circuit plus a software module. This apparatus can be installed in the network device or used in conjunction with the network device. In this embodiment, the example of a network device being used to implement the functions of a network device is provided only and does not constitute a limitation on the solutions described in this embodiment.
[0120] Network devices and / or terminal devices can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can also be deployed in the air on airplanes, balloons, and satellites. This application does not limit the scenario in which the network devices and terminal devices are located. Furthermore, terminal devices and network devices can be hardware devices, or software functions running on dedicated hardware or general-purpose hardware, such as virtualization functions instantiated on a platform (e.g., a cloud platform), or entities that include dedicated or general-purpose hardware devices and software functions. This application does not limit the specific form of the terminal devices and network devices.
[0121] Optionally, the AI node can be deployed in one or more of the following locations within the communication system: access network equipment, terminal equipment, or core network elements. Alternatively, the AI node can also be deployed independently, for example, in a location other than any of the aforementioned devices, such as in the host or cloud server of an over-the-top (OTT) system. The AI node can communicate with other devices in the communication system, which can be, for example, one or more of the following: access network equipment, terminal equipment, or core network elements.
[0122] It is understood that this application does not limit the number of AI nodes. For example, when there are multiple AI nodes, they can be divided based on function, such as different AI nodes being responsible for different functions.
[0123] It can also be understood that AI nodes can be AI network elements or AI modules. AI nodes can be independent devices, or they can be integrated into the same device to implement different functions. Alternatively, they can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the AI nodes described above.
[0124] Figure 3 illustrates a possible application framework in a communication system. As shown in Figure 3, network elements in the communication system are connected via interfaces (e.g., next-generation (NG) interfaces, Xn interfaces) or air interfaces. The NG interface is the interface between the radio access network and the 5G core network. The Xn interface is the interface between access network devices, and the air interface is the interface between access network devices and terminal devices. These network element nodes, such as core network devices, RAN nodes, terminal devices, or one or more devices in operation administration and maintenance (OAM), are equipped with one or more AI modules (only one is shown in Figure 3 for clarity). The access network node can be a single RAN node or can include multiple RAN nodes, for example, including CU, DU, and RU. One or more of the RU, CU, or DU can also be equipped with one or more AI modules. Optionally, the CU can be further divided into CU-CP and CU-UP. One or more AI modules are configured in CU-CP and / or CU-UP.
[0125] AI modules are used to implement corresponding AI functions. AI modules deployed in different network elements can be the same or different. Depending on the parameter configuration, the AI module can implement different functions. The AI module model can be configured based on one or more of the following parameters: structural parameters (e.g., at least one of the following: number of neural network layers, neural network width, inter-layer connections, neuron weights, neuron activation function, or bias in the activation function), input parameters (e.g., at least one of the following: type of input parameter, input dimension, number of input ports), or output parameters (e.g., at least one of the following: type of output parameter, output dimension, number of output ports). The bias in the activation function can also be called the bias of the neural network. Input dimension can refer to the size of an input data set; for example, when the input data is a sequence, the input dimension corresponding to that sequence can indicate the length of the sequence. The number of input ports can refer to the quantity of input data. Similarly, output dimension can refer to the size of an output data set; for example, when the output data is a sequence, the output dimension corresponding to that sequence can indicate the length of the sequence. The number of output ports can refer to the quantity of output data.
[0126] An AI module can have one or more models. A model can infer an output, which includes one or more parameters. The learning, training, or inference processes of different models can be deployed on different nodes or devices, or they can be deployed on the same node or device.
[0127] Network devices can be network devices equipped with one or more AI modules. These network devices can include one or more devices in the core network, RAN, or OAM as shown in Figure 3. For example, the AI module can be a RAN intelligent controller (RIC) as shown in Figure 4, such as a near-real-time RIC (near-RT RIC) or a non-real-time RIC (non-RT RIC). For instance, a near-real-time RIC is located in a RAN node (e.g., in a CU or DU), while a non-real-time RIC is located in the OAM, a cloud server, a core network device, or other access network devices. The RIC can obtain subsets from multiple terminal devices from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU), reassemble them into a dataset, and train based on the dataset. Exemplarily, near-real-time RICs and non-real-time RICs can also be configured as separate network elements, and access network devices can be either near-real-time or non-real-time RICs.
[0128] Figure 4 illustrates another possible application framework in a communication system. In addition to access network nodes (CU, DU, and RU are shown in the figure) and terminals, the communication system shown in Figure 4 also includes an RIC (Regulator-Integrated Circuit). For example, the RIC could be the AI module shown in Figure 3, which can be used to implement AI-related functions. The RIC includes near-real-time RICs and non-real-time RICs. Non-real-time RICs primarily process non-real-time information, such as data that is not sensitive to latency, with latency on the order of seconds. Real-time RICs primarily process near-real-time information, such as data that is relatively sensitive to latency, with latency on the order of tens of milliseconds.
[0129] Near real-time (NRT) RICs are used for model training and inference. For example, they are used to train AI models and then use those models for inference. NRT RICs can obtain network-side and / or terminal-side information from RAN nodes (e.g., CUs, CU-CPs, CU-UPs, DUs, and / or RUs) and / or terminals. This information can be used as training data or inference data. Optionally, the NRT RIC can deliver inference results to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CUs and DUs, and / or between DUs and RUs. For example, the NRT RIC delivers inference results to a DU, which then forwards them to an RU.
[0130] Non-real-time RICs are also used for model training and inference. For example, they can be used to train AI models and then use those models for inference. Non-real-time RICs can obtain network-side and / or terminal-side information from RAN nodes (e.g., one or more of CU, CU-CP, CU-UP, DU, or RU) and / or terminals. This information can be used as training data or inference data, and the inference results can be delivered to the RAN nodes and / or terminals. Optionally, inference results can be exchanged between CUs and DUs, and / or between DUs and RUs; for example, a non-real-time RIC delivers inference results to a DU, which then forwards them to an RU.
[0131] Near real-time RICs and non-real-time RICs can also be configured as separate network elements. Alternatively, near real-time RICs and non-real-time RICs can also be part of other devices. For example, near real-time RICs can be set in RAN nodes (e.g., CU, DU), while non-real-time RICs can be set in OAM, cloud servers, core network devices, or other network devices.
[0132] With the development of AI, AI / ML technology has been introduced into the wireless air interface, for example, for the compression and reconstruction of wireless channel information, beam management and positioning enhancement, and can also be used for data processing in transmitters and / or receivers.
[0133] Figure 5 is a schematic diagram of the AI / ML functional framework for wireless air interfaces. Figure 5 shows the data collection unit, model training unit, management unit, inference unit, and model storage unit, etc.
[0134] The data collection unit can collect data from network devices and / or terminal devices. AI-based air interface or RAN systems require different types of data at various stages of the AI model development process. For example, training the model requires a large amount of diverse data, model inference requires real-time data, and model monitoring requires near real-time data. For wireless or air interface tasks, the data collected by the data collection unit via the wireless air interface includes, but is not limited to, one or more of the following: measurement signals transmitted by network devices, measurement results measured and reported by terminal devices, reference signals transmitted by terminal devices, and measurement results measured by network devices. The data collected by the data collection unit, after preprocessing and other operations, can be used for one or more of the following: model training, model inference, model monitoring, or analysis. The data collected by the data collection unit can be applied to different stages, such as training data for model training, performance monitoring data for monitoring model performance, and so on.
[0135] The model training unit can train or retrain (or update) the model based on the training data from the data collection unit, and can send the trained model or the updated model to the model storage unit for storage.
[0136] The management unit can monitor the model's performance based on monitoring data from the data collection unit, and perform actions such as activating / deactivating, switching, selecting, and rolling back the model (or model function) based on performance. For example, it can send performance feedback or retraining requests to the model training unit. The management unit can also send model transfer requests to the model storage unit to request that the model storage unit send the model to the inference unit.
[0137] The inference unit can perform model inference based on the model from the model storage unit and the inference data from the data collection unit, and output the inference results. The inference unit can also receive instructions from the model management unit.
[0138] Intelligent air interface technology can be understood as the application of AI models, or neural network (NN) models, in wireless communication systems to optimize the performance of wireless communication systems.
[0139] Use cases of AI models in wireless communication include, but are not limited to, channel prediction, channel compression and reconstruction, channel estimation, multi-input multi-output (MIMO) equalization, modulation and demodulation, coding and decoding, etc. These use cases are described below with reference to the accompanying figures.
[0140] Channel Prediction: Figure 6 illustrates the AI model used to implement channel prediction. As shown, the model's input is historical channel information, and its output is the predicted future channel information, i.e., the predicted value of the channel information. Model training data includes historical channel information and the ground truth values of future channel information. Model training can involve updating model parameters with the goal of minimizing the error between the predicted and ground truth values (e.g., minimum square error (MSE)) or maximizing the similarity (e.g., square generalized cosine similarity (SGCS)). The model's performance is evaluated by the error (e.g., MSE) or the similarity (e.g., SGCS) between the predicted and actual future channel information; smaller errors or higher similarities result in better performance.
[0141] Channel compression and reconstruction: This can be achieved through channel compression models and channel reconstruction models. Figure 7 is a schematic diagram of an AI model used to implement channel compression and reconstruction (or reconstruction) functions. As shown in the figure, the input of the channel compression model can be the original channel information, such as the channel matrix, precoding matrix, or eigenvector, and the output can be the compressed bits; the input of the channel reconstruction model (or channel reconstruction model) can be the compressed bits, and the output can be the reconstructed (or rebuilt) channel information. The model training data includes the original channel information. The model training can include updating the model parameters with the goal of minimizing the error (e.g., MSE) or maximizing the similarity (e.g., SGCS) between the reconstructed channel information obtained through the channel compression model and the channel reconstruction model and the original channel information. The performance of the model is evaluated by the error (e.g., MSE) or similarity (e.g., SGCS) between the reconstructed channel and the original channel; the smaller the error or the higher the similarity, the better the performance.
[0142] Channel Estimation: Figure 8 illustrates the AI model used to implement channel estimation. As shown, the model input is the received signal, including multiple resource elements (REs). Some of these REs carry reference signals (RS), also known as pilot signals. These REs can be called reference signal REs or pilot REs. The model output is the estimated channel information corresponding to multiple REs. Model training data can include the channel information corresponding to the received signal or pilot REs, and the channel information of all REs (e.g., true channel information, i.e., the ground truth of channel information). Model training can involve updating model parameters with the goal of minimizing the error (e.g., MSE) or maximizing the similarity (e.g., SGCS) between the estimated channel information and the true channel information. The model performance is evaluated by the error (e.g., MSE) or similarity (e.g., SGCS) between the estimated channel information and the true channel information; the smaller the error or the higher the similarity, the better the performance.
[0143] MIMO equalization: This can be simply called equalization. Figure 9 is a schematic diagram of an AI model used to implement MIMO equalization. As shown, the model's input is the signal (or symbol) received through Nr antennas, and the output is the equalized signal (or symbol) from Nt streams. Here, Nr represents the number of receiving antenna ports, and Nt represents the number of data streams, or transmission layers (it should be understood that Nt data streams are obtained by mapping data onto Nt transmission layers; therefore, a data stream can be considered as data mapped to one transmission layer, such as a symbol). Both Nr and Nt are positive integers greater than or equal to 1. The model training data includes the received signal and the corresponding transmitted symbol, which is then used to generate the equalized symbol output by the model. Model training can include minimizing the error between the equalized symbol and the transmitted symbol (such as MSE or cross entropy (CE)) to update the model parameters. The model's performance is evaluated by the error (such as MSE or CE) between the equalized symbol and the transmitted symbol; the smaller the error, the better the performance.
[0144] Demodulation: Figure 10 is a schematic diagram of the AI model used to implement the demodulation function. As shown in the figure, the input of the model is the received symbol or the equalized symbol, and the output is the demodulated bit corresponding to the symbol (as shown in Figure 10(a)) or the log-likelihood ratio (LLR) corresponding to the demodulated bit (as shown in Figure 10(b)). N in the figure represents N demodulated bits, and N is a positive integer greater than 1. The received symbol can be an un-pre-encoded symbol or a pre-encoded symbol. This model can also be used to implement the equalization function, so the received symbol can be equalized first and then demodulated.
[0145] The model training data includes received symbols and their corresponding transmitted bits (i.e., known bits). The model parameters are updated with the objective of minimizing the error between the estimated demodulated bits and the transmitted bits (such as MSE, CE, or bit error rate (BER)). The model's performance is evaluated by the error between the demodulated bits and the transmitted bits (such as MSE, CE, or BER); the smaller the error, the better the performance.
[0146] Decoding: Figure 11 is a schematic diagram of the AI model used to implement the decoding function. As shown in the figure, the input of the model is the demodulated signal (the bits to be decoded shown in the figure), and the output is the estimated decoded bits. In the figure, N represents N bits to be decoded; K represents K decoded bits; N is greater than K, and N and K are positive integers. The model training data includes the received signal or the demodulated signal, and the corresponding transmitted bits (i.e., known bits). The training of the model may include updating the model parameters with the goal of minimizing the error (such as MSE, CE, or BER) between the estimated decoded bits output by the model and the transmitted bits. The performance of the model is evaluated by the error (such as MSE, CE, or BER) between the estimated decoded bits and the transmitted bits; the smaller the error, the better the performance.
[0147] Figure 12 illustrates a wireless communication system comprising a neural network-based transceiver (including a transmitter and a receiver) provided in an embodiment of this application. The transceiver in the system shown in Figure 12, employing a neural network model, can optimize the performance of transmitting and receiving signals through data-driven training.
[0148] Transceivers employing neural network models can optimize signal transmission and reception performance through data-driven training. This means optimizing one or more modules within the transceiver used for signal processing, such as channel coding, modulation, waveform generation, and pilot signals, using the neural network model. In the diagram, the neural network (e.g., a deep neutral network (DNN)) can replace one or more modules in the transmitter and receiver to perform one or more operations. More specifically, a neural network-based transmitter can be used to perform one or more of the following in data transmission processing: channel coding, modulation, waveform shaping, precoding, time-domain signal processing, or reference signal (or pilot) optimization. Time-domain signal processing can include various possible operations, such as digital predistortion, pre-equalization, nonlinear compensation, and beamforming, and the neural network can be used to perform one or more of these operations. A neural network-based receiver can be used to perform one or more of the following in data reception processing: time-domain signal processing, equalization, waveform shaping, demodulation, channel decoding, or RS-based channel estimation. More specific examples can be found in Figures 13 to 15.
