Communication method and related apparatus

Abstract

A communication method and a related apparatus. The communication method comprises: a model usage side performs channel estimation and equalization processing on the basis of an AI channel estimation model to obtain a symbol S1, and performs channel estimation and equalization processing on the basis of a non-AI channel estimation algorithm to obtain a symbol S2; and the model usage side determines a performance indicator of the AI channel estimation model on the basis of a degree of dispersion between the symbol S1 and a corresponding transmitted symbol and a degree of dispersion between the symbol S2 and a corresponding transmitted symbol.

Description

Communication methods and related devices

[0001] This application claims priority to Chinese Patent Application No. 202411990834.1, filed with the China National Intellectual Property Administration on December 30, 2024, entitled "Communication Method and Related Apparatus", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of communication technology, and in particular to a communication method and related apparatus. Background Technology

[0003] In communication systems, reference signals are a crucial component of system design, primarily used for functions such as channel state measurement, data demodulation, and beam training. With the development of artificial intelligence (AI), these functions can be implemented by combining AI models. For example, by using an AI model in conjunction with a demodulation reference signal (DMRS) to output the channel response estimating the time-frequency resources occupied by the data, and then using this channel response for data demodulation, this AI model can be called an AI channel estimation model.

[0004] Machine learning is an important technological approach to achieving artificial intelligence (AI). Before an AI model is run, its performance is judged by observing its behavior on a pre-provided dataset (collected from historical / non-real-world environments). After training on a static dataset (i.e., training data), the model is then put into dynamic / real-world, constantly changing scenarios / data for inference tasks. Because of the difference between the static data used during training and the dynamically changing data used in actual applications, the model's performance may degrade over time.

[0005] Therefore, how to synchronously conduct model performance evaluation on a running model is an urgent problem to be solved. Summary of the Invention

[0006] This application provides a communication method and related apparatus for evaluating the performance of an AI channel estimation model in operation.

[0007] Firstly, this application provides a communication method. This method can be applied to a first device, which is a device on the AI ​​model user side, such as a terminal or network device deploying the AI ​​model. The method can also be applied to a chip or chip module in the first device, or to a module or unit capable of realizing all or part of the functions of the first device. Taking the first device as an example, the method includes:

[0008] The first device performs channel estimation and equalization processing on the first received signal based on the AI ​​channel estimation model to obtain symbol S1; it performs channel estimation and equalization processing on the second received signal based on the non-AI channel estimation algorithm to obtain symbol S2; based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol, it determines the performance index of the AI ​​channel estimation model; the first transmitted symbol and the second transmitted symbol have the same modulation method.

[0009] Based on this method, since the transmitted symbol is a true value, the closer the estimated channel is to the actual channel, the closer the equalized symbol should be to the transmitted symbol under the same equalization algorithm. Since the first and second transmitted symbols have the same modulation scheme, comparing the dispersion between symbol S1 and the first transmitted symbol with the dispersion between symbol S2 and the second transmitted symbol can determine the performance metrics of the AI ​​channel estimation model, such as whether it outperforms the performance of non-AI channel estimation algorithms.

[0010] Furthermore, based on this method, the true value of the transmitted symbol is used as the comparison object to measure the model performance. Compared with using the channel estimation results of non-AI channel estimation algorithms as the comparison object, if the performance of non-AI channel estimation algorithms is not as good as that of AI channel estimation models, it will lead to incorrect evaluation results of the performance of AI channel estimation models. This method can improve the accuracy of performance evaluation.

[0011] In one optional implementation, the first device determines the performance index of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol, including: calculating a first parameter value based on symbol S1 and the first transmitted symbol; and calculating a second parameter value based on symbol S2 and the second transmitted symbol; the first parameter value and the second parameter value are used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol; and the performance index of the AI ​​channel estimation model is determined based on the first parameter value and the second parameter value.

[0012] Based on this method, the first device uses parameter values ​​to indicate the degree of dispersion between the equalized symbol and the transmitted symbol, thereby facilitating the determination of the performance index of the AI ​​channel estimation model.

[0013] Optionally, in this embodiment, the first parameter value is the Euclidean distance between symbol S1 and the first transmitted symbol; the second parameter value is the Euclidean distance between symbol S2 and the second transmitted symbol.

[0014] In one optional implementation, the first device determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the ratio between the first parameter value and the second parameter value is less than or equal to a first threshold, the first device determines that the AI ​​channel estimation model does not need to be updated; if the ratio between the first parameter value and the second parameter value is greater than the first threshold, the first device determines that the AI ​​channel estimation model needs to be updated.

[0015] Based on this method, if the ratio between the first parameter value and the second parameter value is less than or equal to the first threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is lower than the dispersion between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running normally, its performance has not degraded, and no model update is needed. If the ratio between the first parameter value and the second parameter value is greater than the first threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is higher than the dispersion between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is worse than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running abnormally, its performance has degraded, and a model update is needed.

[0016] In another optional implementation, the first device determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the difference between the first parameter value and the second parameter value is less than or equal to a second threshold, it is determined that the AI ​​channel estimation model does not need to be updated; if the difference between the first parameter value and the second parameter value is greater than the second threshold, it is determined that the AI ​​channel estimation model needs to be updated.

[0017] Based on this method, if the difference between the first parameter value and the second parameter value is less than or equal to the second threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is lower than the dispersion between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running normally, its performance has not degraded, and no model update is needed. If the difference between the first parameter value and the second parameter value is greater than the second threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is higher than that between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is worse than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running abnormally, its performance has degraded, and a model update is needed.

[0018] In another optional implementation, the first device determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: the first device determines that the AI ​​channel estimation model does not need to be updated if the first parameter value is less than or equal to a third threshold and the ratio between the first parameter value and the second parameter value is less than or equal to the first threshold; the first device determines that the AI ​​channel estimation model needs to be updated if the first parameter value is greater than the third threshold or the ratio between the first parameter value and the second parameter value is greater than the first threshold.

[0019] Based on this method, the first device also needs to consider the relationship between the first parameter value and the third threshold to avoid the situation where the dispersion indicated by the first parameter value is too large, even if the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model may run abnormally.

[0020] In another optional implementation, the first device determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: the first device determines that the AI ​​channel estimation model does not need to be updated if the first parameter value is less than or equal to a third threshold and the difference between the first parameter value and the second parameter value is less than or equal to the second threshold; the first device determines that the AI ​​channel estimation model needs to be updated if the first parameter value is greater than the third threshold or the difference between the first parameter value and the second parameter value is greater than the second threshold.

[0021] Based on this method, the first device also needs to consider the relationship between the first parameter value and the third threshold to avoid the situation where the dispersion indicated by the first parameter value is too large, even if the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model may run abnormally.

[0022] In one optional implementation, the first device further receives first configuration information, which indicates a calculation method and a corresponding threshold between a first parameter value and a second parameter value. Alternatively, the first configuration information may indicate not only the first or second threshold corresponding to the calculation method, but also a third threshold, which is used to evaluate the performance of the AI ​​channel estimation model in conjunction with the first parameter value.

[0023] Based on this method, the first device can determine the performance of the AI ​​channel estimation model according to the calculation method indicated by the first configuration information and the corresponding threshold.

[0024] Optionally, the threshold corresponding to the calculation method between the first and second parameter values ​​is related to the modulation scheme of the transmitted symbol. For example, different modulation schemes may correspond to different thresholds. Similarly, the third threshold may also be related to the modulation scheme of the transmitted symbol.

[0025] In one optional implementation, the first device further receives second configuration information, which is used to indicate a first resource for channel estimation based on an AI channel estimation model and a second resource for channel estimation based on a non-AI channel estimation algorithm.

[0026] Based on this method, the first device can perform channel estimation on the first resource based on the AI ​​channel estimation model, and equalize the signal received on the first resource based on the channel estimation result to obtain symbol S1; the first device can perform channel estimation on the second resource based on the non-AI channel estimation algorithm, and equalize the signal received on the second resource based on the channel estimation result to obtain symbol S2.

[0027] Optionally, the first resource and the second resource are the same. Correspondingly, the first transmitted symbol and the second transmitted symbol are also the same. Alternatively, the first resource and the second resource are different; correspondingly, the first transmitted symbol and the second transmitted symbol are signals mapped to their respective resources, and can be the same or different. Optionally, the second configuration information can configure a resource pool for performance evaluation, which is used for channel estimation during performance evaluation. The first resource and the second resource can be determined from this resource pool.

[0028] Optionally, the first resource can be a reference signal resource configured for use in the AI ​​channel estimation model; the second resource can be a reference signal resource specifically configured for non-AI channel estimation.

[0029] Based on this method, when non-AI channel estimation fails to estimate an effective channel on the reference signal resources used in AI channel estimation, additional reference signal resources can be configured for non-AI channel estimation. For example, in cases where reference signal resources are superimposed on resource units to improve resource utilization, additional reference signal resources need to be configured or the system needs to revert to a non-superimposed state for non-AI channel estimation.

[0030] In one optional implementation, the first device further reports performance metrics of the AI ​​channel estimation model, including whether the AI ​​channel estimation model does not need updating or needs updating. Optionally, if the AI ​​channel estimation model does not need updating, the first device may not report performance metrics; if an update is required, the performance metrics will then be reported.

[0031] Based on this method, the first device can report the performance indicators of the AI ​​channel estimation model to the network side or the core network side, which is conducive to timely updating of the AI ​​channel estimation model and improving the accuracy of channel estimation.

[0032] Secondly, this application also provides a communication method, which can be applied to a second device, a side-to-side device for using AI models, such as a network device that distributes AI models to terminals, or a core network device that distributes AI models to network devices. This method can also be applied to chips or chip modules in the second device, or to modules or units that can implement all or part of the functions of the second device. The method is described using a second device as an example. In this method:

[0033] The second device determines and transmits a first transmitted symbol and a second transmitted symbol based on the same modulation scheme. The second device receives the performance index of the AI ​​channel estimation model. This performance index is determined based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol. Symbol S1 is obtained by performing channel estimation and equalization processing on the first received signal based on the AI ​​channel estimation model; symbol S2 is obtained by performing channel estimation and equalization processing on the second received signal based on a non-AI channel estimation algorithm.

[0034] Based on this method, the performance index of the AI ​​channel estimation model is determined by comparing the dispersion between symbol S1 and the first transmitted symbol with the dispersion between symbol S2 and the second transmitted symbol. The performance index of the AI ​​channel estimation model can also be determined based on the constellation point positions of the equalized symbols.

[0035] Furthermore, based on this method, the performance metrics of the AI ​​channel estimation model are obtained by using the true value of the transmitted symbol as the comparison object. Compared with using the channel estimation results of non-AI channel estimation algorithms as the comparison object, if the performance of non-AI channel estimation algorithms is not as good as that of AI channel estimation models, it will lead to incorrect evaluation results of the performance of AI channel estimation models. This method can improve the accuracy of performance evaluation.

[0036] In one optional embodiment, the second device further transmits first configuration information, which indicates the calculation method between the first parameter value and the second parameter value, as well as the corresponding threshold. Optionally, the first parameter value and the second parameter value are used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. Optionally, the first parameter value is the Euclidean distance between symbol S1 and the first transmitted symbol; the second parameter value is the Euclidean distance between symbol S2 and the second transmitted symbol.

[0037] Based on this method, it is beneficial for the model user to evaluate the performance of the AI ​​channel estimation model based on the calculation method and the corresponding threshold.

[0038] In one optional implementation, the second device further sends second configuration information, which indicates a first resource for channel estimation based on an AI channel estimation model and a second resource for channel estimation based on a non-AI channel estimation algorithm. For details regarding the first and second resources, please refer to the relevant content described in the first aspect, and they will not be elaborated upon here.

[0039] Based on this method, it is beneficial for the model user to evaluate the performance of the AI ​​channel estimation model based on the first and second resources.

[0040] Other alternative implementations or beneficial effects of this communication method can be found in the relevant content described in the third aspect, and will not be detailed here.