[0149] Figure 13 is an example of a transceiver including a neural network based embodiment of this application. The transmitter's neural network can be used to optimize the modulation module and / or time-domain signal processing. The information bits to be transmitted are channel-coded to obtain coded bits. After the coded bits are modulated by the neural network, modulation symbols are obtained. The modulation symbols are then RE-mapped to map them onto physical resources. After inverse fast Fourier transform (IFFT), the time-domain signal is obtained. After further time-domain signal processing by the neural network, the signal is transmitted. The receiver's neural network can be used to optimize the demodulation module and time-domain signal processing. The received signal is processed by the neural network for time-domain signal processing and fast Fourier transform (FFT) to obtain modulation symbols. After demodulation by the neural network, the bits to be decoded are obtained. The bits to be decoded are then channel-decoded to obtain estimated information bits.
[0150] Figure 14 is another example of a transceiver including a neural network based embodiment of this application. The transmitter's neural network is used for joint optimization of encoding and modulation, and time-domain signal neural network processing. The information bits to be transmitted are processed by the neural network to obtain modulation symbols, which are then transmitted after IFFT, time-domain signal processing by the neural network, and other operations. The receiver's neural network can be used for joint optimization of decoding and demodulation, as well as time-domain signal processing. The received signal, after time-domain signal processing and FFT by the neural network, can obtain modulation symbols. The neural network can estimate the information bits based on the modulation symbols to obtain the estimated information bits.
[0151] Figure 15 is another example of a transceiver including a neural network based on an embodiment of this application. The transmitter's neural network is used to optimize the reference signal. The transmitter generates a reference signal based on the neural network, which is then processed by IFFT and a time-domain signal neural network before being transmitted. The receiver's neural network is used for channel estimation. The received signal is processed by the time-domain signal neural network and then by FFT to obtain the received reference signal. The neural network can perform channel estimation based on the received reference signal.
[0152] Although not shown in the figures, it can be understood that since waveform shaping is primarily performed after the IFFT or before the FFT, and optionally also before or after the IFFT, the waveform shaping operations shown in Figures 13 to 15 can be included within the IFFT or FFT, and are not shown separately. Furthermore, neural networks are not limited to performing the modulation and time-domain signal processing shown in Figure 13, and the coded modulation and time-domain signal processing shown in Figure 14; as previously mentioned, neural networks can also be used to optimize waveforms.
[0153] It should be understood that the wireless communication systems with neural network-based transceivers shown in Figures 13 to 15 are merely examples. In practical applications, neural network-based transceivers can also be used to implement more functions, such as precoding and waveform shaping. For the sake of simplicity, these will not be illustrated with accompanying figures here.
[0154] Figure 16 is a schematic diagram of the training process of the intelligent air interface provided in an embodiment of this application. As shown in the figure, the training of the neural network model for the wireless communication system includes the training of the transmitting-end neural network model and the receiving-end neural network model. The transmitting-end neural network model performs forward inference on the training data and outputs the forward inference result (as shown in the figure, z). x The forward inference result is sent to the receiving neural network model; the receiving neural network model performs forward inference and backward gradient calculation based on the forward inference result, and obtains the feedback result (as shown in the figure). The receiving neural network model sends the feedback result to the sending neural network model; the sending neural network model performs backpropagation gradient calculation based on the feedback result, and updates its model parameters based on the gradient obtained from the backpropagation gradient calculation. For example, the t-th round update of the sending neural network model's parameters satisfies:
[0155] Among them, w t+1 It refers to the parameters in round t+1, i.e., the parameters updated in round t; w t It is the parameter for round t; z x It is the output of the sending neural network model corresponding to the x-th training sample, l x η is the loss function for round t, and X is the number of samples trained in this round; t It is the learning rate; It is the gradient of the loss function with respect to the output of the sending neural network model.
[0156] In other words, during any training round, the sending neural network model first performs forward inference to obtain the forward inference result and intermediate results; the intermediate inference results are cached; then, the loss function is calculated from the last layer, and gradients are calculated layer by layer according to the chain rule. The gradient calculation process requires the results of forward inference from each layer. The gradients calculated from multiple training data are summed and used to update the model parameters of the sending neural network model.
[0157] In the use of AI models, it is usually necessary to monitor (or evaluate) their performance. Taking the performance monitoring of an AI model deployed on a terminal device by a network device as a management node as an example, the performance monitoring process is as follows: The network device can issue monitoring resources (or evaluation resources), which can be used to transmit reference signals or label data. The terminal device can obtain measurement results (e.g., as label data) by measuring the reference signals on the monitoring resources, or directly obtain label data from the monitoring resources. Then, by comparing the prediction results output by the AI model with the measurement results, the performance index value is obtained, which is then reported to the network device.
[0158] Optionally, the terminal device can report the value of a performance metric to the network device. Here, a performance metric can be understood as a function defined to measure the performance of the model. "Performance metric" is just one possible name; it can also be replaced by a performance measurement function, a predefined function, a loss function, etc. The value of a performance metric refers to the specific numerical value calculated by substituting data into the performance metric based on the defined function. The value of a performance metric is just one possible name; it can also be replaced by a performance measurement value, a function value, etc. This application does not limit this.
[0159] For example, performance metrics can be functions used to represent the difference between the labeled data and the model output, such as MSE, normalized mean square error (NMSE), generalized cosine similarity (GCS), SGCS, CE, BER, etc.
[0160] For example, when the training process shown in Figure 16 is applied to an AI model for channel prediction, the loss function can be the error (e.g., MSE), similarity (e.g., SGCS), or a function of similarity (e.g., 1-SGCS) between the predicted and true values of future channel information; when the training process is applied to an AI model for channel compression and reconstruction, the loss function can be the error (e.g., MSE), similarity (e.g., SGCS), or a function of similarity (e.g., 1-SGCS) between the reconstructed and original channel information; when the training process is applied to an AI model for channel estimation, the loss function can be the channel information corresponding to all estimated REs. The loss function can be the error (e.g., MSE), similarity (e.g., SGCS), or a function of similarity (e.g., 1-SGCS) between the information and the true channel information corresponding to all REs; when this training process is applied to an AI model for implementing MIMO equalization, the loss function can be the error between the equalized symbol and the transmitted symbol (e.g., MSE or CE); when this training process is applied to an AI model for implementing demodulation, the loss function can be the estimated error between the demodulated bit and the transmitted bit (e.g., MSE, BER, or CE); when this training process is applied to an AI model for implementing decoding, the loss function can be the estimated error between the decoded bit and the transmitted bit (e.g., MSE, BER, or CE).
[0161] It should be understood that the examples of loss functions applied to AI models with different functionalities above are merely illustrative and should not constitute any limitation on this application. For example, error can also be represented by NMSE, and similarity can also be represented by GCS, etc. This application includes, but is not limited to, these methods.
[0162] Figure 17 illustrates a schematic diagram of model performance monitoring. As shown in Figure 17, on one hand, the signals carried on the monitoring resources serve as input to the AI function, and the signals are output after processing by the AI model. On the other hand, the monitoring resources can also carry label data or signals used to obtain label data, for example, by processing the signals through a fallback function to obtain label data. The performance monitoring of this AI model can be obtained by calculating the difference or correlation between the AI model's output and the label data. Specifically, the fallback function can be implemented using traditional methods, or it can be implemented using other methods that can provide label data, such as using a default or general AI model relative to the AI function being monitored.
[0163] As mentioned earlier, when management nodes perform performance monitoring on multiple models, they configure resources for monitoring, resources for reporting, performance metrics, and so on for different models. In other words, the management node monitors model performance at the individual model level. When multiple AI models are deployed in a device to implement different functions, performance monitoring may incur significant signaling overhead.
[0164] In view of this, this application provides a communication method that monitors the performance of multiple AI functions as a whole, and continues to monitor the performance of some of the AI functions when performance requirements are not met. This allows for configuring monitoring resources once for each performance monitoring session, rather than configuring monitoring resources for each AI function individually, thereby reducing signaling overhead.
[0165] The method provided in this application will be described in detail below with reference to the accompanying drawings. The following embodiments use the interaction between a first device and a second device as an example to describe the specific flow of the method. The first device can be a terminal device, or a component within a terminal device, such as internal circuits, chips, chip systems, processors, or communication modules; or it can be a logic module or software capable of implementing all or part of the functions of the terminal device, etc. The second device can be a network device, or a component within a network device, such as internal circuits, chips, chip systems, processors, or communication modules; or it can be a logic module or software capable of implementing all or part of the functions of the network device, etc. The second device can act as a management node to manage the AI functions deployed in other devices (such as terminal-side devices).
[0166] It should be noted that the AI functionality in the embodiments of this application can also be called AI feature. An AI functionality or an AI feature can be implemented by one or more models. A model can also be used to implement multiple AI functionality.
[0167] Figure 18 is a schematic flowchart of a communication method provided in an embodiment of this application. As shown in Figure 18, the method 1800 may include steps 1801 to 1802, and optionally one or more of steps 1803 to 1813. The various steps in Figure 18 are described in detail below.
[0168] In step 1801, the second device sends first information to the first device, which indicates first monitoring resources configured for N AI functions. Correspondingly, the first device receives the first information from the second device.
[0169] Where N is a positive integer greater than 1. For example, N is 2; or N is 3, and so on, without limitation.
[0170] In this embodiment, the first monitoring resource is a monitoring resource configured for N AI functions. It can be used to monitor the performance of the N AI functions as a whole. Monitoring the performance of the N AI functions as a whole means treating them as a single entity, obtaining their output, and comparing this output with the labeled data to determine the differences or correlations between them. Simply put, the AI function 1 shown in Figure 17 above can be replaced with the N AI functions in this embodiment.
[0171] In step 1802, the second device transmits or receives the first signal based on the first monitoring resource. Correspondingly, the first device receives or transmits the first signal based on the first monitoring resource.
[0172] The first monitoring resource can be used to transmit the first signal, or in other words, to carry the first signal. This first signal can be used to obtain the first monitoring result for performance monitoring of the N AI functions as a whole.
[0173] Optionally, the second device transmits a first signal based on the first monitoring resource, and the first device receives the first signal based on the first monitoring resource.
[0174] For example, the first signal can be used as input when the N AI functions are treated as a whole.
[0175] In one possible implementation, the first signal is used as input to N AI functions as a whole, and the first performance metric is determined by the output of the N AI functions as a whole. For example, following the signal flow, the N AI functions can be defined as the 1st (or first) AI function, the 2nd AI function, ..., the Nth (or last) AI function. The output of the nth AI function can be used as the input of the (n+1)th AI function, where n is a positive integer from 1 to (N-1). Therefore, using the first signal as input to N AI functions as a whole can mean that the first signal can be used as input to the first AI function among the N AI functions. The output of the N AI functions as a whole can mean the output of the last AI function among the N AI functions.
[0176] Since the first monitoring resource is used to transmit the first signal, the first device can receive the first signal based on the first monitoring resource. The first signal is used as input when the N AI functions are used as a whole. This can mean that the first signal can be used as input when the N AI functions are used as a whole, or it can be used to obtain other signals, and the other signals obtained from the first signal can be used as input when the N AI functions are used as a whole.
[0177] Optionally, the first device transmits a first signal based on the first monitoring resource, and the second device receives the first signal based on the first monitoring resource.
[0178] For example, the first signal can be the output of one of the N AI functions (e.g., denoted as AI function 1), and can also be used as the input of another AI function (e.g., denoted as AI function 1) among the N AI functions. AI function 1 and AI function 2 can be deployed in different devices; for example, AI function 1 can be deployed on the terminal side, and AI function 2 can be deployed on the network side.
[0179] For ease of illustration, step 1802 is represented by a double-headed arrow in the figure, that is, the first signal can be sent from the second device to the first device, or it can be sent from the first device to the second device.
[0180] It should be noted that all N AI functions can be deployed on the terminal side, or some of the N AI functions can be deployed on the terminal side and some on the network side. Since the first signal is transmitted to the first device, the first device uses the first signal as the input when the N AI functions are treated as a whole. This means that the first AI function among the N AI functions can be deployed on the terminal side.
[0181] The terminal side can also be called the UE side, terminal side device, or UE side equipment, etc., and may include: terminal equipment (or user equipment, terminal, etc.), components deployed in the terminal equipment (such as circuits or chips inside the terminal equipment), devices deployed outside the terminal equipment (such as the host or cloud server of the OTT system, hereinafter referred to as the OTT system server), or components deployed in devices outside the terminal equipment (such as circuits or chips inside the device).
[0182] The network side (NW side) can also be referred to as network-side equipment, and may include: network equipment that communicates with terminal equipment, components deployed within network equipment (such as circuits or chips with near real-time radio access network (RAN) intelligent control functions within the network equipment), equipment deployed outside the network equipment (such as intelligent network elements, for example, intelligent network elements with near real-time RAN intelligent control functions), or components deployed within intelligent network elements (such as circuits or chips within the intelligent network element). Network equipment may include: access network equipment, core network equipment, or operation administration and maintenance (OAM) equipment.
[0183] In this embodiment, all N AI functions are deployed on the terminal side, meaning that the N AI functions can be deployed on terminal-side devices, such as those listed above. Alternatively, some of the N AI functions are deployed on the terminal side and some on the network side; that is, at least one of the N AI functions is deployed on a terminal-side device, such as those listed above, while the other AI functions are deployed on network-side devices, such as those listed above.
[0184] In step 1803, the second device sends third information to the first device, which indicates N AI functions. Accordingly,
[0185] The first device receives third information from the second device.
[0186] The second device can use this third information to instruct the first device which AI functions the first monitoring resources should be used for performance monitoring.
[0187] In one possible design, the third information and the aforementioned first information can be carried in the same signaling. In this case, for the second device, steps 1801 and 1803 can be combined into a single transmission step, and for the first device, steps 1801 and 1803 can be combined into a single reception step.
[0188] In another possible implementation, the N AI functions can also be determined by the first device. For example, the N AI functions can be all the AI functions deployed in the first device; or, for example, the N AI functions can be some of the AI functions deployed in the first device, and the first device can indicate the N AI functions to the second device through other signaling, such as through other signaling before step 1801.
[0189] One possible scenario is that all N AI functions are deployed on the terminal side, so the first device can determine the first monitoring result. The method may include one or more of the following steps 1804 to 1806.
[0190] In step 1804, the first device determines the first monitoring result.
[0191] Since all N AI functions are deployed on the terminal side, the first device can obtain the output of the N AI functions as a whole. Therefore, based on the output and the tag data, the first monitoring result can be determined based on the first performance indicator.
[0192] In this embodiment, the first performance indicator can be used to monitor the performance of N AI functions. In other words, if the N AI functions are monitored as a whole, the difference or correlation between the output of the N AI functions as a whole and the label data can be determined based on the first performance indicator, thereby obtaining the first monitoring result.
[0193] The first performance indicator may be indicated by the second device, such as by the third information described below, or it may be determined by the first device or predefined, such as by protocol predefined. This application does not limit this.
[0194] In step 1805, if the first monitoring result meets the first condition, the first device sends the first monitoring result to the second device on the first resource. Accordingly, the second device receives the first monitoring result from the first device on the first resource.
[0195] The first condition is the condition that must be met for the first monitoring result to be sent. That is, if the first monitoring result meets the first condition, then the first monitoring result is sent; if the first monitoring result does not meet the first condition, then the first monitoring result is not sent. Therefore, step 1805 is not necessarily required to be executed.
[0196] For example, the first condition may be that the first monitoring result is greater than or equal to the upper bound corresponding to the first performance index. For example, if the first performance index is MSE, the first condition may be that the first monitoring result MSE is greater than or equal to the upper bound of MSE. Alternatively, the first condition may be that the first monitoring result is less than or equal to the lower bound corresponding to the first performance index. For example, if the first performance index is SGCS, the first condition may be that the first monitoring result SGCS is less than or equal to the lower bound of SGCS.
[0197] The first condition may be indicated by the second device, or it may be determined by the first device itself, or it may be predefined by the protocol. This application does not limit this.
[0198] In another implementation, the first condition can also be a condition that prevents the first monitoring result from being sent. That is, if the first monitoring result meets the first condition, then the first monitoring result is not sent; if the first monitoring result does not meet the first condition, then the first monitoring result is sent. For example, the first condition is that the first monitoring result is less than the upper bound corresponding to the first performance indicator, or the first condition is that the first monitoring result is greater than the lower bound corresponding to the first performance indicator.
[0199] Therefore, step 1805 can also be replaced by the first device sending the first monitoring result to the second device if the first monitoring result does not meet the first condition. For ease of understanding and explanation, unless otherwise specified, the first condition can be understood according to the former implementation method, that is, the first monitoring result is sent if the first condition is met.
[0200] The first condition can also be that the first monitoring result indicates a performance decline.
[0201] Optionally, performance degradation can be determined based on the trend of monitoring results. For example, if the primary performance indicator is MSE, and the MSE obtained in this monitoring (i.e., the first monitoring result) is larger than the MSE obtained in the previous monitoring, then performance degradation is indicated. As another example, if the primary performance indicator is SGCS, and the SGCS obtained in local monitoring (i.e., the first monitoring result) is smaller than the SGCS obtained in the previous monitoring, then performance degradation is indicated.
[0202] Optionally, whether performance has degraded can be determined based on the trend and magnitude of changes in the monitoring results. For example, if the primary performance indicator is MSE, and the MSE obtained in this monitoring (i.e., the first monitoring result) is larger than the MSE obtained in the previous monitoring, and the increase reaches a preset threshold, then performance has degraded. As another example, if the primary performance indicator is SGCS, and the SGCS obtained in the local monitoring (i.e., the first monitoring result) is smaller than the SGCS obtained in the previous monitoring, and the decrease reaches a preset threshold, then performance has degraded.
[0203] It should be understood that the above description uses MSE and SGCS as two examples of primary performance metrics for ease of understanding and explanation only, but this should not constitute any limitation on this application. MSE can also be replaced by other metrics that can be used to represent the difference between output and label data, and SGCS can also be replaced by other metrics that can be used to identify the degree of correlation between output and label data.
[0204] It should also be understood that the first performance indicator is not limited to one indicator; it can be multiple indicators. If the first performance indicator is multiple, the first condition can also include more conditions accordingly. For the sake of brevity, no further examples will be provided.
[0205] In another implementation, the first device may also periodically send the first monitoring result, rather than necessarily sending it only when the first condition is met or not met. It should be understood that the first monitoring result periodically sent by the first device can be obtained from the most recent performance monitoring of the N AI functions.
[0206] In step 1806, the second device sends second information to the first device, the second information indicating one or more of the following: a first performance indicator, a first condition, or a first resource. Accordingly, the first device receives the second information from the second device.
[0207] As mentioned above, the first performance indicator can be used to determine the first monitoring result. That is, the first device can monitor the performance of N AI functions, and based on the first performance indicator, determine the difference or correlation between the output of the N AI functions as a whole and the label data, thereby obtaining the first monitoring result.
[0208] When the first performance indicator is indicated by the second device, the second device can indicate the first performance indicator to the first device via the second information. When the first performance indicator is determined by the first device or predefined by the protocol, the second information may not be used to indicate the first condition.
[0209] The first condition is the condition that must be met for the transmission of the first monitoring result. That is, when the first monitoring result is obtained by the first device, the first device can send the first monitoring result if the first monitoring result meets the first condition.
[0210] When the first condition is indicated by the second device, the second device can indicate the first condition to the first device via the second information. When the first condition is determined by the first device or predefined by a protocol, the second information may not be used to indicate the first condition.
[0211] The first resource is the resource used to transmit the first monitoring results.
[0212] Optionally, the first resource is pre-configured by the second device for the first device. In this case, the second information and the aforementioned first information and / or third information can be carried in the same signaling. When the first, second, and third information are carried in the same signaling, steps 1801, 1803, and 1806 can be combined into a single transmission step for the second device, and steps 1801, 1803, and 1806 can be combined into a single reception step for the first device.
[0213] Optionally, the first resource is configured by the second device for the first device when it determines that the first monitoring result meets the first condition. If the first resource is configured when the first monitoring result meets the first condition, the second device may not configure the first resource if the first monitoring result does not meet the first condition. In this case, the second information may not be used to indicate the first resource. Furthermore, if the first resource is configured when the first monitoring result meets the first condition, and both the first condition and the first resource are indicated by the second device, the second device may first indicate the first condition via signaling, and then, if the first monitoring result meets the first condition, indicate the first resource via signaling. In this case, the information used to indicate the first condition and the information used to indicate the first resource can be carried in different signaling messages; in other words, the second information can be carried in one or more signaling messages. Further, the information used to indicate the first condition and the aforementioned first information and / or third information can be carried in the same signaling message.
[0214] Another possibility is that the N AI functions are partially deployed on the terminal side and partially deployed on the network side, so the second device can also determine the first monitoring result. The method may also include step 1807, in which the second device determines the first monitoring result.
[0215] Since the N AI functions are deployed on the terminal side and the network side respectively, the output of some AI functions on the terminal side can be transmitted to the network side as input to another part of the AI functions on the network side, thus obtaining the output of the N AI functions as a whole. In this case, since the first monitoring result is determined by the second device, the second device does not need to instruct the first performance indicator, the first condition, and the first resource to the first device. In other words, steps 1804 to 1806 can be omitted.
[0216] Steps 1801 to 1807 above can be regarded as the first round of performance monitoring of the AI function. The first round of performance monitoring is to monitor the performance of the N AI functions as a whole.
[0217] In step 1808, the second device sends fourth information to the first device, which indicates M AI functions out of N AI functions, and the second monitoring resources configured for the M AI functions. Correspondingly, the first device receives the fourth information from the second device.
[0218] In one possible implementation, the second device can configure the first and second monitoring resources together. That is, the first and fourth information can be carried in the same signaling message. For the second device, steps 1801 and 1808 can be combined into a single transmission step, and for the first device, steps 1801 and 1808 can be combined into a single reception step.
[0219] In another possible implementation, the second device can send a fourth message to the first device if it determines that the first monitoring result does not meet the performance requirements. That is, the first and fourth messages can be carried in different signaling messages.
[0220] The performance requirement can be the opposite of the first condition, for example, the performance requirement and the first condition can have the same threshold value but different values. The first monitoring result not meeting the performance requirement can mean that the first monitoring result meets the first condition. Alternatively, the performance requirement can have a different threshold value than the first condition; for example, a different threshold value (such as the upper or lower bound of the performance indicators listed above in conjunction with the first condition) can be used as the basis for judgment.
[0221] Where M is a positive integer greater than or equal to 1. For example, M is 1; or M is 2, and so on, without limitation.
[0222] The second monitoring resource is configured for M AI functions, and it can be used to monitor the performance of these M AI functions as a whole. These M AI functions are a subset of N AI functions. Monitoring the performance of the M AI functions as a whole means monitoring the overall performance of a subset of the N AI functions.
[0223] Similar to the first monitoring resource, the second monitoring resource can be used to send or receive a second signal, which can be used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole. In other words, the second monitoring resource is used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole. Similar to the first monitoring result, the second monitoring result is determined based on the output of the M AI functions as a whole.
[0224] It should be noted that if the performance monitoring results obtained from the M AI functions (e.g., denoted as the second monitoring result) do not meet the performance requirements, it can be determined that at least one of the M AI functions does not meet the performance requirements. The scope can be further narrowed down by treating a subset of the M AI functions as a whole for performance monitoring. Conversely, if the second monitoring result obtained from the performance monitoring of the M AI functions meets the performance requirements, it can be determined that at least one of the other (NM) AI functions needs parameter adjustment. The other (NM) AI functions can then be treated as a whole for performance monitoring.
[0225] In step 1809, the second device sends fifth information to the first device, which indicates one or more of the following: a second performance indicator, a second condition, a second resource, or a triggering condition. Accordingly, the first device receives the fifth information from the second device.
[0226] This second performance metric can be used to monitor the performance of M AI functions. In other words, if the M AI functions are monitored as a whole, the difference or correlation between the output of the M AI functions as a whole and the label data can be determined based on this second performance metric, thereby obtaining the second monitoring result.
[0227] The second device can use this fifth information to instruct the first device on what performance indicators to use for performance monitoring of the M AI functions.
[0228] In one possible design, the fifth information and the aforementioned fourth information can be carried in the same signaling. In this case, for the second device, steps 1808 and 1809 can be combined into a single transmission step, and for the first device, steps 1808 and 1809 can be combined into a single reception step.
[0229] In another possible implementation, the second performance indicator may also be determined by the first device, or it may be predefined by the protocol. In this case, the fifth information may not be used to indicate the second performance indicator. If the second performance indicator is determined by the first device, the first device may also indicate the second performance indicator to the second device via other signaling.
[0230] The second device can also instruct the first device, through the fifth information, one or more of the following: a second condition, a second resource, or a triggering condition. The second condition is the condition satisfied for sending the second monitoring result; the second resource is the resource used to transmit the second monitoring result; and the triggering condition is the trigger condition for performing performance monitoring on the M AI functions. It can be understood that the function of the second condition is similar to that of the first condition, and the function of the second resource is similar to that of the first resource. The difference is that the second condition and the second resource are configurations specific to the second monitoring result, while the first condition and the first resource are configurations specific to the first monitoring result. For an understanding of the second condition and the second resource, please refer to the above explanation of the first condition and the first resource; it will not be repeated here.
[0231] Understandably, if all M AI functions are deployed on the terminal side, the second monitoring result can be determined by the first device. In this case, the first device may send the second monitoring result to the second device, and therefore the second device can indicate the second condition and / or the second resource to the first device. If some of the M AI functions are deployed on the terminal side and some on the network side, the second monitoring result may be determined by the first device. In this case, the second device may not indicate the second condition and / or the second resource to the first device, or in other words, the fifth information is not used to indicate the second condition and / or the second resource.
[0232] Alternatively, the second condition can be determined by the first device or predefined by the protocol. In the case where the second condition is determined by the first device, the first device can also indicate the second condition to the second device through other signaling.
[0233] Triggering conditions can be used to trigger the first device to perform performance monitoring on M AI functions. For example, triggering conditions can be the same as or different from the first condition described above. For instance, triggering conditions can have a higher upper bound and / or a lower lower bound than the first condition, or a higher lower bound and / or a lower upper bound. Triggering conditions can also be opposite to the performance requirements described above; for example, the performance requirements and triggering conditions can have the same threshold value but different values. The first monitoring result not meeting the performance requirements can mean that the first monitoring result meets the triggering conditions. Alternatively, triggering conditions can have different threshold values than the performance requirements; for example, a threshold value different from the performance requirements can be used as the basis for judgment.
[0234] As mentioned earlier, the second monitoring resource and the first monitoring resource can be configured through the same signaling. The first device can determine whether to continue monitoring the M AI functions based on the triggering conditions. The second monitoring resource can also be configured when the first monitoring result does not meet the performance requirements. In this case, the first device can directly perform performance monitoring on the M AI functions based on the second monitoring resource, without having to determine whether to perform performance monitoring on the M AI functions based on the triggering conditions. Therefore, the fifth information may not be used to indicate the triggering conditions.
[0235] Alternatively, the triggering condition may be determined by the first device or predefined by the protocol. In this case, the fifth information may not be used to indicate the triggering condition.
[0236] In step 1810, the second device transmits or receives a second signal based on the second monitoring resource. Correspondingly, the first device receives or transmits the second signal based on the second monitoring resource.
[0237] The second monitoring resource can be used to transmit the second signal, or in other words, to carry the second signal. This second signal can be used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole.
[0238] Optionally, the second device transmits a second signal based on the second monitoring resources, and the first device receives the second signal based on the second monitoring resources.
[0239] For example, the second signal can be used as input when the M AI functions are considered as a whole. The function of the second signal is similar to that of the first signal, except that the second signal is used as input when the M AI functions are considered as a whole, while the first signal is used as input when the N AI functions are considered as a whole. Therefore, the second signal can be designed based on the first AI function among the M AI functions. Thus, the second signal may be the same as the first signal, or it may be a different signal.
[0240] The second signal is used as the input when the M AI functions are treated as a whole. This can mean that the second signal can be used as the input when the M AI functions are treated as a whole, or it can be used to obtain other signals, and the other signals obtained by the second signal can also be used as the input when the M AI functions are treated as a whole.