[0041] Thirdly, this application also provides a communication method, which can be applied to a first device, which is a device on the AI ​​model user side, such as a terminal or network device deploying the AI ​​model. The method can also be applied to a chip or chip module in the first device, or to a module or unit capable of realizing all or part of the functions of the first device. Taking the first device as an example, in this method:

[0042] The first device determines first configuration information at a first moment. This first configuration information indicates the measured value of a second parameter corresponding to the measured value range of the first parameter. The first parameter indicates channel quality. The second parameter, determined based on an AI channel estimation model, indicates the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. Based on the AI ​​channel estimation model and the first configuration information, the first device determines the measured value range of the corresponding first parameter and the measured value of the second parameter at a second moment. Based on the measured value of the second parameter at the second moment and the measured value of the second parameter at the first moment, the first device determines the performance index of the AI ​​channel estimation model.

[0043] Based on this method, the first device takes the measurement value of the second parameter corresponding to the measurement range of the first parameter indicated by the first configuration information as the comparison object. Based on the measurement range of the first parameter and the measurement value of the second parameter at the second time, it learns the change in the degree of dispersion between the equalized symbol determined by the AI ​​channel estimation model and the corresponding transmitted symbol, so as to determine the performance index of the AI ​​channel estimation model.

[0044] This method uses the degree of dispersion between the equalized symbol and the transmitted symbol determined by the AI ​​channel estimation model at a previous time as the comparison object, avoiding the channel estimation results of non-AI channel estimation algorithms as the comparison object. If the performance of non-AI channel estimation algorithms is not as good as that of AI channel estimation models, it will lead to incorrect evaluation results of the performance of AI channel estimation models, thus improving the accuracy of performance evaluation.

[0045] Furthermore, based on this method, there is no need to use a non-AI channel estimation algorithm for channel estimation. Therefore, there is no need to configure additional reference signal resources for non-AI channel estimation, thereby improving resource utilization.

[0046] For example, if the measured value of the second parameter at the second time step is not equal to (e.g., greater than) the indicated measured value of the second parameter, it indicates that the dispersion between the symbol and the corresponding transmitted symbol after equalization at the second time step has changed (e.g., more dispersed than at the first time step), indicating that the performance of the AI ​​channel estimation model has decreased compared to the first time step, and a model update is required. If the measured value of the second parameter at the second time step is equal to or less than the indicated measured value of the second parameter, it indicates that the dispersion between the symbol and the corresponding transmitted symbol after equalization at the second time step remains unchanged or is less dispersed, indicating that the performance of the AI ​​channel estimation model remains unchanged or improves compared to the first time step, and no model update is required.

[0047] For example, the first configuration information indicates the measurement range of the second parameter. If the measured value of the second parameter at the second time is not included in the measurement range of the second parameter (e.g., the measured value of the second parameter at the second time is greater than the maximum measured value in the measurement range of the second parameter), it indicates that the dispersion between the symbol and the corresponding transmitted symbol after equalization at the second time has changed (e.g., it is more dispersed than at the first time), indicating that the performance of the AI ​​channel estimation model has decreased compared to the first time, and a model update is required. If the measured value of the second parameter at the second time is included in the measurement range of the second parameter, it indicates that the dispersion between the symbol and the corresponding transmitted symbol after equalization at the second time remains unchanged or is smaller, indicating that the performance of the AI ​​channel estimation model remains unchanged or improves compared to the first time, and no model update is required.

[0048] Optionally, the second moment is a moment following the first moment. The first moment can be the initial deployment moment of the AI ​​channel estimation model, the moment the AI ​​channel estimation model is distributed, the moment the first configuration information is received, or the moment the AI ​​channel estimation model is triggered to be used, etc. The second moment is a moment during the operation or performance evaluation of the AI ​​channel estimation model, or the moment the first information is received, which is used to instruct the AI ​​channel estimation model to perform a performance evaluation.

[0049] Optionally, the first moment can be the initial deployment of the AI ​​channel estimation model; the second moment can be during the operation or performance evaluation of the AI ​​channel estimation model.

[0050] Optionally, the first parameter can be one of the following: reference signal received power (RSRP), reference signal received quality (RSRQ), signal to interference & noise ratio (SINR), received signal strength indicator (RSSI), etc.

[0051] Optionally, the second parameter can be the mean square error (MSE) or mean absolute error (MAE) between the equalized symbol and the corresponding transmitted symbol.

[0052] In one optional implementation, the first device determines first configuration information at a first moment, including: the first device estimating a range of measured values ​​for a first parameter based on a reference signal at the first moment; corresponding to the range of measured values ​​for the first parameter, the first device calculates the MSE or MAE value between the equalized symbol and the corresponding transmitted symbol, and uses this as the measured value of a second parameter corresponding to the range of measured values. Wherein, as the value of the first parameter increases, the measured value of the second parameter decreases.

[0053] Based on this method, the first device can interact with the other side to establish a correspondence between the measurement range of the first parameter and the measurement value of the second parameter.

[0054] In another optional implementation, the first device determines the first configuration information at a first moment, including: the first device receives the first configuration information from the opposite side at the first moment, the first configuration information being used to indicate the measurement value of the second parameter corresponding to the measurement value range of the first parameter.

[0055] Based on this method, the first device can directly obtain the correspondence between the measured value range of the first parameter and the measured value of the second parameter from the other side.

[0056] In one optional implementation, the first device further reports performance metrics of the AI ​​channel estimation model, including whether the AI ​​channel estimation model needs updating or requires updating. Optionally, if the AI ​​channel estimation model does not need updating, then performance metrics do not need to be reported.

[0057] Based on this method, the first device can report the performance indicators of the AI ​​channel estimation model to the network side or the core network side, which is conducive to timely updating of the AI ​​channel estimation model and improving the accuracy of channel estimation.

[0058] In one alternative implementation, the first device receives second configuration information, which is used to indicate the equalization algorithm for determining the second parameter.

[0059] Based on this method, the measured value of the second parameter indicated by the first configuration information and the measured value of the second parameter at the second time point are calculated using the same equalization algorithm, thereby improving the accuracy of model performance evaluation.

[0060] In one optional implementation, the first configuration information is further used to configure the measurement value of the second parameter corresponding to the measurement range of the first parameter under different equalization algorithms.

[0061] Based on this method, the first device can select the measurement value of the second parameter corresponding to the measurement range of the first parameter under its own balancing algorithm, and perform performance evaluation to improve the accuracy of model performance evaluation.

[0062] Fourthly, this application also provides a communication method, which can be applied to a second device, a side-to-side device for using AI models, such as a network device that distributes AI models to terminals, or a core network device that distributes AI models to network devices. This method can also be applied to chips or chip modules in the second device, or to modules or units that can implement all or part of the functions of the second device. The method is described using a second device as an example. In this method:

[0063] The second device transmits first configuration information at a first moment. This first configuration information indicates the measured value of a second parameter corresponding to the measurement range of the first parameter. The first parameter indicates channel quality; the second parameter, determined based on an artificial intelligence (AI) channel estimation model, indicates the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. The measured value of the second parameter at the first moment is used to determine the performance index of the AI ​​channel estimation model at a second moment corresponding to the measurement range of the first parameter; the second device receives the performance index of the AI ​​channel estimation model. The performance index includes whether the AI ​​channel estimation model needs updating or not.

[0064] Based on this method, the model user uses the measurement value of the second parameter corresponding to the measurement range of the first parameter at the first time as the comparison object. Based on the measurement range of the first parameter and the measurement value of the second parameter at the second time, the performance index of the AI ​​channel estimation model is determined.

[0065] This method allows the model user to use the degree of dispersion between the equalized symbol and the transmitted symbol determined by the AI ​​channel estimation model at a previous time as the comparison object, avoiding the channel estimation results of non-AI channel estimation algorithms as the comparison object. If the performance of non-AI channel estimation algorithms is not as good as that of AI channel estimation models, it will lead to incorrect evaluation results of the performance of AI channel estimation models, thus improving the accuracy of performance evaluation.

[0066] Furthermore, based on this method, there is no need to use the channel estimation results of non-AI channel estimation algorithms as comparison objects. Therefore, the second device does not need to be additionally configured with reference signal resources for non-AI channel estimation, thereby improving resource utilization.

[0067] In one optional implementation, the second device further sends second configuration information, which is used to indicate the equalization algorithm for determining the second parameter.

[0068] In one optional implementation, the first configuration information is further used to configure the measurement value of the second parameter corresponding to the measurement range of the first parameter under different equalization algorithms.

[0069] Other alternative implementations or beneficial effects of this communication method can be found in the relevant content described in the third aspect, and will not be detailed here.

[0070] Fifthly, embodiments of this application also provide a communication device. This communication device is the first device described in the foregoing aspects, or a device capable of being used in conjunction with the first device. In one possible implementation, the communication device includes a functional module, which is either hardware circuitry, software, or a combination of hardware circuitry and software.

[0071] In one possible implementation, the communication device includes one or more functional units, such as a processing unit, wherein the processing unit is used to perform channel estimation and equalization processing on a first received signal based on an AI channel estimation model to obtain symbol S1; to perform channel estimation and equalization processing on a second received signal based on a non-AI channel estimation algorithm to obtain symbol S2; and to determine the performance index of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol; the first transmitted symbol and the second transmitted symbol have the same modulation scheme.

[0072] Optionally, possible implementations of the communication device can be found in the relevant description in the first aspect, and will not be detailed here.

[0073] In another possible implementation, in this communication device, the processing unit is configured to determine first configuration information at a first moment, the first configuration information indicating the measured value of a second parameter corresponding to the measured value range of a first parameter; the first parameter indicates channel quality; the second parameter is determined based on an AI channel estimation model and indicates the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. The processing unit is further configured to determine the measured value range of the corresponding first parameter and the measured value of the second parameter at a second moment based on the AI ​​channel estimation model and the first configuration information. The processing unit is also configured to determine the performance index of the AI ​​channel estimation model based on the measured value of the second parameter at the second moment and the measured value of the second parameter at the first moment.

[0074] Optionally, possible implementations of the communication device can be found in the relevant description in the third aspect, which will not be detailed here.

[0075] Sixthly, embodiments of this application also provide a communication device. This communication device is the second device described in the foregoing aspects, or a device capable of being used in conjunction with the second device. In one possible implementation, the communication device includes a functional module, which is either hardware circuitry, software, or a combination of hardware circuitry and software.

[0076] In one possible implementation, the communication device includes one or more functional units, such as a communication unit and a processing unit. The processing unit determines a first transmitted symbol and a second transmitted symbol based on the same modulation scheme. The communication unit transmits the first transmitted symbol and the second transmitted symbol. The communication unit also receives a performance index of an AI channel estimation model. This performance index is determined based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol. Symbol S1 is obtained by performing channel estimation and equalization processing on the first received signal based on the AI ​​channel estimation model; symbol S2 is obtained by performing channel estimation and equalization processing on the second received signal based on a non-AI channel estimation algorithm.

[0077] Optionally, possible implementations of the communication device can be found in the relevant description in the second aspect, and will not be detailed here.

[0078] In another possible implementation, the communication device includes a communication unit that transmits first configuration information at a first moment. This first configuration information indicates the measured value of a second parameter corresponding to the measured value range of a first parameter. The first parameter indicates channel quality; the second parameter, determined based on an artificial intelligence (AI) channel estimation model, indicates the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. The measured value of the second parameter at the first moment is used to determine the performance index of the AI ​​channel estimation model at a second moment corresponding to the measured value range of the first parameter. The communication unit also receives the performance index of the AI ​​channel estimation model. The performance index includes whether the AI ​​channel estimation model needs updating or not.

[0079] Optionally, possible implementations of the communication device can be found in the relevant description in the fourth aspect, which will not be detailed here.

[0080] For the fifth and sixth aspects, as examples, the processing unit can be a processing unit or can be embodied as a processing circuit or logic circuit; the communication unit can be an input / output interface, interface circuit, output circuit, input circuit, pin or related circuit on the chip or chip system.