[0241] Optionally, the first device transmits a second signal based on the second monitoring resources, and the second device receives the second signal based on the second monitoring resources.
[0242] For example, the second signal can be the output of one of the M AI functions (e.g., AI function 3) and can be used as the input of another AI function (e.g., AI function 4). AI function 3 and AI function 4 can be deployed in different devices; for example, AI function 3 can be deployed on the terminal side, and AI function 4 on the network side. For ease of illustration, step 1810 is represented by a double-headed arrow in the figure, meaning that the second signal can be sent from the second device to the first device, or vice versa.
[0243] One possibility is that all M AI functions are deployed on the terminal side, so the method also includes step 1811, whereby the first device determines the second monitoring result.
[0244] Optionally, the method further includes step 1812, whereby the first device sends the second monitoring result to the second device on the second resource if the second monitoring result meets the second condition. Accordingly, the second device receives the second monitoring result from the first device on the second resource.
[0245] Another possibility is that the M AI functions are partially deployed on the terminal side and partially deployed on the network side, so the method also includes step 1813, whereby the second device determines the second monitoring result.
[0246] For a more detailed explanation of steps 1811 to 1813, please refer to steps 1804, 1805 and 1807 above for further understanding, which will not be repeated here.
[0247] Steps 1808 to 1813 above can be regarded as the second round of performance monitoring of AI functions. The second round of performance monitoring is to monitor the performance of M AI functions out of N AI functions as a whole.
[0248] Understandably, if the second monitoring result does not meet the performance requirements, the second device can further indicate, through the sixth information, which AI functions need to be monitored in the next round of performance monitoring. This could be, for example, a subset of the M AI functions, as well as the monitoring resources for the next round of performance monitoring. This process is repeated until performance monitoring is completed for each subset of the N AI functions, or until it is determined that the performance indicators of the smallest subset of the N AI functions do not meet the performance requirements.
[0249] The method described above, which involves monitoring the overall performance of multiple AI functions in one or more rounds, can be called sequential monitoring. The AI functions monitored in each round of performance monitoring can be determined according to preset rules. For example, the AI functions monitored in the next round of performance monitoring are one or more AI functions from the multiple AI functions monitored in the previous round, excluding the AI function that is executed first or last in the execution order.
[0250] For example, the first round of performance monitoring monitors N AI functions, such as AI function 1, AI function 2, AI function 3, ..., AI function N; the second round of performance monitoring monitors (N-1) AI functions, such as AI function 2, AI function 3, ..., AI function N; the third round of performance monitoring monitors (N-2) AI functions, such as AI function 3, ..., AI function N; and so on, without further listing.
[0251] For example, the first round of performance monitoring might monitor N AI functions, such as AI function 1, AI function 2, AI function 3, ..., AI function N; the second round of performance monitoring might monitor (N-1) AI functions, such as AI function 1, AI function 2, AI function 3, ..., AI function (N-1); the third round of performance monitoring might monitor (N-2) AI functions, such as AI function 1, AI function 2, AI function 3, ..., AI function (N-2); and so on, without further listing.
[0252] For ease of understanding, Figure 19 shows an example of sequential monitoring.
[0253] First, the N AI functions can be treated as a whole for performance monitoring. A first signal on the first monitoring resource can be used to determine a first monitoring result, which can be determined based on a first performance indicator. When the first monitoring result meets the triggering condition (e.g., denoted as triggering condition N), performance monitoring is performed on a subset of the N AI functions. The figure shows the performance monitoring of AI functions 1 to AI function (N-1) among the N AI functions.
[0254] The second signal on the second monitoring resource can be used to determine the second monitoring result, which is the monitoring result of the performance monitoring of (N-1) AI functions (including AI function 1 to AI function (N-1)) as a whole. This second monitoring result can be determined based on the second performance index. When the second monitoring result meets the triggering condition (e.g., denoted as triggering condition (N-1)), the performance of some AI functions among the (N-1) AI functions is monitored. The figure shows the performance monitoring of AI function 1 to AI function (N-2) among the (N-1) AI functions.
[0255] The third signal on the third monitoring resource can be used to determine the third monitoring result. The third monitoring result is the monitoring result of the performance monitoring of (N-2) AI functions (including AI function 1 to AI function (N-2)) as a whole. This third monitoring result can be determined based on the third performance index. When the third monitoring result meets the triggering condition (e.g., denoted as triggering condition (N-2)), the performance of some AI functions among the (N-2) AI functions is monitored.
[0256] This process continues until it is determined which of the N AI functions needs parameter adjustments.
[0257] The x-th monitoring resource and x-th performance metric can be used to determine the x-th monitoring result when (N-x+1) AI functions are considered as a whole. If the x-th monitoring result satisfies the trigger condition (N-x+1), performance monitoring can be further performed on some of the (N-x+1) AI functions. It should be understood that the sequential monitoring shown in Figure 19 is determined according to the following rule: the AI function monitored in the next round of performance monitoring is one or more AI functions from the multiple AI functions monitored in the previous round of performance monitoring, excluding the AI function with the latest execution order. This should not constitute any limitation on this application. Many more possible examples can be obtained based on different rules, which are not listed here for brevity.
[0258] In another possible implementation, if the first monitoring result meets the triggering condition, the second device can also configure monitoring resources for each of the N AI functions using the fourth information. That is, if the performance monitoring of the N AI functions as a whole meets the triggering condition, performance monitoring can be performed on each AI function individually. The method of performance monitoring for each AI function can be called parallel monitoring. The specific process can be referred to in the description above combined with Figure 17, and will not be repeated here.
[0259] Based on the above scheme, since the N AI functions can be monitored as a whole, the second device can configure the first monitoring resources for the performance monitoring of the N AI functions as a whole using the first information. If the performance monitoring of the N AI functions as a whole meets the performance requirements, the subsequent configuration for performance monitoring of individual AI functions can be eliminated, and it is also unnecessary to configure monitoring resources for each AI function. This reduces signaling overhead and monitoring overhead, which is beneficial for monitoring the overall performance of the N AI functions.
[0260] To facilitate understanding, the following explanation will be provided in conjunction with the accompanying diagrams and several specific examples.
[0261] Example 1
[0262] Figure 20 is a schematic diagram illustrating the performance monitoring of N AI functions deployed on the terminal side. As shown in Figure 20, the N AI functions include AI channel estimation and AI channel prediction, which is an example where N is 2. The performance monitoring of these N AI functions can be divided into the following three stages:
[0263] Phase 1: Joint performance monitoring of N AI functions.
[0264] For example, the first monitoring resource can be used to transmit a first reference signal (i.e., an example of the first signal) and a second reference signal. The second device can transmit the first reference signal and the second reference signal based on the first monitoring resource, and the first device can receive the first reference signal and the second reference signal based on the first monitoring resource. The density of the first reference signal on the resource is lower than the density of the second reference signal on the resource. For example, the resource density of the first reference signal in at least one dimension in the time domain, frequency domain, or spatial domain is lower than the resource density of the second reference signal in the corresponding dimension. In other words, the first reference signal occupies fewer resources than the second reference signal.
[0265] Wherein, the resource density of the first reference signal in at least one dimension of the time domain, frequency domain, or spatial domain is less than the resource density of the second reference signal in the corresponding dimension, which may include one or more of the following: 1) The number of ports corresponding to the first reference signal is less than the number of ports corresponding to the second reference signal; 2) The number of REs occupied by a port of the first reference signal in one or more RBs is less than the number of REs (or subcarriers) occupied by a port of the second reference signal in the same number of one or more resource blocks (RBs); 3) The number of symbols occupied by a port of the first reference signal in one time slot is less than the number of symbols occupied by a port of the second reference signal in one time slot; 4) The transmission period of a port of the first reference signal in the time domain is less than the transmission period of a port of the second reference signal in the time domain; 5) The transmission interval of a port of the first reference signal in the time domain is greater than the transmission interval of a port of the second reference signal in the time domain; 6) The number of transmissions of a port of the first reference signal in the time domain is less than the number of transmissions of a port of the second reference signal in the time domain.
[0266] It should be understood that the first reference signal and the second reference signal can be interpreted as a low-density reference signal and a high-density reference signal, respectively. High density and low density are relative terms and do not limit the amount of resources occupied by the first reference signal and the second reference signal in the time domain, frequency domain, and spatial domain, respectively.
[0267] In Example 1, the first reference signal may serve as the input when the N AI functions are taken as a whole. Specifically, the first reference signal may be used as the input for AI channel estimation, and through processing of AI channel estimation, estimated channel information is output; the estimated channel information may be used as the input for AI channel prediction, and through processing of AI channel prediction, predicted channel information 1 is output. The predicted channel information 1 is also the output when the N AI functions are taken as a whole.
[0268] The second reference signal may be used to obtain label data. In Phase 1 of this example, the label data is channel information (for example, denoted as label channel information 1). Since the density of the second reference signal on resources is higher than the density of the first reference signal on resources, the channel information obtained by performing channel estimation based on the second reference signal may correspond to more resources, so that richer channel information can be obtained. In addition, the second reference signal may be the same in time domain as the time instant corresponding to the predicted channel information 1. If the time instant corresponding to the channel information output by the AI channel estimation is referred to as a prediction time instant, the second reference signal may be a reference signal sent at the prediction time instant.
[0269] The first apparatus may determine a first monitoring result based on the label channel information 1, the predicted channel information 1 (that is, the output when the N AI functions are taken as a whole), and a first performance indicator.
[0270] In one possible design, the first performance indicator is a difference or a degree of correlation between the label channel information 1 and the predicted channel information 1, and for example, may be MSE (predicted channel information 1, label channel information 1), SGCS (predicted channel information 1, label channel information 1), or the like.
[0271] In another possible design, the first performance indicator is a difference or a degree of correlation between an actually received second reference signal and a predicted received second reference signal determined based on the second reference signal (it should be understood that the second reference signal is known to the receiver) and the predicted channel information 1, such as MSE or SGCS. The second reference signal may be represented as S, the actually received second reference signal may be represented as Y, and the predicted channel information 1 may be represented as The predicted received second reference signal determined based on the predicted channel information 1 may be represented as and satisfies: For example, the difference between the actually received second reference signal and the predicted second reference signal (it should be understood that the predicted second reference signal may refer to the predicted received second reference signal) may be determined by or the like for representation, and the difference between the actually received second reference signal and the predicted second reference signal may be determined by or the like for representation.
[0272] Phase Two: Performance monitoring of some AI functions.
[0273] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions.
[0274] For example, Figure 20 shows an example of performance monitoring for AI channel estimation. A second monitoring resource can be used to transmit a third reference signal. A second device can send the third reference signal based on the second monitoring resource, and a first device can receive the third reference signal based on the second monitoring resource. This third reference signal is used to obtain stage two tag data (e.g., denoted as Tag Channel Information 2). A portion of the reference signal in this third reference signal (i.e., an example of the second signal) can be used as input for AI channel estimation. Through AI channel estimation processing, estimated channel information is output.
[0275] The first device can determine the second monitoring result based on tag channel information 2, estimated channel information, and a second performance indicator. For example, the second monitoring result can be the difference or correlation between the estimated channel information and tag channel information 2, such as MSE (estimated channel information, tag channel information 2), SGCS (estimated channel information, tag channel information 2), etc.
[0276] It should be understood that the third reference signal and a portion of the third reference signal can be understood as a high-density reference signal and a low-density reference signal. Specifically, "partial" can refer to a portion in at least one dimension of the time domain, frequency domain, or spatial domain.
[0277] Phase 3: Performance monitoring of another part of the AI functions.
[0278] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring has already been performed for AI channel estimation in phase two, so performance monitoring can be performed on another AI function (i.e., AI channel prediction).
[0279] For example, the third monitoring resource can be used to transmit a fourth reference signal (i.e., an example of the third signal) and a fifth reference signal. The second device can transmit the fourth and fifth reference signals based on the third monitoring resource, and the first device can receive the fourth and fifth reference signals based on the third monitoring resource. The fourth reference signal can be used as input for AI channel prediction, and the fifth reference signal can be used to obtain the tag data of stage three (e.g., denoted as tag channel information 3). The fourth and fifth reference signals can have the same density but occupy different time-domain resources. The fifth reference signal can be a reference signal transmitted at the prediction time, and the fourth reference signal can be a reference signal transmitted before the prediction time. On one hand, AI channel prediction can perform channel prediction based on the input fourth reference signal to obtain the predicted channel information 3 at the prediction time. On the other hand, tag channel information 3 can be obtained by performing channel estimation on the fifth reference signal transmitted at the prediction time. The first device can determine the third monitoring result based on the tag channel information 3, the predicted channel information 3, and the third performance index. For example, the third monitoring result is the difference or correlation between the predicted channel information 3 and the tag channel information 3, such as MSE (predicted channel information 3, tag channel information 3), SGCS (predicted channel information 3, tag channel information 3), etc.
[0280] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the parameters of the AI channel estimation may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0281] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI channel estimation. Performance monitoring of another part of AI functions (such as AI channel prediction) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e., AI functions) in Phase 2 and Phase 3 are swapped.
[0282] Example 2
[0283] Figure 21 is another schematic diagram illustrating the performance monitoring of N AI functions. As shown in Figure 21, the N AI functions include: AI channel prediction and AI channel compression and reconstruction, which is an example where N is 2. AI channel compression and reconstruction can be implemented using AI channel compression models and AI channel reconstruction models. These models can be deployed on the terminal side and the network side, respectively; for example, the AI channel compression model can be deployed on the terminal side, and the AI channel reconstruction model on the network side. Alternatively, the terminal side may also have both AI channel compression and AI channel reconstruction models simultaneously.
[0284] Performance monitoring of these N AI functions can be divided into the following three stages:
[0285] Phase 1: Joint performance monitoring of N AI functions.