[0081] In implementation, the processor can be used for, but is not limited to, baseband-related processing, and the transceiver or communication interface can be used for, but is not limited to, radio frequency transceiver. These devices can be disposed on separate chips, or at least partially or entirely on the same chip. For example, the processor can be further divided into analog baseband processors and digital baseband processors. The analog baseband processor can be integrated with the transceiver (or communication interface) on the same chip, while the digital baseband processor can be disposed on a separate chip. With the continuous development of integrated circuit technology, more and more devices can be integrated on the same chip. For example, a digital baseband processor can be integrated with multiple application processors (e.g., but not limited to graphics processors, multimedia processors, etc.) on the same chip. Such a chip can be called a System on a Chip (SoC). Whether the devices are disposed independently on different chips or integrated on one or more chips often depends on the needs of the product design. This application does not limit the implementation form of the above-mentioned devices.

[0082] In a seventh aspect, this application provides a communication device that may include a processing circuit and a transceiver circuit connected together. The transceiver circuit is used for exchanging (or sending, receiving, or inputting / outputting) information or data, and the processing circuit is used for executing program instructions to cause the communication device to perform the methods described in any possible embodiment of the first, second, third, or fourth aspect above. The transceiver circuit may be a communication interface, an input / output interface, or a transceiver. The transceiver may be a radio frequency module in the communication device, or a combination of a radio frequency module and an antenna. The transceiver circuit may be an input / output interface of a chip or circuit.

[0083] Eighthly, this application provides a communication device including a processor for executing the methods described in any possible implementation of the first, second, third, fourth, or any of the above-described aspects. Alternatively, the processor is configured to execute a program stored in a memory, wherein when the program is executed, the methods described in any possible implementation of the first, second, third, fourth, or any of the above-described aspects are executed.

[0084] In one possible implementation, the memory is located outside the aforementioned communication device.

[0085] In one possible implementation, the memory is located within the aforementioned communication device.

[0086] In one possible implementation, the processor and memory can also be integrated into a single device; that is, the processor and memory can be integrated together. For example, the communication device can be a chip or a chip system.

[0087] In one possible implementation, the communication device further includes a transceiver for receiving or transmitting first information. Exemplarily, the transceiver can also be used to receive or transmit a reference signal. Exemplarily, the communication device can be a terminal device or a network device.

[0088] Ninthly, this application provides a computer-readable storage medium storing program instructions that, when executed on a computer, cause the computer to perform the method described in any possible implementation of the first aspect, the second aspect, the third aspect, the fourth aspect, or any of the aspects.

[0089] In a tenth aspect, this application provides a program product containing program instructions that, when executed, causes the method described in any possible implementation of the first aspect, or the second aspect, or the third aspect, or the fourth aspect, or any of the aspects, to be performed.

[0090] Eleventhly, this application provides an apparatus, which can be implemented in the form of a chip or a device, including a processing circuit. The processing circuit is used to read and execute a program stored in a memory to execute one or more of the communication methods provided in the first aspect, the second aspect, the third aspect, the fourth aspect, or any possible implementation thereof. Optionally, the apparatus further includes a memory connected to the processing circuit via a circuit. Further optionally, the apparatus includes a communication interface connected to the processing circuit. The communication interface is used to receive information to be processed, the processing circuit obtains the information from the communication interface, processes the information, and outputs the processing result through the communication interface. The communication interface can be an input / output interface.

[0091] Optionally, the aforementioned processing circuitry and memory can be physically independent units, or the memory can be integrated with the processing circuitry.

[0092] In a twelfth aspect, this application provides a communication system comprising a first communication device and a second communication device; the first communication device is configured to perform the method described in any possible implementation of the first aspect, the third aspect, or any of the above aspects, and the second communication device is configured to perform the method described in any possible implementation of the second aspect, the fourth aspect, or any of the above aspects. Attached Figure Description

[0093] Figure 1 is a schematic diagram of a communication system;

[0094] Figure 2 is a schematic diagram of another communication system;

[0095] Figure 3 is a schematic diagram of a possible application framework in a communication system;

[0096] Figure 4 is a schematic diagram of another possible application framework in a communication system;

[0097] Figure 5 is a schematic diagram of a possible DMRS pattern;

[0098] Figure 6 is a schematic diagram of a neural network;

[0099] Figure 7 is a schematic diagram of the structure of a neuron;

[0100] Figure 8 is a schematic diagram of the main process of a communication system;

[0101] Figure 9 is a constellation diagram of 16QAM and QPSK;

[0102] Figure 10 is a schematic diagram of the performance evaluation of a DMRS-based model;

[0103] Figure 11 is a schematic diagram of model performance evaluation provided in an embodiment of this application;

[0104] Figure 12 is a flowchart of a communication method 100 provided in an embodiment of this application;

[0105] Figure 13 is another flowchart of the communication method 100 provided in an embodiment of this application;

[0106] Figure 14 is a flowchart of a communication method 200 provided in an embodiment of this application;

[0107] Figure 15 is a schematic diagram of the ORAN implementation provided in the embodiments of this application;

[0108] Figure 16 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0109] Figure 17 is a schematic diagram of the structure of a terminal provided in an embodiment of this application. Detailed Implementation

[0110] To facilitate a clear description of the technical solutions of the embodiments of this application, the following points will be explained before introducing the solutions of this application.

[0111] (1) "And / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone, where A and B can be singular or plural. In the description of this application, unless otherwise stated, "multiple" means two or more. "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 represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0112] (2) “Instruction” can include direct instruction, indirect instruction, explicit instruction, and implicit instruction. When describing a certain instruction information or when an instruction is used to instruct A, it can be understood that the instruction information carries A, directly instructs A, or indirectly instructs A.

[0113] The instruction information, or the information that the instruction indicates, is called the instruction-to-instruction information. In practical implementation, there are many ways to instruct the instruction-to-instruction information, such as, but not limited to, directly instructing the instruction-to-instruction information itself or its index. It can also indirectly instruct the instruction-to-instruction information by instructing other information, where there is a correlation between the other information and the instruction-to-instruction information. Furthermore, it can instruct only a part of the instruction-to-instruction information, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing instruction overhead to some extent. In addition, the instruction-to-instruction information can be sent as a whole or divided into multiple sub-information pieces, and the sending period and / or timing of these sub-information pieces can be the same or different.

[0114] (3) "Send" and "receive" indicate the direction of signal transmission. For example, "send information to XX" can be understood as the destination of the information being XX, which can include sending directly through the air interface or sending indirectly through the air interface by other units or modules. "receive information from YY" can be understood as the source of the information being YY, which can include receiving directly from YY through the air interface or receiving indirectly from YY through the air interface by other units or modules.

[0115] "Sending" can also be understood as the "output" of a chip interface, and "receiving" can be understood as the "input" of a chip interface. In other words, sending and receiving can occur between devices, such as between network devices and terminal devices, or within a device, such as between components, modules, chips, software modules, or hardware modules within the device via buses, traces, or interfaces. Furthermore, unless otherwise specified, "transmission" includes receiving and / or sending. For example, transmitting signals can include receiving signals and / or sending signals.

[0116] For example, in this application, "sending information" can be understood as one device sending information to another device, or it can also be understood as one logical module within a device sending information to another logical module. For example, "terminal sending information" can be understood as a terminal sending information to another device (such as an access network device), or it can be understood as logical module 1 in the terminal sending information to logical module 2 in the terminal.

[0117] For example, in this application, "receiving information" can be understood as one device receiving information from another device, or it can also be understood as a logical module within a device receiving information from another logical module. For example, "terminal device receiving information" can be understood as a terminal device receiving information from another device (such as a network device), or it can be understood as logical module 1 in the terminal receiving information from logical module 2 in the network device.

[0118] In this application, phrases such as "sending information to... (e.g., a terminal)" or related illustrations in the accompanying drawings can be understood as indicating that the destination of the information is an access network device. This can include sending information directly or indirectly to an access network device. Similarly, phrases such as "receiving information from... (e.g., a network device)," "receiving information from... (e.g., a network device)," or "receiving information sent by (e.g., a network device)," or related illustrations in the accompanying drawings, can be understood as indicating that the source of the information is a network device. This can include receiving information directly or indirectly from a network device. Information may undergo necessary processing between the source and destination, such as format changes, but the destination can understand the valid information from the source. Similar expressions in this application can be interpreted similarly and will not be elaborated further here.

[0119] (4) Information C is used to determine information D, which includes both information D being determined solely based on information C and information D being determined based on information C and other information. In addition, information C can also be used to determine information D indirectly, for example, information D is determined based on information E, and information E is determined based on information C.

[0120] (5) "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product or device.

[0121] (6) In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terms and / or descriptions of different embodiments are consistent and can be referenced by each other. The technical features of different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0122] (7) In this application, "first" and "second" are used for convenience of description to distinguish objects and are not intended to limit the scope of the embodiments of this application, nor are they used to describe the order or sequence of features. It should be understood that the objects described in this way can be interchanged where appropriate so as to describe solutions other than those in the embodiments of this application.

[0123] (8) The words “exemplary” or “for example” are used to indicate that they are examples, illustrations, or descriptions. Any embodiment or design that is described as “exemplary” or “for example” in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words “exemplary” or “for example” is intended to present the relevant concepts in a specific manner.

[0124] (9) "Information", such as first information, can be a message or the content of a message.

[0125] (10) “Network device sends to terminal”, correspondingly, “terminal receives from network device” or “terminal receives from network device”; similarly, “network device receives from terminal”, correspondingly, “terminal sends to network device” or “network device receives from terminal”, which will not be elaborated here.

[0126] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0127] I. Communication System

[0128] The technical solutions provided in this application can be applied to various communication systems, such as: 5th generation (5G) or new radio (NR) systems, long term evolution (LTE) systems, LTE frequency division duplex (FDD) systems, LTE 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.

[0129] In a communication system, one 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, equipment, communication device, communication module, node, communication node, etc. This application describes the device as an example. For instance, the first device is a terminal device, and the second device is a network device. The 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.

[0130] In wireless communication networks, such as mobile communication networks, the services supported by the networks are becoming increasingly diverse, and the requirements they need to meet are also becoming more varied. For example, networks need to support ultra-high speeds, ultra-low latency, and / or massive connectivity. These requirements make network planning, network configuration, and / or resource scheduling increasingly complex. To meet this challenge, 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 into the network.

[0131] For example, Figure 1 is a schematic diagram of a communication system. As shown in Figure 1, the communication system may include at least one network device, such as network device 110 shown in Figure 1; the communication system may also include at least one terminal device, such as terminal device 120 and terminal device 130 shown in Figure 1. Network device 110 and terminal devices (such as terminal devices 120 and 130) can communicate via a wireless link. The communication devices in this communication system, for example, network device 110 and terminal device 120, can communicate via multi-antenna technology.

[0132] As an example, a terminal device is a device that provides voice or data and has wireless connectivity. A terminal device can be called a terminal, user equipment (UE), mobile station (MS), mobile terminal (MT), etc., and can be a device with wireless transceiver capabilities. It can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; it can also be deployed on water (such as on ships); and it can also be deployed in the air (e.g., on airplanes, balloons, and satellites). Terminal devices can be used to connect people, objects, and machines. Terminal devices can be widely used in various scenarios, such as cellular communication, device-to-device (D2D), vehicle-to-everything (V2X), peer-to-peer (P2P), machine-to-machine (M2M), machine-type communication (MTC), Internet of Things (IoT), virtual reality (VR), augmented reality (AR), industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, smart homes, remote sensing, passive sensing, positioning, navigation, autonomous delivery and mobility, etc.

[0133] As examples, terminal devices can be UEs conforming to the 3rd Generation Partnership Project (3GPP) standards, fixed devices, mobile devices, handheld devices, wearable devices, cellular phones, smartphones, Session Initiation Protocol (SIP) phones, tablets, laptops, PDAs, personal computers, mobile internet devices (MIDs), VR devices, AR devices, smart books, vehicles, satellites, Global Positioning System (GPS) devices, drones, robots, helicopters, aircraft, ships, remote control devices, wireless terminals or industrial equipment in industrial control, wireless terminals in self-driving vehicles, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals or smart home devices in smart homes, cellular phones, cordless phones, SIP phones, and wireless local loops. The terminal device can also be a communication device in a future wireless communication system.

[0134] As an example, 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 only hardware devices but can 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 application function and require the use of other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0135] Optionally, the device used to implement the terminal's functions can be the terminal itself; it can also be a device capable of supporting the terminal in implementing these functions, such as a chip system, a communication module, or a modem, which can be installed in the terminal. In this embodiment, the chip system can be composed of chips or may include chips and other discrete devices. This embodiment does not limit the specific technology or device form used in the terminal device. In one possible implementation, the UE can act as a base station. For example, the UE can act as a scheduling entity, providing sidelink signals between UEs in V2X, D2D, or P2P, without relaying communication signals through a base station. In another possible implementation, the UE can also act as a relay node. For example, the UE can act as a relay device or an integrated access and backhaul (IAB) node to provide wireless backhaul services to the terminal device.

[0136] As an example, a network device is a device used to communicate with a terminal device, an entity on the network side used to transmit or receive signals, such as a base station (BS). A network device can also be a radio access network (RAN) node (or device) through which a terminal device accesses a wireless network. A BS can be a device deployed in a RAN capable of wireless communication with a terminal. Base stations can take many forms, such as macro base stations, micro base stations, relay stations, and access points. Exemplarily, the base station involved in this application embodiment can be a base station in 5G, a base station in a 6th generation (6G) mobile communication system, an access network device or module of an access network device in an open radio access network (O-RAN) system, a base station in a future mobile communication system or an access node in a WiFi system, or an evolved node B (eNB) in LTE, etc. Among these, a base station in 5G can also be called a transmission reception point (TRP) or a 5G base station (next-generation node B, gNB). Base stations can also be replaced by the following names, such as: wireless access point, node B, transmitting point (TP), master MeNB, secondary SeNB, multi-standard radio (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), centralized unit (CU), distributed unit (DU), location node, IAB donor, etc. Base stations can also be mobile switching centers and devices that perform base station functions in D2D, V2X, and M2M communications. Base stations can support networks using the same or different access technologies. Optionally, RAN nodes can also be servers, wearable devices, vehicles, or in-vehicle equipment. For example, in V2X technology, RAN equipment can be a roadside unit (RSU).

[0137] 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.

[0138] In some deployments, the network devices mentioned in the embodiments of this application may be devices including CU, DU, or CU and DU, or devices with control plane CU nodes (central unit-control plane (CU-CP)) and user plane CU nodes (central unit-user plane (CU-UP)) and DU nodes. For example, the network devices may include gNB-CU-CP, gNB-CU-UP, and gNB-DU.

[0139] In some deployments, multiple radio access network (RAN) nodes collaborate to assist terminals in achieving radio access, with different RAN nodes 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 equipment or radio unit units, such as RRUs, AAUs, or RRHs.

[0140] 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, some downlink and / or uplink baseband functions, such as, for downlink, precoding, digital beamforming (BF), or one or more of inverse fast Fourier transform (IFFT) / cyclic prefix addition (CP), are moved from the DU to the RU; and for uplink, digital beamforming (BF), or one or more of fast Fourier transform (FFT) / cyclic prefix removal (CP), are moved from the DU to the RU. In one possible implementation, the interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the segmentation between DU and RU differs, corresponding to different categories (Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, F.

[0141] Taking eCPRI Cat A as an example, for downlink transmission, the DU is configured to implement one or more functions before and after layer mapping (i.e., coding, rate matching, scrambling, modulation, and layer mapping), while other functions after layer mapping (e.g., resource element (RE) mapping, digital beamforming (BF), or one or more functions of inverse fast Fourier transform (IFFT) / adding cyclic prefix (CP)) are moved to the RU. For uplink transmission, the DU is configured to implement one or more functions before and after demapping (i.e., decoding, rate matching de-matching, descrambling, demodulation, inverse discrete Fourier transform (IDFT), channel equalization, and demapping), while other functions after demapping (e.g., digital BF or one or more functions of fast Fourier transform (FFT) / removing CP) are moved to the RU. It is understandable that the functional descriptions of the DU and RU corresponding to various types of eCPRI can be found in the eCPRI protocol, and will not be elaborated here.

[0142] 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.

[0143] 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 ORAN system, CU can also be called O-CU (open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called 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.

[0144] In this embodiment, the device for implementing the functions of the network device can be the network device itself; or it can be a device capable of supporting the network device in implementing the functions, such as a chip system, hardware circuit, software module, or hardware circuit plus software module. This device can be installed in the network device or used in conjunction with the network device.

[0145] For example, Figure 2 is a schematic diagram of another communication system. Compared to the communication system shown in Figure 1, the communication system shown in Figure 2 also includes an AI network element 140. The AI ​​network element 140 is used to perform AI-related operations, such as building training datasets or training AI models.

[0146] In one possible implementation, network device 110 can send data related to the training of the AI ​​model to AI network element 140, which then constructs a training dataset and trains the AI ​​model. For example, the data related to the training of the AI ​​model may include data reported by the terminal device. AI network element 140 can send the results of operations related to the AI ​​model to network device 110, which then forwards them to the terminal device. 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. Exemplarily, a portion of the trained AI model may be deployed on network device 110, and another portion on the terminal device. Alternatively, the trained AI model may be deployed on network device 110. Or, the trained AI model may be deployed on the terminal device.

[0147] Figure 2 illustrates the example of AI network element 140 being directly connected to network device 110. In other scenarios, AI network element 140 can also be connected to a terminal device. Alternatively, AI network element 140 can be connected to both network device 110 and a terminal device simultaneously. Alternatively, AI network element 140 can also be connected to network device 110 through a third-party network element. This application embodiment does not limit the connection relationship between AI network elements and other network elements. It is understood that this application embodiment does not limit the number of AI network elements. For example, when there are multiple AI network elements, these multiple AI network elements can be divided based on function, such as different AI network elements being responsible for different functions.

[0148] Optionally, the AI ​​network element 140 can be an AI node or an AI module.

[0149] AI network element 140 can be an independent device, or it can be integrated into the same device to implement different functions. Alternatively, it can be a network element in a hardware device, a software function running on dedicated hardware, or a virtualization function instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the AI ​​network element 140.

[0150] For example, AI network element 140 can also be configured as a module in network devices and / or terminal devices, such as in network device 110 or terminal device shown in Figure 1. Alternatively, AI network element 140 can also be deployed as a module in the core network device of the communication system, or in a location other than terminal devices, network devices, and core network devices, such as in the host or cloud server of an over-the-top (OTT) system. AI network element 140 can communicate with other devices in the communication system, which can be one or more of the following: network devices, terminal devices, or core network elements, etc.

[0151] 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 network devices or multiple terminal devices. The embodiments of this application do not limit the number of network devices and terminal devices included in the communication system.

[0152] 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., NG, Xn) or air interfaces. These network element nodes, such as core network equipment, access network nodes (RAN nodes), terminals, or one or more devices in the operation, administration, and maintenance (OAM) systems, 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 and DU. The CU and / 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 models are configured in the CU-CP and / or CU-UP.

[0153] The AI ​​module is 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., type and / or dimension of input parameters), or output parameters (e.g., type and / or dimension of output parameters). The bias in the activation function can also be referred to as the neural network bias.

[0154] In one example, the neural network mentioned above can be a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), or a generative adversarial network (GAN).

[0155] Deep Neural Networks (DNNs) are artificial neural network architectures with multiple layers of nonlinear transformation units stacked in a hierarchical structure to form deep computational models. Compared to shallow neural networks, deep neural networks have more hidden layers, allowing the network model to capture more complex data structures and higher-level abstract features.

[0156] A CNN is a deep neural network with a convolutional structure. A CNN contains a feature extractor consisting of convolutional layers and subsampling layers. This feature extractor can be viewed as a filter, and the convolution process can be seen as performing convolution between a trainable filter and an input image or a convolutional feature map.

[0157] RNN is a type of recursive neural network that takes sequence data as input, recursively moves along the direction of sequence evolution, and connects all nodes (recurrent units) in a chain-like manner.

[0158] GAN is a deep learning model. It consists of a generator and a discriminator, and is trained through adversarial learning. Its purpose is to estimate the potential distribution of data samples and generate new data samples.

[0159] 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.

[0160] Figure 4 illustrates another possible application framework in a communication system. As shown in Figure 4, the communication system includes a RAN intelligent controller (RIC). For example, the RIC can be the AI ​​modules 117 and 118 shown in Figure 3, used to implement AI-related functions. The RIC includes near-real-time RICs (near-RT RICs) and non-real-time RICs (non-RT RICs). Non-real-time RICs primarily process non-real-time information, such as data that is not sensitive to latency, with latency in the order of seconds. Real-time RICs primarily process near-real-time information, such as data that is relatively sensitive to latency, with latency in the order of tens of milliseconds.

[0161] 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.

[0162] 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., CUs, CU-CPs, CU-UPs, DUs, and / or RUs) and / or terminals. This information can be used as training data or inference data, and the inference results can be delivered to 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.

[0163] Near real-time RICs and non-real-time RICs can also be configured as separate network elements. Optionally, the near real-time RICs and non-real-time RICs can also be part of other devices. For example, the near real-time RIC can be set in a RAN node (e.g., in a CU or DU), while the non-real-time RIC can be set in an OAM, a cloud server, a core network device, or other network devices.

[0164] For example, the network device can be one or more devices in the core network device, access network node (RAN node), or OAM as shown in Figure 3. For instance, the AI ​​module can be the RIC shown in Figure 4, such as a near real-time RIC or a non-real-time RIC. For example, the near real-time RIC is set in the RAN node (e.g., in the CU, DU), while the non-real-time RIC is set in the OAM, cloud server, core network device, or other network device. For example, the near real-time RIC and the non-real-time RIC can also be set up separately as a network element; the network device can be a near real-time RIC or a non-real-time RIC.

[0165] It should be noted that Figures 1 to 4 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 to 4. In practical applications, the communication system may include multiple network devices or multiple terminal devices. The embodiments of this application do not limit the number of network devices and terminal devices included in the communication system.

[0166] II. Examples of Concepts That May Be Involved in the Embodiments of This Application

[0167] 1. Reference signal

[0168] Reference signals, as a crucial component of communication system design, are primarily used for channel state measurement, data demodulation, and beam training. The design of reference signals mainly includes the design of random sequence generation, time-frequency resource mapping, and the power of the corresponding transmission sequence, forming a complete reference signal pattern design. In the embodiments of this application, the design or configuration of the reference signal pattern may at least include the configuration of the reference signal's position, power, and sequence.

[0169] Reference signals refer to different specific signals in uplink and downlink transmissions, as shown in Table 1. The channel state information-reference signal (CSI-RS) is mainly used for downlink channel measurement, obtaining downlink channel state information, beam management, mobility management, rate matching, etc. The demodulation reference signal (DMRS) is used for uplink or downlink channel estimation to demodulate the corresponding physical channel. The sounding reference signal (SRS) is mainly used for uplink channel measurement, time and frequency synchronization, beam management, etc.

[0170] Table 1 shows the reference signals that refer to different specific signals in uplink and downlink transmission.

[0171] One function of a reference signal is to aid in channel estimation. During communication, the receiver needs prior knowledge of the wireless channel information in the antennas at both ends to coherently monitor and decode the data transmitted by the transmitter. Typically, the reference signal is sparse in the time, frequency, and spatial domains. Therefore, after estimating the channel at the time-frequency resource unit containing the reference signal using a channel estimation algorithm, it is also necessary to estimate the wireless channel on the time-frequency spatial resources where the reference signal has not been transmitted. Common channel estimation algorithms include Wiener filtering, least squares (LS), linear minimum mean square error (LMMESE), and compressed sensing (CS) algorithms.

[0172] Since reference signals carry almost no useful information, their overhead is often a consideration to balance channel estimation performance with the available time-frequency resources for data transmission. Many reference signals serve different purposes in 5G systems; for example, DMRS is used for data demodulation, estimating the channel response and the time-frequency resources occupied by the data. There is a trade-off between the accuracy of channel estimation and the density / overhead of DMRS. If the channel exhibits significant frequency selectivity (i.e., large frequency domain variations), the density of DMRS in the frequency domain should be increased. Similarly, if the channel varies rapidly in the time domain, more resources need to be allocated in the time domain to deploy reference signals.