[0286] The first monitoring resource can be used to transmit a first reference signal (i.e., another example of the first signal) and a second reference signal. The second device can transmit the first and second reference signals based on the first monitoring resource, and the first device can receive the first and second reference signals based on the first monitoring resource. The first reference signal can be used as input when the N AI functions are considered as a whole; in this example, the first reference signal can be used as input for AI channel prediction. After processing by AI channel prediction, predicted channel information is output. This predicted channel information is then input to AI channel compression and reconstruction. After compression and reconstruction of the channel information, reconstructed channel information 1 can be output. This reconstructed channel information 1 is also the output when the N AI functions are considered as a whole. The second reference signal can be used to obtain tag data. In stage one of this example, the tag data is the channel information at the prediction time (e.g., denoted as tag channel information 1). Therefore, the second reference signal can be a reference signal transmitted at the prediction time, and the first reference signal can be a reference signal transmitted before the prediction time; the two occupy different resources in the time domain. The relationship between the first reference signal and the second reference signal is similar to the relationship between the fourth reference signal and the fifth reference signal in stage three of Example 1 above. Please refer to the above for understanding, and it will not be repeated here.
[0287] If the AI channel compression model is deployed on the terminal side and the AI channel reconstruction model is deployed on the network side, the first monitoring result of the joint performance monitoring of the N AI functions can be obtained by the second device. At this time, the first device can also send the tag channel information 1 to the second device. If the terminal side has both the AI channel compression model and the AI channel reconstruction model, the first monitoring result of the joint performance monitoring of the N AI functions can be obtained by the first device.
[0288] The first or second device can determine the first monitoring result based on the tag channel information 1, the reconstructed channel information 1, and a first performance indicator. For example, the first performance indicator is the difference or correlation between the tag channel information 1 and the reconstructed channel information 1, such as MSE (reconstructed channel information 1, tag channel information 1), SGCS (reconstructed channel information 1, tag channel information 1), etc.
[0289] Phase Two: Performance monitoring of some AI functions.
[0290] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions.
[0291] For example, Figure 21 shows an example of performance monitoring for AI channel prediction. A second monitoring resource can be used to transmit a third reference signal (i.e., another example of the second signal) and a fourth reference signal. A second device can send the third and fourth reference signals based on the second monitoring resource, and a first device can receive the third and fourth reference signals based on the second monitoring resource. The third reference signal can be used as input for AI channel prediction, and the fourth reference signal can be used to obtain the tag data for stage two (e.g., denoted as Tag Channel Information 2). The relationship between the third and fourth reference signals is similar to the relationship between the fourth and fifth reference signals in stage three of Example 1 above, and can be understood by referring to the above; it will not be repeated here.
[0292] The first device can determine the second monitoring result based on tag channel information 2, predicted channel information, and a second performance indicator. For example, the second performance indicator is the difference or correlation between tag channel information 2 and predicted channel information, such as MSE (predicted channel information, tag channel information 2), SGCS (predicted channel information, tag channel information 2), etc.
[0293] Phase 3: Performance monitoring of another part of the AI functions.
[0294] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring has already been performed for AI channel prediction in phase two, so performance monitoring can be performed on another AI function (i.e., AI channel compression and reconstruction).
[0295] For example, the third monitoring resource can be used to transmit the fifth reference signal (i.e., another example of the third signal). The second device can send the fifth reference signal based on the third monitoring resource, and the first device can receive the fifth reference signal based on the third monitoring resource. This fifth reference signal can be used as input for AI channel compression and reconstruction, and also for obtaining stage three tag data (e.g., denoted as tag channel information 3). Specifically, by performing channel estimation on the fifth reference signal, the original channel information can be obtained. This original channel information can be used as input for AI channel compression and reconstruction. After compression and reconstruction of the channel information by AI channel compression and reconstruction, the reconstructed channel information 3 can be output. The first device can determine the third monitoring result based on the tag channel information 3, the reconstructed channel information 3, and a third performance indicator. For example, the third monitoring result is the difference or correlation between the reconstructed channel information 3 and the tag channel information 3, such as MSE (reconstructed channel information 3, tag channel information 3), SGCS (reconstructed channel information 3, tag channel information 3), etc.
[0296] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the parameters of the AI channel prediction may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0297] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI channel prediction. Performance monitoring of another part of AI functions (such as AI channel compression and reconstruction) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e. AI functions) in Phase 2 and Phase 3 are swapped.
[0298] Example 3
[0299] Figure 22 is another schematic diagram illustrating the performance monitoring of N AI functions deployed on the terminal side. As shown in Figure 22, the N AI functions include AI channel estimation and AI demodulation, which is an example where N is 2. The performance monitoring of these N AI functions can be divided into the following three stages:
[0300] Phase 1: Joint performance monitoring of N AI functions.
[0301] The first monitoring resource can be used to transmit a first reference signal and a data signal (i.e., another example of the first signal). The second device can transmit the first reference signal and the data signal based on the first monitoring resource, and the first device can receive the first reference signal and the data signal based on the first monitoring resource.
[0302] Since AI channel estimation can be used to obtain estimated channel information based on a first reference signal, and the estimated channel information can be used to equalize the received data signal, the data signal transmitted by the first monitored resource in this example can be a known data signal. Furthermore, equalization can be implemented using AI functionality or conventional algorithms; this application does not limit this approach. This example first discusses the case where equalization is implemented using conventional algorithms.
[0303] Both the first reference signal and the data signal received by the first device can be used as inputs when the N AI functions are considered as a whole. Specifically, the first reference signal can be used as input for AI channel estimation. After processing by AI channel estimation, estimated channel information is output. The estimated channel information can be used to equalize the received data signal to obtain an equalized data signal. The equalized data signal can be used as input for AI demodulation. After processing by AI demodulation, demodulated bits or the log-likelihood ratio corresponding to the demodulated bits can be output. The demodulated bits or the log-likelihood ratio corresponding to the demodulated bits output by AI demodulation are the output when the N AI functions are considered as a whole.
[0304] The first device can further process the AI demodulated output to obtain estimated bits. These estimated bits can be obtained by decoding the demodulated bits using a conventional decoding algorithm, or they can be determined based on the log-likelihood ratio corresponding to the demodulated bits. This application does not limit the specific method used.
[0305] The first device can determine a first monitoring result based on known bits and estimated bits, as well as a first performance indicator. For example, the first monitoring result is the error between the estimated bits and known bits, such as BER, MSE, or CE, such as BER(estimated bits, known bits), MSE(estimated bits, known bits), or CE(estimated bits, known bits).
[0306] It should be understood that the known bit is a data bit that the first device already knows in advance, corresponding to the data signal transmitted via the first monitoring resource mentioned above. In Phase One, the known bit can be used as tag data.
[0307] Phase Two: Performance monitoring of some AI functions.
[0308] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions.
[0309] For example, Figure 22 shows an example of performance monitoring for AI channel estimation. A second monitoring resource can be used to transmit a second reference signal. A second device can send the second reference signal based on the second monitoring resource, and a first device can receive the second reference signal based on the second monitoring resource. This second reference signal can be used to obtain stage two tag data (i.e., tag channel information). A portion of the second reference signal (i.e., another example of the second signal) can be processed by AI channel estimation to output estimated channel information as input. The first device can determine a second monitoring result based on the tag channel information, the estimated channel information, and a second performance indicator.
[0310] It should be understood that the second reference signal and the portion of the second reference signal can be understood as a high-density reference signal and a low-density reference signal. Specifically, "partial" can refer to a portion in at least one dimension of the time domain, frequency domain, or spatial domain.
[0311] Phase 3: Performance monitoring of another part of the AI functions.
[0312] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring for AI channel estimation has already been performed in phase two, so performance monitoring can be performed on another AI function (i.e., AI demodulation).
[0313] For example, a third monitoring resource can be used to transmit data signals (i.e., another example of a third signal). A second device can transmit data signals based on the third monitoring resource, and a first device can receive data signals based on the first monitoring resource. The data signal can be a known data signal, and the received data signal is equalized and then input to the AI demodulation. After AI demodulation, the bit to be decoded or the log-likelihood ratio corresponding to the bit to be decoded can be output. The first device further processes the output of the AI demodulation to obtain an estimated bit.
[0314] The first device can determine the third monitoring result based on the known bits (i.e., the tag data of stage three) and the estimated bits, as well as the third performance metric. For example, the third monitoring result can be the error between the estimated bits and the known bits, such as BER (estimated bits, known bits), MSE (estimated bits, known bits), or CE (estimated bits, known bits), etc.
[0315] It should be understood that the known bits used as tag data in Phase 3 are data bits that the first device already knew in advance, corresponding to the data signals transmitted through the third monitoring resource mentioned above.
[0316] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the parameters of the AI channel estimation may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0317] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI channel estimation. Performance monitoring of another part of AI functions (such as AI demodulation) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e. AI functions) in Phase 2 and Phase 3 are swapped.
[0318] Example 4
[0319] Example 4 is an extension of Example 3.
[0320] Optionally, the N AI functions in Example 3 may also include AI equalization. That is, the N AI functions include AI channel estimation, AI equalization, and AI demodulation. This is an example where N is 3. Performance monitoring of these N AI functions can be divided into the following four stages:
[0321] Phase 1: Joint performance monitoring of N AI functions.
[0322] Similar to Example 3, the first monitoring resource can be used to transmit a first reference signal and a data signal, the data signal being a data signal corresponding to a known bit. The second device can transmit the first reference signal and the data signal based on the first monitoring resource, and the first device can receive the first reference signal and the data signal based on the first monitoring resource. The first reference signal and the data signal received by the first device (i.e., another example of the first signal) can both be used as inputs when the N AI functions are considered as a whole. Specifically, the first reference signal can be used as input for AI channel estimation, and after processing by AI channel estimation, estimated channel information is output. The estimated channel information and the received data signal can be used as inputs for AI equalization, and after processing by AI equalization, equalized data signals are output. The equalized data signal can be used as input for AI demodulation, and after processing by AI demodulation, demodulated bits or the log-likelihood ratio corresponding to the demodulated bits can be output. The demodulated bits output by AI demodulation or the log-likelihood ratio corresponding to the demodulated bits are the output when the N AI functions are considered as a whole. It should be understood that demodulated bits are bits obtained by demodulating symbols (such as AI demodulation).
[0323] The subsequent operations are the same as in stage one of Example 3, as described above, and will not be repeated here.
[0324] Phase Two: Performance monitoring of some AI functions.
[0325] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions, such as AI channel estimation.
[0326] The second monitoring resource can be used to transmit a second reference signal. The second device can send the second reference signal based on the second monitoring resource, and the first device can receive the second reference signal based on the second monitoring resource. This second reference signal can be used to obtain stage two tag data (i.e., tag channel information). A portion of the second reference signal (i.e., another example of the second signal) can be used as input for AI channel estimation. Through AI channel estimation processing, estimated channel information is output. The first device can determine the second monitoring result based on the tag channel information, the estimated channel information, and the second performance index.
[0327] It should be understood that the second reference signal and the portion of the second reference signal can be understood as a high-density reference signal and a low-density reference signal. Specifically, "partial" can refer to a portion in at least one dimension of the time domain, frequency domain, or spatial domain.
[0328] Phase 2 can be the same as Phase 2 in Example 3 above. Please refer to Figure 22 for a schematic diagram of Phase 2. It will not be described again.
[0329] Phase 3: Performance monitoring of another part of the AI functions.
[0330] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 3, and performance monitoring for AI channel estimation has already been performed in phase two, so performance monitoring can be performed on another AI function (i.e., AI equalization and AI demodulation).
[0331] For example, a third monitoring resource can be used to transmit data signals (i.e., another example of a third signal), a second device can transmit data signals based on the third monitoring resource, and a first device can receive data signals based on the third monitoring resource. The first device can use the received data signal as input when considering AI equalization and AI demodulation as a whole, and the output bit to be decoded, or the log-likelihood ratio corresponding to the bit to be decoded, is the output when considering AI equalization and AI demodulation as a whole. The first device can further process this output to obtain estimated bits.
[0332] The first device can determine the third monitoring result based on the known bits (i.e., the tag data of stage three) and the estimated bits, as well as the third performance metric. For example, the third monitoring result is the error between the estimated bits and the known bits, such as BER (estimated bits, known bits), MSE (estimated bits, known bits), or CE (estimated bits, known bits), etc.
[0333] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the parameters of the AI channel estimation may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0334] Phase 4: Performance monitoring of some AI functions within another part of the AI functionality.
[0335] If the third monitoring result meets the performance requirements, performance monitoring can be performed on some AI functions within another part of the AI functionality. In this example, N is 3, and performance monitoring has already been performed for AI channel estimation, AI equalization, and AI demodulation in stages one and two, respectively. Therefore, performance monitoring can be performed on either AI equalization or AI demodulation. For example, performance monitoring can be performed on AI demodulation.
[0336] The process of monitoring the performance of AI demodulation can be found in the relevant instructions in Stage 3 of Example 3 above, and will not be repeated here.
[0337] Understandably, if the fourth monitoring result obtained from the performance monitoring of AI demodulation meets the performance requirements, performance monitoring of AI equalization can still be performed; if the fourth monitoring result obtained from the performance monitoring of AI demodulation does not meet the performance requirements, performance monitoring of AI equalization can be performed, or performance monitoring of AI equalization can be skipped first, and the parameters of AI demodulation can be adjusted first and then stage one can be executed again. If the first monitoring result obtained from executing stage one again meets the performance requirements, then performance monitoring of AI equalization can be discontinued.
[0338] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI channel estimation. Performance monitoring of another part of AI functions (such as AI equalization and AI demodulation) can also be performed first. The specific implementation process is similar to the example above, except that the monitored object (i.e., AI function) in Phase 2 is adjusted. The monitored object in subsequent phases can be determined based on whether the performance requirements are met by the second monitoring result. The specific method can also be referred to above, and will not be repeated here.
[0339] In addition, in some other possible implementations, AI equalization can be combined with AI channel estimation into a single function, for example, by the same AI model, for which the input can be the reference signal and data signal received on the first monitoring resource, and the output can be the equalized data signal.
[0340] Example 5
[0341] Example 5 illustrates the performance monitoring of N AI functions deployed on the terminal side. These N AI functions include AI channel estimation and AI decoding, representing an example where N is 2. Performance monitoring of these N AI functions can be divided into the following three stages.
[0342] Phase 1: Joint performance monitoring of N AI functions.
[0343] Similar to Example 3, the first monitoring resource can be used to transmit a first reference signal and a data signal (i.e., another example of a first signal), the data signal being a data signal corresponding to known bits. The second device can transmit the first reference signal and the data signal based on the first monitoring resource, and the first device can receive the first reference signal and the data signal based on the first monitoring resource.