[0173] After determining the DMRS density in the time-frequency domain, it is necessary to further consider the position of the DMRS in the time-frequency resource block. For example, under the condition of channel stability, in order to reduce interpolation error and reduce implementation complexity, the DMRS signal can be evenly distributed in the frequency and time domains. Since the DMRS itself does not transmit any data signals useful to the user, it is necessary to allocate the DMRS with an appropriate density to maximize throughput. The position of the DMRS in a single time-frequency resource block (RB) is mainly determined by the following parameters: (1) Mapping type: There are type A and type B. The two mapping types have different restrictions on the starting symbol position and the number of PDSCH symbols in the physical downlink shared channel (PDSCH); (2) DMRS configuration type: determines the frequency domain resource position of the DMRS; (3) DMRS-additional position: determines whether there is an additional DMRS in the time domain; (4) maximum length: determines whether it is a single-symbol DMRS or a double-symbol DMRS.

[0174] For example, on a single RB, one possible DMRS pattern is shown in Figure 5. The horizontal axis represents 14 orthogonal frequency division multiplexing (OFDM) symbols, and the vertical axis represents 12 subcarriers. The black time-frequency resource blocks represent the locations carrying DMRS signals. The parameters corresponding to this pattern are: mapping type: DMRS mapping type A, DMRS configuration type: DMRS configuration type 1, maxLength = 1, DMRS-Additional Position = 0, DMRS Type APosition = 2. Here, DMRS Type A Position indicates the starting OFDM symbol position of the DMRS. It can be seen that the protocol generates a finite number of DMRS patterns by providing the possible values ​​(or ranges) of each parameter.

[0175] The method for generating DMRS random sequences depends on the specific waveform used. For example, 5G supports both CP-OFDM and DFT-s-OFDM waveforms. The method for mapping DMRS sequences to physical time-frequency resource units is also clearly defined in standard protocols, and will not be detailed here.

[0176] 2. Artificial intelligence, machine learning, AI models, and neural networks

[0177] Artificial intelligence (AI) enables machines to possess human-like intelligence; for example, it allows machines to use computer hardware and software to simulate certain intelligent human behaviors. Generally, AI refers to the technology of exhibiting human intelligence through ordinary computer programs. AI can be defined as a machine or computer that imitates humans and possesses cognitive functions related to human thinking, such as learning and problem-solving. AI can learn from past experiences, make rational decisions, and respond quickly. The goal of AI is to understand intelligence by constructing computer programs that demonstrate symbolic reasoning or logical reasoning.

[0178] Machine learning (ML) is an important technological approach to achieving artificial intelligence, using machine learning to solve problems within artificial intelligence. Machine learning theory primarily involves designing and analyzing algorithms that allow computers to automatically "learn." Machine learning algorithms are a class of algorithms that automatically analyze data to obtain patterns and use these patterns to predict unknown data. Because machine learning algorithms involve a significant amount of statistical theory, they are closely related to inferential statistics and are also known as statistical learning theory. Machine learning can be categorized into supervised learning, unsupervised learning, and reinforcement learning.

[0179] Supervised learning, on the other hand, uses machine learning algorithms to learn the mapping relationship between sample values ​​and labels based on collected sample values ​​and labels, and then expresses this learned mapping relationship using a machine learning model. During training, the model parameters are optimized by calculating the error between the model's predicted values ​​and the true labels. Unsupervised learning (self-supervised learning) relies solely on collected sample values, using algorithms to discover the inherent patterns in the samples themselves. During training, the model parameters are optimized by calculating the error between the model's predicted values ​​and the samples themselves. Reinforcement learning, unlike supervised learning, is a type of algorithm that learns problem-solving strategies through interaction with the environment. For example, in downlink power control, a reinforcement learning model adjusts the downlink transmission power of each user based on the total system throughput feedback from the wireless network, aiming to achieve a higher system throughput. The goal of reinforcement learning is also to learn the mapping relationship between the environment state and the optimal decision action. However, because the label of the "correct action" cannot be obtained in advance, the network cannot be optimized by calculating the error between the action and the "correct action." Reinforcement learning training is achieved through iterative interaction with the environment.

[0180] An AI model is an algorithm or computer program that enables AI functionality. It represents the mapping relationship or function between the model's input and output. AI models can be used for reasoning (or prediction), meaning they can be used to predict the output corresponding to a given input. This output can also be called the reasoning result. AI models can be neural networks or other machine learning models.

[0181] Neural networks (NNs) are a specific implementation of machine learning, a mathematical model that mimics the behavioral characteristics of animal neural networks for information processing. According to the general approximation theorem, neural networks can theoretically approximate any continuous function, thus enabling them to learn arbitrary mappings. Therefore, neural networks can accurately abstract and model complex, high-dimensional problems. Traditional communication systems rely on extensive expert knowledge to design communication modules, while communication systems based on deep neural networks (DNNs) can discover hidden pattern structures in large datasets, establish mapping relationships between data, and achieve performance superior to traditional modeling methods.

[0182] Neural networks typically consist of multiple layers, each layer containing one or more logical decision units called neurons. Increasing the depth and / or width of a neural network enhances its expressive power, providing more robust information extraction and abstract modeling capabilities for complex systems. The depth of a neural network can be understood as the number of layers it comprises, while the number of neurons in each layer can be referred to as the width of that layer.

[0183] Figure 6 is a schematic diagram of a neural network provided in an embodiment of this application. In one implementation, the neural network includes an input layer and an output layer. The input layer of the neural network processes the received input through neurons and then passes the result to the output layer, which obtains the output result of the neural network. In another implementation, the neural network includes an input layer, a hidden layer, and an output layer. The input layer of the neural network processes the received input through neurons and then passes the result to the intermediate hidden layer. The hidden layer then passes the calculation result to the output layer or an adjacent hidden layer, and finally, the output layer obtains the output result of the neural network. A neural network may include one or more hidden layers connected in sequence. This application does not limit the number of hidden layers in a neural network. DNNs typically include multiple hidden layers. Hidden layers usually affect the ability to extract information and fit functions. Increasing the number of hidden layers or expanding the width of each layer can improve the function fitting ability of the DNN. The neural network shown in Figure 6 includes one input layer, one hidden layer, and one output layer, where the input layer has 3 neurons, the hidden layer has 4 neurons, and the output layer has 2 neurons. It is understood that the number of layers and the number of neural network elements in each layer of the neural network shown in Figure 6 are only examples.

[0184] Each connection between neurons corresponds to a weight (its value is called a weight), and these weights can be updated through training. Each neuron can also correspond to a bias value, which can be updated through training. Updating a neural network means updating these weights and bias values. Knowing the structure of a neural network—that is, how the output of one neuron is input into another—as well as the weights and biases, we know all the information about the neural network.

[0185] As shown in Figure 6, a neuron may have multiple input connections, and each neuron calculates its output based on its input. For example, each neuron performs a weighted summation of its input values, and the result is passed through an activation function to produce the output. A neuron may have multiple output connections, and the output of one neuron serves as the input to the next neuron. It should be understood that the input layer only has output connections; each neuron in the input layer receives the value input to the neural network, and the value of each neuron serves as the input to all output connections. The output layer only has input connections.

[0186] Figure 7 is a schematic diagram of a neuron. As shown in Figure 7, assume the neuron's input is x = [x0, x1, ..., x...]. n The weights corresponding to each input are d = [d0, d1, ..., d2]. n Where n is a positive integer, d i and x i It can be any possible type, such as a decimal, an integer (e.g., 0, a positive integer, or a negative integer), or a complex number. i As xi The weights are used to assign weights to x. i Weighting is applied. The bias value is b, which is used to sum the input values ​​according to the weights. Assuming the activation function is represented by f(z), the output y of the neuron shown in Figure 7 is:

[0187] The bias value b can be a decimal, an integer (0, a positive integer, or a negative integer), or a complex number, among other values. The activation functions of different neurons in a neural network can be the same or different. The form of the activation function f(z) can vary, and these are not all listed here. This application does not restrict the form of the activation function.

[0188] Based on the way the network is constructed, DNNs can be divided into feedforward neural networks (FNN), convolutional neural networks (CNN), and recurrent neural networks (RNN).

[0189] The defining characteristic of FNN networks is that neurons in adjacent layers are completely connected pairwise, which typically requires a large amount of storage space and results in high computational complexity. CNN is a type of neural network specifically designed to process data with a grid-like structure. For example, time-series data (discrete sampling along the time axis) and image data (two-dimensional discrete sampling) can both be considered grid-like data. CNNs do not use all the input information at once; instead, they use a fixed-size window to extract a portion of the information for convolution operations, significantly reducing the computational cost of model parameters. Furthermore, depending on the type of information extracted by the window (e.g., people and objects in an image represent different types of information), each window can use different convolution kernels, allowing CNNs to better extract features from the input data. RNN is a type of DNN network that utilizes feedback time-series information. Its input includes the new input value at the current time step and its own output value at the previous time step. RNNs are suitable for acquiring temporally correlated sequence features, and are particularly suitable for applications such as speech recognition and channel coding / decoding.

[0190] The aforementioned FNN, CNN, and RNN are common neural network structures, all built upon neurons. Each neuron performs a weighted summation of its input values, and the result is passed through a nonlinear function to produce the output. The weights and nonlinear function used in the weighted summation operation of neurons in a neural network are called the network's parameters. Taking a neuron with max{0,x} as an example of a nonlinear function, we can... The parameters of the operated neuron are: weights d = [d0, ..., d...]. nThe weighted summation bias is b, and the nonlinear function is max{0,x}. The parameters of all neurons in a neural network constitute the parameters of that neural network.

[0191] 3. Transmitted symbols, received signals, and equalized symbols

[0192] The transmitted symbols, received signals, and equalized symbols are obtained from different processing flows in the communication system. To facilitate understanding, the main processes of the communication system will be briefly described first.

[0193] Figure 8 is a schematic diagram of the main process of a communication system. Taking point-to-point communication as an example, the transmitting end needs to encode and modulate the signal to form a signal suitable for transmission in the channel. The channel is taken as a wireless channel. The receiving end needs to perform equalization, demodulation, and decoding to obtain the original signal. It can be seen that the processing at the receiving end is the reverse process of the transmitting end. The positions of modulation and demodulation in the main process of the communication system are shown in Figure 8. Modulation maps a discrete stream of bits consisting of 0s and 1s into modulation symbols in a specific way for signal transmission. Demodulation demodulates the equalized signal or the equalized symbols and decodes them to obtain the corresponding signal, such as the bit stream consisting of 0s and 1s.

[0194] Commonly used modulation methods include Quadrature Phase Shift Keying (QPSK), Amplitude Shift Keying Modulation (ASK), Frequency Shift Keying Modulation (FSK), Phase Shift Keying Modulation (PSK), Quadrature Amplitude Modulation (QAM), and Quadrature Amplitude Modulation (QAM).

[0195] The set of all possible modulation symbols corresponding to a modulation scheme, and the mapping relationship between modulation symbols and bits under that modulation scheme, is called a constellation set, constellation, or constellation diagram. The number of all possible modulation symbols corresponding to each modulation scheme is usually a power of 2, i.e., 2^3 ... mEach modulation symbol represents m bits of information, where m can also be called the order of the modulation scheme. Taking 16QAM and QPSK as examples, refer to the constellation diagram of 16QAM shown on the left in Figure 9. The number of possible modulation symbols for this scheme is 16, and each symbol represents 4 bits of information; the order of this modulation scheme is 4. Refer to the constellation diagram of QPSK shown on the right in Figure 9. The number of possible modulation symbols for this scheme is 4, and each symbol represents 2 bits of information; the order of this modulation scheme is 2.

[0196] As shown in Figure 8, the transmitted symbol is a symbol obtained by modulating the encoded bit stream; it is the modulation symbol in the modulation method. The transmitted symbol can also be called the original symbol, the original constellation point, or the original signal.

[0197] Received signals are the observations of the receiver on the corresponding resources, such as observations on DMRS symbols or PDSCH resources.