[0344] Both the first reference signal and the received data signal can be used as inputs when the N AI functions are considered as a whole. Specifically, the first reference signal can be used as input for AI channel estimation. After processing by AI channel estimation, estimated channel information is output. The estimated channel information can be used to equalize the received data signal to obtain an equalized data signal. The equalized data signal, after demodulation, yields the bits to be decoded. These bits to be decoded can be used as input for AI decoding. After processing by AI decoding, estimated bits are output. These estimated bits are the output when the N AI functions are considered as a whole. Alternatively, these bits to be decoded can be used as input for AI decoding. After processing by AI decoding, the probability of each bit value corresponding to each bit to be decoded is output. The probability of each bit value is the output when the N AI functions are considered as a whole, and the estimated bits can be determined based on the probabilities of each bit value.
[0345] The first device can determine a first monitoring result based on known bits and estimated bits, as well as a first performance indicator. For example, the first monitoring result is the error between the estimated bits and known bits, such as BER (estimated bits, known bits), MSE (estimated bits, known bits), or CE (estimated bits, known bits), etc.
[0346] It should be understood that the known bit is a data bit that the first device already knows in advance, corresponding to the data signal transmitted via the first monitoring resource mentioned above. In Phase One, the known bit can be used as tag data.
[0347] Phase Two: Performance monitoring of some AI functions.
[0348] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions, such as AI channel estimation.
[0349] The second monitoring resource can be used to transmit a second reference signal. The second device can send the second reference signal based on the second monitoring resource, and the first device can receive the second reference signal based on the second monitoring resource. This second reference signal can be used to obtain stage two tag data (i.e., tag channel information). A portion of the second reference signal (i.e., another example of the second signal) can be used as input for AI channel estimation. Through AI channel estimation processing, estimated channel information is output. The first device can determine the second monitoring result based on the tag channel information, the estimated channel information, and the second performance index.
[0350] It should be understood that the second reference signal and the portion of the second reference signal can be understood as a high-density reference signal and a low-density reference signal. Specifically, "partial" can refer to a portion in at least one dimension of the time domain, frequency domain, or spatial domain.
[0351] Phase 2 can be the same as Phase 2 in Example 3 above. Please refer to Figure 22 for a schematic diagram of Phase 2. It will not be described again.
[0352] Phase 3: Performance monitoring that triggers another part of the AI functions.
[0353] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring has already been performed for AI channel estimation in phase two, so performance monitoring can be performed on another AI function (i.e., AI decoding).
[0354] For example, a third monitoring resource can be used to transmit data signals (i.e., another example of a third signal). A first device can transmit a data signal based on the third monitoring resource, and a second device can receive a data signal based on the third monitoring resource. This data signal can be a data signal corresponding to a known bit, which, after equalization and demodulation, becomes the bit to be decoded and is used as input to AI decoding. After AI decoding, an estimated bit can be output, or the probability of each bit value can be output. The probabilities of each bit value output by the AI decoding can be used to obtain the estimated bit.
[0355] The first device can determine the third monitoring result based on the estimated bits and known bits, as well as the third performance indicator. For example, the third monitoring result is the error between the estimated bits and known bits, such as BER (estimated bits, known bits), MSE (estimated bits, known bits), or CE (estimated bits, known bits), etc.
[0356] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the parameters of the AI channel estimation may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0357] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI channel estimation. Performance monitoring of another part of AI functions (such as AI decoding) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e. AI functions) in Phase 2 and Phase 3 are swapped.
[0358] Optionally, the N AI functions may also include AI equalization and / or AI demodulation. The N AI functions may include AI channel estimation, AI equalization, and AI decoding, i.e., an example where N is 3; or, the N AI functions may include AI channel estimation, AI demodulation, and AI decoding, i.e., another example where N is 3; or, the N AI functions may include AI channel estimation, AI equalization, AI demodulation, and AI decoding, i.e., an example where N is 4.
[0359] When N is greater than 2, performance monitoring may involve more stages, as illustrated in Example 4 above, which includes four stages. Combining the examples above, those skilled in the art can derive the process for performance monitoring of these N AI functions; for the sake of brevity, it will not be detailed further.
[0360] Example 6
[0361] Example 6 is an example of performance monitoring of one or more AI functions deployed on the network side and one or more AI functions deployed on the terminal side. The N AI functions may include one or more AI functions deployed on the network side: channel coding or modulation, and one or more AI functions deployed on the terminal side: demodulation or decoding; or, the N AI functions may include one or more AI functions deployed on the terminal side: channel coding or modulation, and one or more AI functions deployed on the network side: demodulation or decoding.
[0362] Taking N AI functions, including AI encoding and AI decoding, as an example (N=2), the AI encoding and AI decoding can be used unpaired. For example, the sending end uses AI encoding to implement the encoding function, and the receiving end uses traditional decoding methods to perform decoding; or, for another example, the sending end uses traditional encoding methods to perform encoding, and the receiving end uses AI decoding to implement the decoding function.
[0363] Performance monitoring of these N AI functions can be divided into the following three stages:
[0364] Phase 1: Joint performance monitoring of N AI functions.
[0365] For example, the second device can use known bit 1 (i.e., the tag data of phase one) as input to the AI encoding on the network side, and output encoded bits after AI encoding processing. The encoded bits are then modulated, pre-encoded, and other operations to obtain data signal 1, which is then transmitted. The first monitoring resource can be used to transmit the data signal 1 (i.e., another example of the first signal), the second device can transmit the data signal 1 based on the first monitoring resource, and the first device can receive the data signal 1 based on the first monitoring resource.
[0366] The first device can perform equalization, demodulation, and other operations on the received data signal 1 to obtain the bit to be decoded. This bit to be decoded is used as the input to the AI decoding on the terminal side. After AI decoding processing, the estimated bit 1 is output. The first device can determine a first monitoring result based on the known bit and the estimated bit, as well as a first performance index. The first monitoring result can be the error between the estimated bit 1 and the known bit 1, such as BER (estimated bit 1, known bit 1), MSE (estimated bit, known bit 1), or CE (estimated bit 1, known bit 1), etc.
[0367] Phase Two: Trigger performance monitoring of certain AI functions.
[0368] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions, such as the AI encoding.
[0369] For example, the second device can use known bit 2 (i.e., the tag data of stage two) as input to the AI encoding on the network side, and output encoded bits after AI encoding processing. The encoded bits are then modulated, pre-encoded, and other operations to obtain data signal 2, which is then transmitted. The second monitoring resource can be used to transmit this data signal 2 (i.e., another example of the second signal). The second device can transmit the data signal 2 based on the second monitoring resource, and the first device can receive the data signal 2 based on the second monitoring resource.
[0370] The first device can perform equalization, demodulation, and other operations on the received data signal 2 to obtain the bit to be decoded. This bit to be decoded can be decoded using traditional decoding methods to obtain the estimated bit 2. The first device can determine the second monitoring result based on the known bit 2, the estimated bit 2, and a second performance indicator. The second monitoring result can be the error between the estimated bit 2 and the known bit 2, such as BER (estimated bit 2, known bit 2), MSE (estimated bit 2, known bit 2), or CE (estimated bit 2, known bit 2), etc.
[0371] Phase 3: Performance monitoring of another part of the AI functions.
[0372] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring has already been performed for AI channel estimation in phase two, so performance monitoring can be performed on another AI function (i.e., AI decoding).
[0373] For example, the second device can encode the known bit 3 (i.e., the tag data of stage three) using a conventional encoding method and output the encoded bit. After modulation, precoding, and other operations, the encoded bit becomes data signal 3, which is then transmitted. The third monitoring resource can be used to transmit the data signal 3. The second device can transmit the data signal 3 based on the third monitoring resource, and the first device can receive the data signal 3 based on the third monitoring resource.
[0374] The first device can perform equalization, demodulation, and other operations on the received data signal 3 to obtain the bit to be decoded. This bit to be decoded can be used as input for AI decoding. After AI decoding, the estimated bit 3 is output. The first device can determine the third monitoring result based on the known bit 3, the estimated bit 3, and a third performance indicator. The third monitoring result can be the error between the estimated bit 3 and the known bit 3, such as BER (estimated bit 3, known bit 3), MSE (estimated bit 3, known bit 3), or CE (estimated bit 3, known bit 3), etc.
[0375] It should be understood that the tag data in Phase 1, Phase 2 and Phase 3 are all known bits, namely known bit 1, known bit 2 and known bit 3, respectively. They can be the same or different, and this application does not limit them.
[0376] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the AI code parameters may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0377] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI coding. Performance monitoring of another part of the AI functions (such as AI decoding) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e., AI functions) in Phase 2 and Phase 3 are swapped.
[0378] It should also be understood that the above description uses the example of AI encoding being deployed on the network side and AI decoding being deployed on the terminal side. When AI encoding can also be deployed on the terminal side and AI decoding can also be deployed on the network side, the operations of the first and second devices described above can be interchanged. For example, in stage one, the first device can send data signal 1 based on the first monitoring resource. This data signal 1 can be obtained by processing the encoded bits output by the AI encoding deployed on the terminal side based on known bit 1, and the second device can receive data signal 1 based on the first monitoring resource. In stage two, the first device can send data signal 2 based on the second monitoring resource. This data signal 2 can be obtained by processing the encoded bits output by the AI encoding deployed on the terminal side based on known bit 2, and the second device can receive data signal 2 based on the second monitoring resource. In stage three, the first device can send data signal 3 based on the third monitoring resource. This data signal 3 can be obtained by the terminal side after performing operations such as encoding, modulation, and precoding based on known bit 3, and the second device can receive this data signal 3 based on the third monitoring resource.
[0379] Example 7
[0380] Example 7 illustrates performance monitoring of an AI function deployed on the network side and an AI function deployed on the terminal side. The N AI functions include AI reference signal optimization and AI channel estimation. Specifically, AI reference signal optimization can be deployed on the network side, and AI channel estimation can be deployed on the terminal side, or vice versa.
[0381] Performance monitoring of these N AI functions can be divided into the following three stages:
[0382] Phase 1: Joint performance monitoring of N AI functions.
[0383] For example, the AI reference signal optimization on the network side can output a first reference signal (i.e., an example of a first signal), and a first monitoring resource can be used to transmit the first reference signal output by the AI reference signal optimization. A second device can transmit the first reference signal based on the first monitoring resource, and a first device can receive the first reference signal based on the first monitoring resource. Optionally, one or more of the following information can be input into the AI reference signal optimization: the density of the reference signal, historical channel information, or preset values, to obtain the output first reference signal. The preset values can refer to preset parameters of the neural network.
[0384] The first device can input the received first reference signal into the AI channel estimation at the terminal side, process the AI channel estimation, and output the estimated channel information 1. This estimated channel information 1 can be the output of the N AI functions as a whole.
[0385] On the other hand, the second device can transmit a conventional reference signal, and the first device can use a conventional channel estimation algorithm to perform channel estimation on the received conventional reference signal to obtain tag channel information 1.
[0386] The first device can determine a first monitoring result based on tag channel information 1, estimated channel information 1, and a first performance indicator. This first monitoring result can be the difference or correlation between the estimated channel information 1 and tag channel information 1, such as MSE (estimated channel information, tag channel information 1), SGCS (estimated channel information 1, tag channel information 1), etc. Alternatively, the first monitoring result can also be the difference or correlation between the received conventional reference signal and the estimated reference signal determined based on the estimated channel information 1, such as MSE or SGCS. Here, the received conventional reference signal can be represented as S, and the estimated channel information 1 can be represented as... The estimated reference signal determined based on the estimated channel information 1 can be expressed as: satisfy: For example, the difference between the received conventional reference signal and the estimated reference signal can be expressed as: or
[0387] Phase Two: Performance monitoring of some AI functions.
[0388] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions, such as AI channel estimation.
[0389] For example, the second monitoring resource can be used to transmit the second reference signal, the second device can send the second reference signal based on the second monitoring resource, and the first device can receive the second reference signal based on the second monitoring resource. The second reference signal is used to obtain the tag data of stage two (e.g., denoted as tag channel information 2). A portion of the reference signal in the second reference signal (i.e., another example of the second signal) can be used as input for AI channel estimation. Through the processing of AI channel estimation, the estimated channel information 2 is output.
[0390] The first device can determine the second monitoring result based on the tag channel information 2, the estimated channel information 2, and the second performance index. For example, the second monitoring result can be the difference or correlation between the estimated channel information 2 and the tag channel information 2, such as MSE (estimated channel information 2, tag channel information 2), SGCS (estimated channel information 2, tag channel information 2), etc.
[0391] It should be understood that the second reference signal and the portion of the second reference signal can be understood as a high-density reference signal and a low-density reference signal. Specifically, "partial" can refer to a portion in at least one dimension of the time domain, frequency domain, or spatial domain.
[0392] The process of monitoring the performance of AI channel estimation can be seen in Figure 20, which illustrates stage two, and will not be described in detail here.
[0393] Phase 3: Performance monitoring that triggers another part of the AI functions.
[0394] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring has already been performed for AI channel estimation in phase two, so performance monitoring can be performed on another AI function (i.e., AI reference signal optimization).
[0395] For example, the AI reference signal optimization on the network side can output a third reference signal (i.e., an example of a third signal). The third monitoring resource is similar to the first monitoring resource and can be used to transmit the third reference signal output by the AI reference signal optimization. The second device can transmit the third reference signal based on the third monitoring resource, and the first device can receive the third reference signal based on the third monitoring resource. Optionally, one or more of the following information can be input into the AI reference signal optimization: the density of the reference signal, historical channel information, or a preset value. At the transmitting end (e.g., the network side), one or more of the above information can be input into the AI reference signal optimization. Based on the input, the AI reference signal optimization can output a reference signal. The reference signal output by the AI reference signal optimization can be sent to the first device, and the first device can use a conventional channel estimation algorithm to obtain the estimated channel information 3.
[0396] On the other hand, the transmitting end can transmit a traditional reference signal, and the receiving end (such as the terminal side) can use a traditional channel estimation algorithm to perform channel estimation on the received traditional reference signal to obtain the tag channel information 3.