[0198] The equalized symbol is obtained by equalizing the received signal based on the channel estimation results using an equalization algorithm. Optionally, the equalized symbol is also called an equalized symbol, an equalized signal, or an equalized constellation point.

[0199] 4. Model Monitoring

[0200] In this embodiment of the application, model monitoring is also referred to as model evaluation, performance evaluation, or model performance assessment.

[0201] Model monitoring is a technique used to observe and ensure the performance and reliability of running AI / ML models. Before operation, the performance of an AI / ML model is judged solely by observing its performance on pre-provided datasets (historical / non-real-world data). After training on a static dataset (i.e., training data), the model is then put into dynamic / real-world, constantly changing scenarios / data for inference tasks. This difference between the static data during training and the dynamically changing data in real-world use can cause the performance of AI / ML models to degrade over time. Therefore, it is necessary to perform model monitoring concurrently with the running model.

[0202] This application uses an AI model for DMRS channel estimation as an example. This AI model can be called an AI channel estimation model or an AIDMRS estimation model. Correspondingly, traditional methods that use channel estimation algorithms for channel estimation can be called non-AI channel estimation algorithms or non-AIDMRS estimation.

[0203] Third, how to simultaneously conduct model performance evaluation on a running model is an urgent problem to be solved.

[0204] One solution is to evaluate model performance based on the output metrics of the AI ​​model. For example, Figure 10 is a schematic diagram of DMRS-based model performance evaluation. In the upper branch, the received signal or observation Y on the reference signal resource and DMRS are used as inputs to the AI ​​channel estimation model to obtain the channel estimation result H1. In the lower branch, a non-AI channel estimation algorithm with better channel estimation performance, such as Wiener filtering, is used for DMRS estimation to obtain the channel estimation result H2. In this model performance evaluation, the channel estimation result H2 of the non-AI channel estimation algorithm is taken as the ideal true channel value. If the normalized mean square error (NMSE) values ​​of H1 and H2 are large, such as exceeding a certain threshold, it indicates that the performance of the AI ​​channel estimation model has degraded. Here, NMSE is a metric used to evaluate the accuracy of the prediction model, providing the difference between the predicted value and the actual value. The model user can feed back the performance evaluation results of the performance degradation to the other side to complete the entire evaluation process. However, the scheme of evaluating model performance based on the output index of AI model has the following problems: (1) The scheme takes the result of non-AI channel estimation algorithm as the comparison object, that is, the true value or the actual value. If the performance of non-AI channel estimation algorithm is not as good as AI model, the scheme will lead to incorrect performance evaluation results. (2) In some scenarios where non-AI algorithm cannot effectively estimate the equivalent channel (such as superimposed reference signal, data and reference signal are superimposed on RE to improve the resource utilization of physical channel), additional reference signal resources need to be configured (such as DMRS configured according to a specific resource configuration method), or fall back to non-superimposed state for non-AI channel estimation. This may lose physical channel (such as PDSCH or PUSCH) resources and lose communication performance.

[0205] IV. The communication method 100 provided in this application

[0206] This application provides a communication method 100 that calculates an index of the difference between the equalized symbol and the corresponding transmitted symbol to evaluate performance. This communication method 100 does not rely on the equivalent channel H estimated by DMRS, but instead uses the dispersion of the equalized symbol and the transmitted symbol calculated from H to evaluate model performance. Since the transmitted symbol corresponding to the modulation scheme is the true value, the closer the estimated channel is to the actual channel, the closer the equalized symbol should be to the original DMRS or the original data constellation under the same equalization algorithm. Therefore, compared to schemes that evaluate model performance based on the output index of AI models, the communication method 100 provided in this application can improve the accuracy of performance evaluation.

[0207] For example, taking the reference signal resource shown in Figure 11 as an example, on the resource blocks filled in black, the terminal performs channel estimation based on a non-AI channel estimation algorithm; on the resource blocks filled in gray, the terminal performs channel estimation based on an AI channel estimation model. The transmitted symbols on the reference signal resource shown in Figure 11 are DMRS modulated into QPSK. As shown in Figure 11, the equivalent channel estimated by the terminal based on the AI ​​channel estimation model is H1. Based on H1, the received signal on the corresponding resource block is equalized to obtain the equalized symbol S1; the equivalent channel estimated by the terminal based on the non-AI channel estimation algorithm is H2. Based on H2, the received signal on the corresponding resource block is equalized to obtain the equalized symbol S2.

[0208] Taking the difference between the equalized symbol and the corresponding transmitted symbol represented using Euclidean distance, and the transmitted symbol of DMRS on a single resource element (RE) as an example, the comparison object for the terminal's performance evaluation is the four modulation symbols corresponding to QPSK shown in Figure 9, with the codebook as follows:

[0209] Assuming that in Figure 11, the symbol S1 corresponding to the AI ​​channel estimation model is... It is evident that S1 belongs to the first quadrant and is the first codeword in the codebook. For the transmitted symbol corresponding to symbol S1, compare the two and calculate the Euclidean distance between S1 and its corresponding transmitted symbol.

[0210] Assuming that in Figure 11, the symbol S2 corresponding to the non-AI channel estimation algorithm is... It is evident that S2 belongs to the first quadrant and is the first codeword in the codebook. For the transmitted symbol corresponding to symbol S2, calculate the Euclidean distance between S2 and its corresponding transmitted symbol.

[0211] The calculations above show that the Euclidean distance of the AI ​​channel estimation model is smaller than that of the non-AI channel estimation algorithm. Furthermore, since the equalization is performed by the same terminal, their equalization algorithms are consistent, meaning the performance evaluation results are independent of the equalization algorithm. Therefore, the channel estimation results of the AI ​​channel estimation model are closer to the actual channel than those of the non-AI channel estimation algorithm, thus eliminating the need to update the AI ​​channel estimation model.

[0212] As can be seen, in the communication method 100 exemplified in Figure 11, the truth value of the transmitted symbol is used as the comparison object to measure the model performance. Compared with using the result of the non-AI channel estimation algorithm as the comparison object, if the performance of the non-AI channel estimation algorithm is not as good as the performance of the AI ​​channel estimation model, it will lead to an incorrect evaluation result of the performance of the AI ​​channel estimation model. Therefore, the communication method 100 can improve the accuracy of performance evaluation.

[0213] In addition, this communication method still needs to utilize non-AI channel estimation because the degree of dispersion may vary under different channels and signal-to-noise ratios. Therefore, while using the true value of the transmitted symbol as a comparison object, it also combines the degree of dispersion corresponding to the non-AI channel estimation for comparison, which can improve the accuracy of performance evaluation.

[0214] In this embodiment, the communication method 100 can be implemented based on the interaction between a transmitting end and a receiving end. The receiving end is a device that deploys an AI model, i.e., the AI ​​model user side; the transmitting end is a device that sends an AI channel estimation model to the receiving end, i.e., the opposite side of the AI ​​model user side. The relevant functions of the transmitting end can also be applied to a chip or chip module in the transmitting end, or to a module or unit that can realize all or part of the functions of the transmitting end; correspondingly, the relevant functions of the receiving end can be applied to a chip or chip module in the receiving end, or to a module or unit that can realize all or part of the functions of the receiving end.

[0215] For example, in downlink transmission, the receiving end is a terminal, or a chip or chip module within the terminal, or a module or unit capable of implementing all or part of the terminal's functions; the transmitting end is a network device, or a chip or chip module within the network device, or a module or unit capable of implementing all or part of the network device's functions. Similarly, in uplink transmission, the receiving end is a network device, or a chip or chip module within the network device, or a module or unit capable of implementing all or part of the network device's functions; the transmitting end is a core network element, or a chip or chip module within the core network element, or a module or unit capable of implementing all or part of the core network element's functions.

[0216] The following description, with reference to Figures 12 and 13, using downstream transmission as an example, illustrates the communication method 100 described in the embodiments of this application.

[0217] Figure 12 is a flowchart illustrating a communication method 100 provided in an embodiment of this application. This communication method not only performs performance evaluation on the AI ​​channel estimation model in the terminal but also improves the accuracy of the performance evaluation. As shown in Figure 12, the communication method includes, but is not limited to, the following steps:

[0218] 101. The terminal performs channel estimation on the first received signal based on the AI ​​channel estimation model to obtain channel estimation result H1; and performs channel estimation on the second received signal based on the non-AI channel estimation algorithm to obtain channel estimation result H2.

[0219] The method further includes: the network device sending a first transmission symbol and a second transmission symbol to the terminal. The first transmission symbol is the transmission symbol corresponding to the first received signal, and the second transmission symbol is the transmission symbol corresponding to the second received signal.

[0220] The first and second received signals can be observations on reference signal resources, i.e., reference signals, or observations on PDSCH resources, i.e., data signals. Optionally, if the first and second received signals are reference signals, the network device can additionally configure a reference signal for performance evaluation of the AI ​​channel estimation model. The modulation scheme of the reference signals corresponding to the AI ​​channel estimation model and the non-AI channel estimation model is the same. One or more of the configurations of the location, power, or sequence of the reference signal may be different or the same, and this application does not impose any limitations.

[0221] In one optional embodiment, the first received signal and the second received signal are the same received signal, and correspondingly, the first transmitted symbol corresponding to the first received signal and the second transmitted symbol corresponding to the second received signal are the same transmitted symbol. In another optional embodiment, the first received signal and the second received signal are different received signals, and correspondingly, the first transmitted symbol and the second transmitted symbol are also different transmitted symbols, but the modulation scheme is the same.

[0222] In this context, the transmitted symbols corresponding to DMRS are QPSK constellation points, meaning DMRS is primarily modulated into symbols within the QPSK constellation diagram. Data signals, such as the modulation and coding scheme (MCS) on PDSCH resources, are configured with various modulation schemes, such as QPSK, 16QAM, 64QAM, 256QAM, or 1024QAM. Since higher-order modulation schemes correspond to more constellation points, the equalized symbols are more difficult to compare with their corresponding transmitted symbols due to noise. Optionally, if the observed received signal is a data signal, a lower-order modulation scheme can be configured for performance evaluation.

[0223] 102. The terminal performs equalization processing on the first received signal based on the channel estimation result H1 to obtain the equalized symbol S1; and performs equalization processing on the second received signal based on the channel estimation result H2 to obtain the equalized symbol S2.

[0224] Since it is the same terminal, the equalization algorithm used by the terminal for the first received signal and the second received signal is the same.

[0225] 103. The terminal determines the performance index of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol.

[0226] The first and second transmitted symbols are the same transmitted symbol. For example, if an AI channel estimation model is running in a terminal, and after a period of time or an event triggers the terminal to perform a performance evaluation on the AI ​​channel estimation model, then in addition to obtaining H1 by performing channel estimation on the reference signal resource based on the AI ​​channel estimation model, the terminal will also perform channel estimation on the same reference signal resource based on a non-AI channel estimation algorithm to obtain H2. Accordingly, based on the different channel estimation results, such as H1 and H2, the terminal performs equalization processing on the received signal on the reference signal resource to obtain symbol S1 and symbol S2, respectively.

[0227] The first and second transmitted symbols are different transmitted symbols. The network device can allocate reference signal resources for the terminal for non-AI channel estimation. For example, if the terminal triggers a performance evaluation of the AI ​​channel estimation model, but the non-AI channel estimation algorithm cannot accurately estimate the channel using the reference signal resources corresponding to the AI ​​channel estimation model, then the network device will configure another reference signal resource for the terminal to perform channel estimation using the non-AI channel estimation algorithm.

[0228] In one optional implementation, step 103 may include: the terminal calculating a first parameter value based on symbol S1 and the first transmitted symbol; and calculating a second parameter value based on symbol S2 and the second transmitted symbol; the first parameter value and the second parameter value are used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol; and determining the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value. Optionally, in this implementation, the first parameter value is the Euclidean distance between symbol S1 and the first transmitted symbol; and the second parameter value is the Euclidean distance between symbol S2 and the second transmitted symbol. Based on this method, the terminal uses parameter values ​​to indicate the degree of dispersion between the equalized symbol and the transmitted symbol, thereby facilitating the determination of the performance index of the AI ​​channel estimation model.