[0397] The first device can determine a third monitoring result based on tag channel information 3, estimated channel information 3, and a third performance indicator. This third monitoring result can be the difference or correlation between the estimated channel information 3 and the tag channel information 3, such as MSE (estimated channel information 3, tag channel information 3), SGCS (estimated channel information 3, tag channel information 3), etc. Alternatively, the third monitoring result can also be the difference or correlation between the received conventional reference signal and the estimated reference signal determined based on the estimated channel information 3, such as... or Where S3 represents the received conventional reference signal, This represents the estimated reference signal determined based on the estimated channel information 3.
[0398] Understandably, if the first monitoring result obtained in Phase 1 meets the performance requirements, Phase 2 may not be executed. If the second monitoring result obtained in Phase 2 does not meet the performance requirements, Phase 3 may be executed, or it may not be executed first. Instead, the parameters of the AI channel estimation may be adjusted and Phase 1 may be executed again. If the first monitoring result obtained from executing Phase 1 again still does not meet the performance requirements, Phase 3 may be executed. If the first monitoring result obtained from executing Phase 1 again meets the performance requirements, Phase 3 may not be executed.
[0399] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI channel estimation. Performance monitoring of another part of AI functions (such as AI reference signal optimization) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e. AI functions) in Phase 2 and Phase 3 are swapped.
[0400] It should also be understood that the above description uses the example of AI reference signal optimization being deployed on the network side and AI channel estimation being deployed on the terminal side. When AI reference signal optimization can also be deployed on the terminal side and AI channel estimation can be deployed on the network side, the operations of the first device and the second device can be interchanged. For example, in stage one, the first device can send a first reference signal based on the first monitoring resource, which is a reference signal output by AI reference signal optimization, and the second device can receive the first reference signal based on the first monitoring resource; in stage two, the first device can send a second reference signal based on the second monitoring resource, and the second device can receive the second reference signal based on the second monitoring resource; in stage three, the first device can send a third reference signal based on the third monitoring resource, which is a reference signal output by AI reference signal optimization, and the second device can receive the third reference signal based on the third monitoring resource.
[0401] Example 8
[0402] Example 8 is an example of performance monitoring for N AI functions deployed on the network side. The N AI functions include AI encoding and AI modulation, i.e., an example where N is 2. Performance monitoring of these N AI functions can be divided into the following three stages:
[0403] Phase 1: Joint performance monitoring of N AI functions.
[0404] For example, the first monitoring resource can be used to transmit a data signal, which is a known bit 1 that can be used as tag data.
[0405] On one hand, the second device can use known bit 1 (i.e., stage one) as input to the AI encoding on the network side, and after AI encoding processing, output encoded bits. The encoded bits can be used as input to the AI modulation on the network side, and after AI modulation processing, output modulation symbols. The AI modulation symbols are processed (e.g., pre-coding) to obtain data signal 1, which is then transmitted. The first monitoring resource can be used to transmit this data signal 1, the second device can transmit this data signal 1 based on the first monitoring resource, and the first device can receive this data signal 1 based on the first monitoring resource.
[0406] The first device can process the received data signal 1 using conventional demodulation methods and decoding algorithms to obtain an estimated bit 1. Based on this estimated bit 1, the known bit 1, and a first performance indicator, the first device can determine a first monitoring result. This first monitoring result can be, for example, the error between the estimated bit 1 and the known bit 1, such as BER (estimated bit 1, known bit 1), MSE (estimated bit 1, known bit 1), or CE (estimated bit 1, known bit 1), etc.
[0407] Phase Two: Trigger performance monitoring of certain AI functions.
[0408] If the initial monitoring result does not meet the performance requirements, performance monitoring can be performed on a subset of the N AI functions. In this example, N is 2, so performance monitoring can be performed on one of the two AI functions, such as the AI encoding.
[0409] For example, the second device can use known bit 2 (i.e., the tag data of stage two) as input to the AI encoding on the network side, and output encoded bits after AI encoding processing. The encoded bits are then modulated, pre-encoded, and other operations to obtain data signal 2, which is then transmitted. The second monitoring resource can be used to transmit this data signal 2 (i.e., another example of the second signal). The second device can transmit the data signal 2 based on the second monitoring resource, and the first device can receive the data signal 2 based on the second monitoring resource.
[0410] The first device can perform equalization, demodulation, and other operations on the received data signal 2 to obtain the bit to be decoded. This bit to be decoded can be decoded using traditional decoding methods to obtain the estimated bit 2. The first device can determine the second monitoring result based on the known bit 2, the estimated bit 2, and a second performance indicator. The second monitoring result can be the error between the estimated bit 2 and the known bit 2, such as BER (estimated bit 2, known bit 2), MSE (estimated bit 2, known bit 2), or CE (estimated bit 2, known bit 2), etc.
[0411] Phase 3: Performance monitoring of another part of the AI functions.
[0412] If the second monitoring result meets the performance requirements, performance monitoring can be performed on another part of the N AI functions. In this example, N is 2, and performance monitoring has already been performed on AI channel coding in phase two, so performance monitoring can be performed on another AI function (i.e., AI modulation).
[0413] For example, the second device can use the known bit 3 as input to AI modulation on the network side, process it through AI modulation, and output a modulation symbol. The modulation symbol can be processed to obtain a data signal 3 (i.e., an example of a third signal) which is then transmitted. A third monitoring resource can be used to transmit the data signal 3. The second device can transmit the data signal 3 based on the third monitoring resource, and the first device can receive the data signal 3 based on the third monitoring resource.
[0414] The first device can process the received data signal 3 to obtain an estimated bit 3. Based on the estimated bit 3, the known bit 3, and a third performance indicator, the first device can determine a third monitoring result. This third monitoring result can be, for example, the error between the estimated bit 3 and the known bit 3, such as BER (estimated bit 3, known bit 3), MSE (estimated bit 3, known bit 3), or CE (estimated bit 3, known bit 3), etc. It is understood that if the first monitoring result obtained in stage one meets the performance requirements, stage two may not be executed; if the second monitoring result obtained in stage two does not meet the performance requirements, stage three may be executed, or it may not be executed initially, but instead, the AI encoding parameters may be adjusted and stage one executed again. If the first monitoring result obtained from executing stage one again still does not meet the performance requirements, stage three can be executed; if the first monitoring result obtained from executing stage one again meets the performance requirements, stage three may not be executed.
[0415] It should also be understood that the performance monitoring of some AI functions in Phase 2 does not necessarily have to be AI coding. Performance monitoring of another part of AI functions (such as AI modulation) can also be performed first. The specific implementation process is similar to the example above, except that the monitored objects (i.e. AI functions) in Phase 2 and Phase 3 are swapped.
[0416] It should be noted that in the examples above, the monitoring resources are used to transmit data signals, which are data signals corresponding to known bits. The relationship between this data signal and the known bits can be understood as follows: the known bits are bits that the first and second devices know in advance. For example, they can be bits generated based on predefined rules. For instance, the known bits can be some or all of the bits in a random bit sequence generated according to a random number seed and generation algorithm; or, for example, the known bits can be some or all of the bits in a pseudo-random (PN) sequence, and so on, without limitation. The data signal corresponding to this known bit refers to the data signal that the second device encodes, modulates, etc., the known bit, and then transmits, and which is received by the first device. The estimated bits obtained by demodulating and decoding this data signal should be the same as or similar to the known bits.
[0417] The foregoing examples provide possible implementations for performance monitoring of N different AI functions. These examples are provided for ease of understanding only and should not be construed as limiting this application. In general, performance monitoring of these N AI functions may involve the following possible scenarios:
[0418] 1. Some or all of the N AI functions are deployed on the transmitter on the terminal side. The first device can send reference signals or tag data, and the second device can perform performance monitoring based on the received signals.
[0419] 2. Some or all of the N AI functions are deployed on the transmitter on the network side. The second device can send reference signals or tag data, and the first device can perform performance monitoring based on the received signals.
[0420] 3. Some or all of the N AI functions are deployed on the receiver on the terminal side. The second device sends reference signals or tag data, and the first device can perform performance monitoring based on the received signals.
[0421] 4. Some or all of the N AI functions are deployed on the receiver on the network side. The first device can send reference signals or tag data, and the second device can perform performance monitoring based on the received signals.
[0422] The methods provided in the embodiments of this application have been described in detail above with reference to several accompanying drawings. The apparatus provided in the embodiments of this application will now be described with reference to the accompanying drawings.
[0423] Figures 23 to 25 are schematic block diagrams of possible apparatuses provided in embodiments of this application. These apparatuses can be used to implement the functions of the first or second apparatus in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments.
[0424] Figures 23 and 24 are schematic diagrams of possible communication devices provided in the embodiments of this application. These communication devices can be used to implement the functions of the terminal devices and network devices in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments. In the embodiments of this application, the communication device can be the access network device as shown in Figure 1 or Figure 2, the core network device as shown in Figure 3, the access network node as shown in Figure 4, or the terminal device as shown in Figure 1 or Figure 2, or the terminal as shown in Figure 3 or Figure 4.
[0425] As shown in Figure 23, the communication device 2300 includes a processing module 2310 and a communication module 2320. The communication device 2300 is used to implement the functions of the terminal device in the method embodiment shown in Figure 18 above.
[0426] In an optional embodiment, the communication device 2300 can be used to implement the functions of the first device in the embodiment shown in FIG18.
[0427] When the communication device 2300 is used to implement the function of the first device in the embodiment shown in FIG18: the communication module 2320 is used to receive first information, the first information is used to indicate the first monitoring resources configured for N AI functions, where N is a positive integer greater than 1; the processing module 2310 is used to send or receive a first signal based on the first monitoring resources, the first signal is used to obtain a first monitoring result for performance monitoring of the N AI functions as a whole.
[0428] Optionally, the processing module 2310 is also used to determine the first monitoring result.
[0429] Optionally, the communication module 2320 is also used to send the first monitoring result if the first monitoring result meets the first condition.
[0430] Optionally, the communication module 2320 is further configured to receive second information, which indicates one or more of the following: a first performance indicator, a first condition, or a first resource; the first performance indicator is a performance indicator on which the performance monitoring of N AI functions as a whole is based, and the first resource is a resource used to transmit the first monitoring result.
[0431] Optionally, the communication module 2320 is also used to receive second information, the third information being used to indicate N AI functions.
[0432] Optionally, the communication module 2320 is further configured to receive fourth information, which indicates M AI functions out of N AI functions, and a second monitoring resource configured for the M AI functions, the second monitoring resource being used to obtain a second monitoring result for performance monitoring of the M AI functions, where M is a positive integer less than N; and to send or receive a second signal based on the second monitoring resource, the second signal being used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole.
[0433] Optionally, the communication module 2320 is further configured to receive fifth information, which indicates one or more of the following: a second performance indicator, a second condition, a second resource, or a triggering condition; the second performance indicator is the performance indicator on which the performance monitoring of the M AI functions as a whole is based; the second condition is the condition satisfied for sending the second monitoring result; the second resource is the resource used to transmit the second monitoring result; and the triggering condition is the condition for triggering the performance monitoring of the M AI functions as a whole.
[0434] In another alternative embodiment, the communication device 2300 can be used to implement the function of the second device in the embodiment shown in FIG18.
[0435] When the communication device 2300 is used to implement the function of the second device in the embodiment shown in FIG18: the communication module 2320 is used to send first information, the first information is used to indicate the first monitoring resources configured for N AI functions, the first monitoring resources are used to obtain the first monitoring result of the performance monitoring of the N AI functions, where N is a positive integer greater than 1; it is also used to send or receive a first signal based on the first monitoring resources, the first signal is used to obtain the first monitoring result of the performance monitoring of the N AI functions as a whole.
[0436] .
[0437] Optionally, the communication module 2320 is also used to receive the first monitoring result.
[0438] Optionally, the processing module 2310 is also used to determine the first monitoring result.
[0439] Optionally, the communication module 2320 is further configured to send second information, which indicates one or more of the following: a first performance indicator, a first condition, or a first resource; the first performance indicator is a performance indicator on which the performance monitoring of N AI functions as a whole is based, and the first resource is a resource used to transmit the first monitoring result.
[0440] Optionally, the communication module 2320 is also used to send third information, which is used to indicate N AI functions.
[0441] Optionally, the communication module 2320 is further configured to send fourth information, which indicates M AI functions out of N AI functions, and a second monitoring resource configured for the M AI functions. The second monitoring resource is configured to obtain a second monitoring result for performance monitoring of the M AI functions, where M is a positive integer less than N. The communication module 2320 is also configured to send or receive a second signal based on the second monitoring resource. The second signal is configured to obtain a second monitoring result for performance monitoring of the M AI functions as a whole.
[0442] Optionally, the processing module 2310 is further configured to determine, if the first monitoring result does not meet the performance requirements, to perform performance monitoring on M of the N AI functions.
[0443] Optionally, the communication module 2320 is further configured to send fifth information, which indicates one or more of the following: a second performance indicator, a second condition, a second resource, or a triggering condition; the second performance indicator is the performance indicator on which the performance monitoring of the M AI functions as a whole is based; the second condition is the condition satisfied for sending the second monitoring result; the second resource is the resource used to transmit the second monitoring result; and the triggering condition is the condition for triggering the performance monitoring of the M AI functions as a whole.
[0444] For a more detailed description of the processing module 2310 and the communication module 2320 described above, please refer to the relevant description in the method embodiment shown in FIG18.
[0445] It should be noted that the communication module can also be called a transceiver module, transceiver unit, transceiver, transceiver device, or transceiver apparatus, etc. The processing module can also be called a processor, processing board, processing unit, or processing apparatus, etc. Optionally, the communication module is used to execute the sending and receiving operations of the first or second communication device in the above method. The device in the communication module that implements the receiving function can be considered as the receiving module, and the device in the communication module that implements the sending function can be considered as the sending module; that is, the communication module can include both a receiving module and a sending module.
[0446] It should also be noted that, in one possible design, the aforementioned processing module and / or communication module can be implemented through virtual modules. For example, the processing module can be implemented through software functional units or virtual devices, and the communication module can be implemented through software functions or virtual devices. In another possible design, the processing module or communication module can also be implemented through physical devices. For example, if the device is implemented using a chip / chip circuit, the communication module can be an input / output circuit and / or a communication interface, performing input operations (corresponding to the aforementioned receiving operation) and output operations (corresponding to the aforementioned sending operation); the processing module can be an integrated processor, a microprocessor, or an integrated circuit.