[0229] The calculation method between the first parameter value and the second parameter value may include, but is not limited to, the following optional implementation methods:

[0230] In one optional implementation, the terminal determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the ratio between the first parameter value and the second parameter value is less than or equal to a first threshold, the terminal determines that the AI ​​channel estimation model does not need to be updated; if the ratio between the first parameter value and the second parameter value is greater than the first threshold, the terminal determines that the AI ​​channel estimation model needs to be updated.

[0231] Based on this method, if the ratio between the first parameter value and the second parameter value is less than or equal to the first threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is lower than the dispersion between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running normally, its performance has not degraded, and no model update is needed. If the ratio between the first parameter value and the second parameter value is greater than the first threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is higher than the dispersion between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is worse than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running abnormally, its performance has degraded, and a model update is needed.

[0232] In another optional implementation, the terminal determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the difference between the first parameter value and the second parameter value is less than or equal to a second threshold, the terminal determines that the AI ​​channel estimation model does not need to be updated; if the difference between the first parameter value and the second parameter value is greater than the second threshold, the terminal determines that the AI ​​channel estimation model needs to be updated.

[0233] Based on this method, if the difference between the first parameter value and the second parameter value is less than or equal to the second threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is lower than the dispersion between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running normally, its performance has not degraded, and no model update is needed. If the difference between the first parameter value and the second parameter value is greater than the second threshold, it indicates that the dispersion between symbol S1 and the first transmitted symbol is higher than that between symbol S2 and the second transmitted symbol. This means that the performance of the AI ​​channel estimation model is worse than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model is running abnormally, its performance has degraded, and a model update is needed.

[0234] In another optional implementation, the terminal determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the first parameter value is less than or equal to a third threshold and the ratio between the first parameter value and the second parameter value is less than or equal to the first threshold, the terminal determines that the AI ​​channel estimation model does not need to be updated; if the first parameter value is greater than the third threshold, or the ratio between the first parameter value and the second parameter value is greater than the first threshold, the terminal determines that the AI ​​channel estimation model needs to be updated.

[0235] Based on this method, the terminal also needs to consider the relationship between the first parameter value and the third threshold to avoid the situation where the dispersion indicated by the first parameter value is too large, even if the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model may run abnormally.

[0236] In another optional implementation, the terminal determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the first parameter value is less than or equal to a third threshold and the difference between the first parameter value and the second parameter value is less than or equal to the second threshold, the terminal determines that the AI ​​channel estimation model does not need to be updated; if the first parameter value is greater than the third threshold, or the difference between the first parameter value and the second parameter value is greater than the second threshold, the terminal determines that the AI ​​channel estimation model needs to be updated.

[0237] Based on this method, the terminal also needs to consider the relationship between the first parameter value and the third threshold to avoid the situation where the dispersion indicated by the first parameter value is too large, even if the performance of the AI ​​channel estimation model is better than that of the non-AI channel estimation algorithm, the AI ​​channel estimation model may run abnormally.

[0238] In addition, in this communication method, the network device also sends first configuration information to the terminal, and the terminal also receives the first configuration information. The first configuration information indicates the calculation method and corresponding threshold between a first parameter value and a second parameter value. For example, the first configuration information indicates that the calculation method is the ratio between the first parameter value and the second parameter value, and the corresponding threshold is a first threshold. As another example, the first configuration information indicates that the calculation method is the difference between the first parameter value and the second parameter value, and the corresponding threshold is a second threshold. Based on this method, the terminal can determine the performance of the AI ​​channel estimation model according to the calculation method indicated by the first configuration information and the corresponding threshold.

[0239] Optionally, the first configuration information not only indicates the first threshold or the second threshold corresponding to the calculation method, but also indicates the third threshold, which is used to evaluate the performance of the AI ​​channel estimation model in combination with the first parameter value.

[0240] Optionally, the threshold corresponding to the calculation method between the first and second parameter values ​​is also related to the modulation scheme of the transmitted symbol. For example, different modulation schemes may correspond to different thresholds. Since different modulation schemes affect the degree of dispersion between the equalized symbol and the transmitted symbol, this method can improve the accuracy of performance evaluation. Similarly, the third threshold can also be related to the modulation scheme of the transmitted symbol.

[0241] Optionally, the threshold corresponding to the calculation method between the first parameter value and the second parameter value can also be called the monitoring threshold or performance evaluation threshold.

[0242] 104. The terminal sends the performance metrics of the AI ​​channel estimation model to the network device.

[0243] In one optional implementation, the performance metrics include whether the AI ​​channel estimation model needs updating or not. Optionally, if the AI ​​channel estimation model does not need updating, the terminal may not report the performance metrics; the performance metrics will only be reported when an update is required.

[0244] Based on this method, the terminal can report the performance indicators of the AI ​​channel estimation model to the network side or the core network side, which is conducive to timely updating of the AI ​​channel estimation model and improving the accuracy of channel estimation.

[0245] Figure 13 is a schematic flowchart of another communication method 100 provided in an embodiment of this application. In this communication method, the network device may also send second configuration information to the terminal. The second configuration information is used to indicate a first resource for channel estimation based on an AI channel estimation model and a second resource for channel estimation based on a non-AI channel estimation algorithm. As shown in Figure 13, this communication method may include, but is not limited to, the following steps:

[0246] 201. The network device sends the AI ​​channel estimation model to the terminal, and the terminal receives and deploys the AI ​​channel estimation model accordingly.

[0247] 202. The terminal runs the AI ​​channel estimation model, performs channel estimation based on DMRS, and receives PDSCH.

[0248] For example, if the AI ​​channel estimation model has no performance issues in the early stages of deployment, the likelihood of model updates is low, so channel estimation can be performed based on the deployed AI channel estimation model.

[0249] 203. The network device sends first configuration information and second configuration information to the terminal, and the terminal receives the first configuration information and second configuration information accordingly.

[0250] The first configuration information and the second configuration information may be sent together or separately, and this application does not impose any restrictions.

[0251] Optionally, the network device may send first and second configuration information after the model has been running for a period of time, such as periodically or non-periodically, to trigger the terminal to perform a performance evaluation on the AI ​​model.

[0252] The first configuration information is used to indicate the calculation method between the first parameter value and the second parameter value, as well as the corresponding threshold. The first parameter value and the second parameter value are used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. For a description of the first configuration information, the first parameter value, and the second parameter value, please refer to the relevant content in the embodiment shown in Figure 12, which will not be elaborated here.

[0253] The second configuration information is used to indicate a first resource for channel estimation based on an AI channel estimation model and a second resource for channel estimation based on a non-AI channel estimation algorithm. Based on this method, the terminal can perform channel estimation on the first resource based on the AI ​​channel estimation model and equalize the signal received on the first resource based on the channel estimation result to obtain symbol S1; the terminal can perform channel estimation on the second resource based on the non-AI channel estimation algorithm and equalize the signal received on the second resource based on the channel estimation result to obtain symbol S2.

[0254] 204. The terminal performs channel estimation on the first received signal on the first resource based on the AI ​​channel estimation model to obtain channel estimation result H1; and performs channel estimation on the second received signal on the second resource based on the non-AI channel estimation algorithm to obtain channel estimation result H2.

[0255] Optionally, the first resource and the second resource are the same. In the embodiment shown in Figure 12, the AI ​​channel estimation model and the non-AI channel estimation algorithm can be evaluated for performance based on the same received signal and transmitted symbols.

[0256] Optionally, the first resource and the second resource are different. Correspondingly, the first transmission symbol and the second transmission symbol are signals mapped to the corresponding resources, which may be the same or different. Optionally, the second configuration information can configure a resource pool for performance evaluation. This resource pool is used for channel estimation during performance evaluation, and the first resource and the second resource can be determined from this resource pool. Optionally, the first resource and the second resource can also be referred to as performance evaluation resources.

[0257] Optionally, the first resource can be the reference signal resource used in the operation of the AI ​​channel estimation model; the second resource can be a specially configured reference signal resource for non-AI channel estimation.

[0258] Based on this method, when non-AI channel estimation fails to estimate an effective channel on the reference signal resources used in AI channel estimation, additional reference signal resources can be configured for non-AI channel estimation. For example, in cases where reference signal resources are superimposed on resource units to improve resource utilization, additional reference signal resources need to be configured or the system needs to revert to a non-superimposed state for non-AI channel estimation.

[0259] 205. The terminal performs equalization processing on the first received signal based on the channel estimation result H1 to obtain the equalized symbol S1; and performs equalization processing on the second received signal based on the channel estimation result H2 to obtain the equalized symbol S2.

[0260] 206. The terminal determines the performance index of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol.

[0261] The first transmitted symbol is a transmitted symbol on the first resource, and the second transmitted symbol is a transmitted symbol on the second resource. Both have the same modulation scheme.

[0262] 207. When the AI ​​channel estimation model needs to be updated, the terminal sends performance metrics to the network device.

[0263] The optional implementations of steps 205 to 207 can be found in the relevant content of the embodiment shown in Figure 12, and will not be described in detail here.

[0264] In one embodiment of the communication method 100, the Euclidean distances between N equalization symbols and their corresponding transmitted symbols are summed or averaged to obtain the metric results corresponding to the AI ​​channel estimation model and the non-AI channel estimation algorithm. Each of the AI ​​channel estimation model and the non-AI channel estimation algorithm has a metric result calculated after considering multiple REs. Therefore, the performance of the AI ​​channel estimation model can be evaluated based on these metric results.

[0265] For example, taking the reference signal resources shown in Figure 10 as an example, based on the received signal y1 on the two gray-filled reference signal resources, the terminal estimates the equivalent channel H1 using an AI channel estimation model. Based on H1 and the received signal y1, the equalized symbols S1-1 and S1-2 are obtained through an equalizer. Similarly, based on the received signal y2 on the two black-filled reference signal resources, the terminal estimates the equivalent channel H2 using a non-AI channel estimation algorithm. Based on H2 and the received signal y2, the equalized symbols S2-1 and S2-2 are obtained through an equalizer. The comparison object for the terminal's performance evaluation is the four modulation symbols corresponding to the QPSK shown in Figure 9, with the codebook as follows:

[0266] Assume the symbol S1-1 corresponding to the AI ​​channel estimation model is S1-2 is So, the first codeword in the codebook The second codeword is the transmitted symbol corresponding to symbol S1-1. Let S1-1 be the transmitted symbol corresponding to symbol S1-2. Compare the two symbols separately and calculate the Euclidean distance between S1-1 and its corresponding transmitted symbol. The Euclidean distance between S1-2 and its corresponding transmitted symbol is:

[0267] Assume that the symbol S2 corresponding to the non-AI channel estimation algorithm is S2-1. S2-2 is So, the first codeword in the codebook The second codeword is the transmitted symbol corresponding to symbol S2-1. Let S2-1 be the transmitted symbol corresponding to symbol S2-2. Compare the two symbols separately and calculate the Euclidean distance between S2-1 and its corresponding transmitted symbol. The Euclidean distance between S2-2 and its corresponding transmitted symbol is:

[0268] The above calculation results show that the Euclidean distance of the AI ​​channel estimation model is The Euclidean distance of the non-AI channel estimation model is

[0269] Furthermore, by comparing the ratio or difference between the Euclidean distance of the AI ​​channel estimation model and the Euclidean distance of the non-AI channel estimation model with the corresponding threshold, the terminal can determine the performance indicators of the AI ​​channel estimation model, such as whether the model is normal or abnormal.

[0270] As can be seen, this embodiment uses the measurement results of multiple REs for performance evaluation, which can further improve the accuracy of AI channel estimation model performance evaluation.

[0271] V. The communication method provided in this application 200

[0272] This application provides a communication method 200, which uses the measurement range of a first parameter at a first time point, corresponding to the measurement value of a second parameter, as a comparison object to determine the corresponding measurement range of the first parameter, and the measurement value of the second parameter at a second time point. Based on the measurement value of the second parameter at the second time point and the measurement value of the second parameter at the first time point, a performance index of the AI ​​channel estimation model is determined. The first parameter indicates channel quality; the second parameter, determined based on the AI ​​channel estimation model, indicates the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. Therefore, the communication method 200, for the AI ​​channel estimation model, determines the change in the degree of dispersion between the equalized symbol and the corresponding transmitted symbol at different times, and evaluates the performance of the AI ​​channel estimation model.