[0447] The module division in this embodiment is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional modules in the various examples of this embodiment can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.
[0448] Figure 24 is a schematic diagram of a communication device provided in another embodiment of this application. As shown in Figure 24, the communication device 2400 includes a processing circuit 2410 and a communication circuit 2420. The processing circuit 2410 and the communication circuit 2420 are coupled to each other. The processing circuit 2410 may be one or more processors, or all or part of the circuitry in one or more processors used for control or processing functions. It is understood that when the communication device 2400 is a network device or a terminal device, the communication circuit 2420 may be a transceiver circuit, a transceiver, a communication interface, or an input / output interface. When the communication device 2400 is a chip for a network device or a terminal device, the communication circuit 2420 may be an input / output interface, a communication interface, or an input / output circuit. Optionally, the communication device 2400 may further include a memory 2430 for storing instructions executed by the processor, or storing input data required by the processor to execute instructions, or storing data generated after the processor executes instructions.
[0449] When the communication device 2400 is used to implement the method shown in FIG18, the processing circuit 2410 is used to implement the function of the processing module 2310, and the communication circuit 2420 is used to implement the function of the communication module 2320.
[0450] When the aforementioned communication device is a chip applied to a terminal device, the chip implements the functions of the terminal device in the above method embodiments. The chip can receive information from other modules (such as a radio frequency module or antenna) in the terminal device, information sent by the network device to the terminal device; or, the chip can send information to other modules (such as a radio frequency module or antenna) in the terminal device, information sent by the terminal device to the network device.
[0451] When the aforementioned communication device is a module applied to a network device (such as a base station), the module implements the functions of the network device in the above method embodiments. This module can receive information from other modules (such as radio frequency modules or antennas) in the network device, and this information may be sent from the terminal device to the network device; alternatively, the module can send information to other modules (such as radio frequency modules or antennas) in the network device, and this information may be sent from the network device to the terminal device. The module here can be the baseband chip of the base station, or a DU (Digital Unit) or other modules. The DU here can be a DU under an Open Radio Access Network (O-RAN) architecture.
[0452] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), graphics processing units (GPUs), neural processing units (NPUs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.
[0453] For example, the processor used for AI can be one or more of the following: graphics processing unit (GPU), neural processing unit (NPU), tensor processing unit (TPU), and data processing unit (DPU).
[0454] For example, one possible implementation of a processor for AI could be the AI processor 2500 shown in Figure 25. Figure 25 is a schematic diagram of the structure of an AI processor provided in an embodiment of this application.
[0455] As shown in Figure 25, the AI processor 2500 may include one or more of the following: an AI core, a digital vision pre-processing (DVPP) module, a task scheduler (TS), an L3 cache, an AI CPU, a control CPU, an L2 cache, a universal serial bus (USB) interface, a network interface card (NIC), a peripheral component interconnect express (PCIe) interface (PCIe is a high-speed serial computer expansion bus standard), a double data rate (DDR) / high bandwidth memory (HBM) interface, a general purpose input / output (GPIO) / inter-integrated circuit (I2C) bus, etc. It is understood that the specific meanings of these terms are well known to those skilled in the art and will not be elaborated upon here.
[0456] The terms “unit”, “module”, etc., used in this specification may be used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution.
[0457] The method steps in the embodiments of this application can be implemented in hardware or in software instructions executable by a processor. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, compact disc read-only memory (CD-ROM), or any other form of storage medium well known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. The storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in a base station or terminal. The processor and storage medium can also exist as discrete components in the base station or terminal.
[0458] This application also provides a communication system, which includes the first device and the second device described in the embodiments of this application.
[0459] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that a computing device can store, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state drive). The computer-readable storage medium includes instructions or program code that, when executed on a computing device, cause the computing device to perform the methods provided above.
[0460] This application also provides a computer program product, which may be a software or program product containing instructions capable of running on a computing device or stored on any usable medium. When the instructions are executed on the computing device, the computing device performs the methods provided above, or performs the functions of the apparatus provided above.
[0461] This application also provides a chip including at least one processor, which, when program instructions are executed by the at least one processor, causes the at least one processor to perform the method executed by the first device or the method executed by the second device in the above method embodiments.
[0462] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0463] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0464] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0465] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0466] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0467] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0468] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A communication method, characterized in that, include: Receive first information, the first information being used to indicate a first monitoring resource configured for N AI functions, where N is a positive integer greater than 1; Based on the first monitoring resource, a first signal is sent or received, and the first signal is used to obtain a first monitoring result for performance monitoring of the N AI functions as a whole.
2. The method as described in claim 1, characterized in that, The first signal is used as input when the N AI functions are considered as a whole, and the first monitoring result is determined based on the output when the N AI functions are considered as a whole.
3. The method as described in claim 1 or 2, characterized in that, The N AI functions are deployed on the terminal side, and the method further includes: Determine the first monitoring result.
4. The method as described in claim 3, characterized in that, The method further includes: If the first monitoring result meets the first condition, the first monitoring result is sent.
5. The method as described in claim 4, characterized in that, The method further includes: Receive second information, which indicates one or more of the following: a first performance indicator, a first condition, or a first resource; the first performance indicator is a performance indicator on which the performance monitoring of the N AI functions as a whole is based, and the first resource is a resource used to transmit the first monitoring result.
6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Receive third information, which is used to indicate the N AI functions.
7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Receive fourth information, the fourth information being used to indicate M AI functions out of the N AI functions, and a second monitoring resource configured for the M AI functions, wherein M is a positive integer less than N; Based on the second monitoring resource, a second signal is sent or received, and the second signal is used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole.
8. The method as described in claim 7, characterized in that, The second signal is used as the input when the M AI functions are considered as a whole, and the second monitoring result is determined based on the output when the M AI functions are considered as a whole.
9. The method as described in claim 7 or 8, characterized in that, The method further includes: Receive fifth information, the fifth information being used to indicate one or more of the following: a second performance indicator, a second condition, a second resource, or a triggering condition; the second performance indicator is the performance indicator on which the performance monitoring of the M AI functions as a whole is based; the second condition is the condition satisfied for sending the second monitoring result; the second resource is the resource used to transmit the second monitoring result; the triggering condition is the condition for triggering the performance monitoring of the M AI functions as a whole.
10. The method according to any one of claims 1 to 9, characterized in that, The N AI functions include: a first AI function for channel estimation and a second AI function for channel prediction; the first monitoring resource is used to transmit a first reference signal and a second reference signal, wherein the density of the first reference signal on the resource is lower than the density of the second reference signal on the resource; the first reference signal is the input of the N AI functions as a whole, and the output of the N AI functions as a whole is the predicted channel information; the first monitoring result is determined based on the predicted channel information and the channel information determined by the second reference signal, or the first monitoring result is determined based on the predicted received signal of the second reference signal and the received signal of the second reference signal, wherein the predicted received signal of the second reference signal is determined based on the predicted channel information.
11. The method according to any one of claims 1 to 9, characterized in that, The N AI functions include: a first AI function for channel prediction and a second AI function for channel compression and reconstruction; the first monitoring resource is used to transmit a first reference signal and a second reference signal; the first reference signal is the input when the N AI functions are used as a whole, and the output when the N AI functions are used as a whole is the reconstructed channel information; the first monitoring result is determined based on the reconstructed channel information and the channel information determined by the second reference signal.
12. The method according to any one of claims 1 to 9, characterized in that, The N AI functions include: a first AI function for channel estimation and a second AI function for demodulation; the first monitoring resource is used to transmit a first reference signal, which is the input of the N AI functions as a whole; the output of the N AI functions as a whole is at least one estimated demodulated bit or at least one likelihood log-likelihood ratio corresponding to the at least one demodulated bit; the first monitoring result is determined based on the estimated bit and tag data, wherein the estimated bit is obtained by decoding the at least one demodulated bit or determined according to the at least one log-likelihood ratio, and the tag data is a known bit.
13. The method as described in claim 12, characterized in that, The N AI functions also include a third AI function for balancing.
14. The method according to any one of claims 1 to 9, characterized in that, The N AI functions include: a first AI function for channel estimation and a second AI function for decoding; the first monitoring resource is used to transmit a first reference signal, which is the input of the N AI functions as a whole; the first monitoring result is determined based on estimated bits and tag data, where the tag data is a known bit sequence, and the estimated bits are the output of the N AI functions as a whole, or the estimated bits are determined based on the probability of the bit values output by the N AI functions as a whole.
15. The method as described in claim 14, characterized in that, The N AI functions also include a third AI function for equalization and / or a fourth AI function for demodulation.
16. A communication method, characterized in that, include: Send a first message, which is used to indicate a first monitoring resource configured for N AI functions, where N is a positive integer greater than 1; Based on the first monitoring resource, a first signal is sent or received, and the first signal is used to obtain a first monitoring result for performance monitoring of the N AI functions as a whole.
17. The method as described in claim 16, characterized in that, The first signal is used as input when the N AI functions are considered as a whole, and the first monitoring result is determined based on the output when the N AI functions are considered as a whole.
18. The method as described in claim 16 or 17, characterized in that, The method further includes: Receive the first monitoring result.
19. The method as described in claim 18, characterized in that, The method further includes: Send a second message, which indicates one or more of the following: a first performance indicator, a first condition, or a first resource; the first performance indicator is the performance indicator on which the performance monitoring of the N AI functions as a whole is based; the first condition is the condition that must be met to transmit the first monitoring result; and the first resource is the resource used to transmit the first monitoring result.
20. The method as described in claim 16 or 17, characterized in that, Some of the N AI functions are deployed on the network side, and the method further includes: Determine the first monitoring result.
21. The method according to any one of claims 16 to 20, characterized in that, The method further includes: Send a third message, which is used to instruct the N AI functions.
22. The method according to any one of claims 16 to 21, characterized in that, The method further includes: Send a fourth message, which is used to indicate M AI functions out of the N AI functions, and a second monitoring resource configured for the M AI functions, wherein M is a positive integer less than N; Based on the second monitoring resource, the second signal is sent or received, and the second signal is used to obtain a second monitoring result for performance monitoring of the M AI functions as a whole.
23. The method as described in claim 22, characterized in that, The second signal is used as the input when the M AI functions are considered as a whole, and the second monitoring result is determined based on the output when the M AI functions are considered as a whole.
24. The method as described in claim 22 or 23, characterized in that, Before sending the fourth message, the method further includes: If the first monitoring result does not meet the performance requirements, it is determined that performance monitoring will be performed on M of the N AI functions.
25. The method according to any one of claims 22 to 24, characterized in that, The method further includes: Send a fifth message, which indicates one or more of the following: a second performance indicator, a second condition, a second resource, or a triggering condition; the second performance indicator is the performance indicator on which the performance monitoring of the M AI functions as a whole is based; the second condition is the condition satisfied for transmitting the second monitoring result; the second resource is the resource used for transmitting the second monitoring result; the triggering condition is the condition satisfied for triggering the performance monitoring of the M AI functions as a whole.
26. The method according to any one of claims 16 to 25, characterized in that, The N AI functions include: a first AI function for channel estimation and a second AI function for channel prediction; the first monitoring resource is used to transmit a first reference signal and a second reference signal, wherein the density of the first reference signal on the resource is lower than the density of the second reference signal on the resource; the first reference signal is the input of the N AI functions as a whole, and the output of the N AI functions as a whole is the predicted channel information; the first monitoring result is determined based on the predicted channel information and the channel information determined by the second reference signal, or the first monitoring result is determined based on the predicted received signal of the second reference signal and the received signal of the second reference signal, wherein the predicted received signal of the second reference signal is determined based on the predicted channel information.
27. The method according to any one of claims 16 to 25, characterized in that, The N AI functions include: a first AI function for channel prediction and a second AI function for channel compression and reconstruction; the first monitoring resource is used to transmit a first reference signal and a second reference signal; the first reference signal is the input when the N AI functions are used as a whole, and the output when the N AI functions are used as a whole is the reconstructed channel information; the first monitoring result is determined based on the reconstructed channel information and the channel information determined by the second reference signal.
28. The method according to any one of claims 16 to 25, characterized in that, The N AI functions include: a first AI function for channel estimation and a second AI function for demodulation; the first monitoring resource is used to transmit a first reference signal, which is the input of the N AI functions as a whole; the output of the N AI functions as a whole is at least one estimated demodulated bit or at least one likelihood log-likelihood ratio corresponding to the at least one demodulated bit; the first monitoring result is determined based on the estimated bit and tag data, wherein the estimated bit is obtained by decoding the at least one demodulated bit or determined according to the at least one log-likelihood ratio, and the tag data is a known bit.
29. The method as described in claim 28, characterized in that, The N AI functions also include a third AI function for balancing.
30. The method according to any one of claims 16 to 25, characterized in that, The N AI functions include: a first AI function for channel estimation and a second AI function for decoding; the first monitoring resource is used to transmit a first reference signal, which is the input of the N AI functions as a whole; the first monitoring result is determined based on estimated bits and tag data, where the tag data is a known bit sequence, and the estimated bits are the output of the N AI functions as a whole, or the estimated bits are determined based on the probability of the bit values output by the N AI functions as a whole.
31. The method as described in claim 30, characterized in that, The N AI functions also include a third AI function for equalization and / or a fourth AI function for demodulation.
32. A communication device, characterized in that, It includes modules or units for implementing the method as described in any one of claims 1 to 15, or modules or units for implementing the method as described in any one of claims 16 to 31.
33. A communication device, characterized in that, It includes one or more processors and communication circuitry, the communication circuitry being used for at least one of inputting or outputting signals; the one or more processors being used to implement the method as described in any one of claims 1 to 15, or to implement the method as described in any one of claims 16 to 31.
34. The apparatus as claimed in claim 33, characterized in that, The communication device is a network device or a terminal device, or a chip used in the network device or the terminal device.
35. A communication system, characterized in that, include: A first apparatus for performing the method as described in any one of claims 1 to 15; The second apparatus is used to perform the method as described in any one of claims 16 to 31.
36. A readable storage medium, characterized in that, Used to store programs or instructions that, when executed, are implemented in accordance with the method of any one of claims 1 to 15, or in accordance with any one of claims 16 to 31.
37. A computer program product, characterized in that, Includes a program or instructions that, when executed, implement the method as claimed in any one of claims 1 to 15, or implement the method as claimed in any one of claims 16 to 31.