[0273] The communication method 200 uses the degree of dispersion between the equalized symbols determined by the AI ​​channel estimation model and the transmitted symbols at previous times, such as the initial stage of model deployment, as the comparison object. This avoids using the results of non-AI channel estimation algorithms as the comparison object. If the performance of non-AI channel estimation algorithms is not as good as that of AI channel estimation models, it will lead to incorrect evaluation of the performance of AI channel estimation models. The communication method 200 can improve the accuracy of performance evaluation.

[0274] In addition, the communication method 200 does not require channel estimation based on a non-AI channel estimation algorithm. Therefore, it does not require additional configuration of reference signal resources for non-AI channel estimation, thereby improving resource utilization and enhancing communication performance.

[0275] The communication method 200 does not configure reference signal resources for non-AI channel estimation. When the model performance is normal within the same interval of the first parameter, the measured values ​​or measurement range table of the second parameter of the equalization symbol and the transmitted symbol are used. When the model performance evaluation is triggered, the measured values ​​of the second parameter in the table are compared and the performance evaluation results are fed back.

[0276] In this embodiment, the communication method 200 can be implemented based on the interaction between a transmitting end and a receiving end. The receiving end is a device that deploys an AI model, i.e., the AI ​​model user side; the transmitting end is a device that sends an AI channel estimation model to the receiving end, i.e., the opposite side of the AI ​​model user side. The relevant functions of the transmitting end can also be applied to a chip or chip module in the transmitting end, or to a module or unit that can realize all or part of the functions of the transmitting end; correspondingly, the relevant functions of the receiving end can be applied to a chip or chip module in the receiving end, or to a module or unit that can realize all or part of the functions of the receiving end.

[0277] For example, in downlink transmission, the receiving end is a terminal, or a chip or chip module within the terminal, or a module or unit capable of implementing all or part of the terminal's functions; the transmitting end is a network device, or a chip or chip module within the network device, or a module or unit capable of implementing all or part of the network device's functions. Similarly, in uplink transmission, the receiving end is a network device, or a chip or chip module within the network device, or a module or unit capable of implementing all or part of the network device's functions; the transmitting end is a core network element, or a chip or chip module within the core network element, or a module or unit capable of implementing all or part of the core network element's functions.

[0278] The communication method 200 described in this application embodiment will be described below with reference to Figure 14, taking downstream transmission as an example. As shown in Figure 14, the communication method 200 may include, but is not limited to, the following steps:

[0279] 301. The terminal determines the first configuration information at a first moment. The first configuration information is used to indicate the measurement value of the second parameter corresponding to the measurement value range of the first parameter.

[0280] The first parameter is used to indicate channel quality, such as RSRP, RSRQ, SINR, or RSSI.

[0281] The second parameter, determined based on the AI ​​channel estimation model, indicates the degree of dispersion between the equalized symbol and its corresponding transmitted symbol. For example, the second parameter could be Euclidean distance, indicating the dispersion between the equalized symbol and its corresponding transmitted symbol. Alternatively, the second parameter could be the NMSE or MAE between the equalized symbol and its corresponding transmitted symbol.

[0282] The first moment can be the initial deployment moment of the AI ​​channel estimation model, the moment when the AI ​​channel estimation model is distributed, the moment when the first configuration information is received, or the moment when the AI ​​channel estimation model is triggered to be used, etc.

[0283] The measured value of the second parameter indicated by the first configuration information can be a single value or a range of values.

[0284] For example, taking SINR as the first parameter, NMSE or MAE as the second parameter, and the first time point as the initial stage of model deployment, the measurement value of the second parameter corresponding to the measurement range of the first parameter indicated by the first configuration information can be shown in Table 2 below.

[0285] Table 2 shows the measurement range of SINR and its corresponding NMSE or MAE during the initial stage of model deployment.

[0286] For example, the measurement range of the second parameter corresponding to the measurement range of the first parameter indicated by the first configuration information can be shown in Table 3 below.

[0287] Table 3 shows the measurement range of SINR and the corresponding NMSE or MAE during the initial stage of model deployment.

[0288] Optionally, Table 2 can be configured for a single terminal (per UE), where the terminal can determine a specific NMSE / MAE value based on the SINR value. Table 3 can be configured for a single cell, where multiple terminals exist. Different terminals can determine different NMSE / MAE values ​​based on the SINR value, thus using a range of NMSE / MAE values. As the SINR increases, the NMSE / MAE value or the corresponding value within the range decreases, i.e., A>B>C>…>E>… in Table 2, and |A1-A0|>|B1-B0|>|C1-C0|>…>|E1-E0|>… in Table 3. Here, ||| represents an absolute value.

[0289] In one optional implementation, the terminal determines first configuration information at a first moment, including: the terminal estimates a measurement range of a first parameter based on a reference signal at the first moment; the terminal calculates the NMSE or MAE value between the equalized symbol and the corresponding transmitted symbol according to the measurement range of the first parameter, and uses this value as the measurement value of a second parameter corresponding to the measurement range. Wherein, as the value of the first parameter increases, the measurement value of the second parameter decreases. Correspondingly, the network device can configure the reference signal resources at the first moment and transmit the reference signal. Based on this method, the terminal can interact with the network device to construct a correspondence between the measurement range of the first parameter and the measurement value of the second parameter.

[0290] In another optional implementation, the terminal determines the first configuration information at a first moment, including: the terminal receiving the first configuration information from the other side at the first moment, the first configuration information being used to indicate the measured value of the second parameter corresponding to the measured value range of the first parameter. Based on this method, the terminal can directly obtain the correspondence between the measured value range of the first parameter and the measured value of the second parameter from the network device. Optionally, the network device can receive the measured values ​​of the second parameter corresponding to the first parameter values ​​reported by each terminal, such as each terminal in a cell, and construct the correspondence as shown in Table 3.

[0291] 302. The terminal determines the measurement range of the corresponding first parameter and the measurement value of the second parameter at the second time step based on the AI ​​channel estimation model and the first configuration information.

[0292] In one optional implementation, the terminal further receives second configuration information, which indicates the equalization algorithm for determining the second parameter. Based on this method, the measured value of the second parameter indicated by the first configuration information and the measured value of the second parameter at the second time point are calculated using the same equalization algorithm, thereby improving the accuracy of model performance evaluation.

[0293] In one optional implementation, the first configuration information is further used to configure the measured values ​​of the second parameter corresponding to the measured value range of the first parameter under different equalization algorithms. Based on this method, the terminal can select the measured values ​​of the second parameter corresponding to the measured value range of the first parameter under its own equalization algorithm to perform performance evaluation and improve the accuracy of model performance evaluation.

[0294] 303. The terminal determines the performance index of the AI ​​channel estimation model based on the measured value of the second parameter at the second time and the measured value of the second parameter at the first time.

[0295] For example, if the measured value of the second parameter at the second time step is not equal to (e.g., greater than) the measured value of the second parameter at the first time step, it indicates that the degree of dispersion between the equalized symbol and the corresponding transmitted symbol at the second time step has changed (e.g., more dispersed than at the first time step), indicating that the performance of the AI ​​channel estimation model has decreased compared to the first time step. If the measured value of the second parameter at the second time step is equal to or less than the measured value of the second parameter at the first time step, it indicates that the degree of dispersion between the equalized symbol and the corresponding transmitted symbol at the second time step remains unchanged or is smaller, indicating that the performance of the AI ​​channel estimation model remains unchanged or is improved compared to the first time step.

[0296] For example, the first configuration information indicates the measurement range of the second parameter. If the measured value of the second parameter at the second time is not included in the measurement range of the second parameter at the first time (e.g., the measured value of the second parameter at the second time is greater than the maximum measured value in the measurement range of the second parameter at the first time), it indicates that the dispersion between the symbol and the corresponding transmitted symbol after equalization at the second time has changed (e.g., it is more dispersed than at the first time), indicating that the performance of the AI ​​channel estimation model has decreased compared to the first time. If the measured value of the second parameter at the second time is included in the measurement range of the second parameter at the first time, it indicates that the dispersion between the symbol and the corresponding transmitted symbol after equalization at the second time remains unchanged or is smaller, indicating that the performance of the AI ​​channel estimation model remains unchanged or is improved compared to the first time.

[0297] One possible implementation is that the performance metrics of the AI ​​channel estimation model can include performance degradation, unchanged performance, or improvement. Another possible implementation is that the terminal can determine if the AI ​​channel estimation model's performance needs updating when it degrades to a certain level; otherwise, it can determine if an update is unnecessary.

[0298] 304. The terminal reports the performance metrics of the AI ​​channel estimation model to the network device.

[0299] Optionally, if the AI ​​channel estimation model does not need to be updated, then performance metrics do not need to be reported.

[0300] Based on this method, the terminal can determine and report the performance indicators of the AI ​​channel estimation model to the network side, which is conducive to timely updating of the AI ​​channel estimation model and improving the accuracy of channel estimation.

[0301] Both communication method 100 and communication method 200 provided in this application involve equalization algorithms. The equalization algorithm may be an implementation algorithm on the terminal side, and there is a possibility that the algorithm has been enhanced internally by the terminal. In communication method 100, since the processing based on the AI ​​channel estimation model and the processing based on the non-AI channel estimation algorithm are calculated within the same terminal and the same equalization algorithm is used, the impact of equalization can be disregarded. However, communication method 200 can indicate the correspondence between the measured value range of the first parameter and the measured value of the second parameter in the form of a predefined table. There is a possibility that the equalization algorithm of the second parameter when the table is defined may be different from the equalization algorithm used by the terminal after the actual deployment model, which may lead to inaccurate monitoring results. Therefore, communication method 200 can additionally specify the equalization algorithm used for performance evaluation through the second configuration information, for example, a fixed-based maximal ratio combining (MRC) / interference rejection combining (IRC) to provide equalization, so that the equalization algorithm of the second parameter when the table is defined is aligned with the equalization algorithm used by the terminal after the actual deployment model. The communication method 200 can configure a corresponding relationship table under different equalization algorithms through the first configuration information. The terminal can select the measurement value of the second parameter corresponding to the measurement range of the first parameter under its own equalization algorithm to perform performance evaluation and improve the accuracy of model performance evaluation.

[0302] In the communication methods 100 and 200 provided in this application, when the network side and the terminal side perform operations such as DMRS configuration signaling interaction and AI channel estimation model distribution, the deployment of the AI ​​channel estimation model on the network side and the terminal side may not be within the same physical entity as the terminal entity, and the signaling interaction may also be the interaction between other network elements where the model is deployed. The terminal-side AI channel estimation model may be implemented on the AI / GPU chip inside the terminal, or it may be implemented in a location outside the terminal, such as in the host or cloud server of an over-the-top (OTT) system.

[0303] For example, Figure 15 is a schematic diagram of the ORAN implementation provided in this application embodiment. The pre-trained AI channel estimation model is trained based on a static dataset. Training phase: The network side configures the DMRS configuration for training on the terminal and distributes the DMRS configuration; the terminal receives the DMRS based on the DMRS configuration, performs channel estimation, and uploads the corresponding received signal and channel estimation results for training the AI ​​channel estimation model; the network side can distribute the trained AI channel estimation model. The AI ​​channel estimation model is deployed in a location other than the terminal, such as the host of the OTT system or the cloud server. Model inference phase (or running phase): The network device configures and sends the DMRS according to the protocol; the terminal determines the input of the AI ​​channel estimation model, such as the received signal and the DMRS, and sends it to the OTT; the OTT performs channel estimation based on the AI ​​channel estimation model and outputs the channel estimation result to the terminal. Model performance monitoring: As the AI ​​channel estimation model runs in the actual scenario, the terminal monitors the performance of the AI ​​channel estimation model, such as using the communication method 100 and communication method 200 described in this application, and triggers model fine-tuning / redistribution after certain conditions are met.

[0304] The communication method provided by the embodiments of this application has been described above with reference to Figures 11 to 15. In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terms and / or descriptions between the various embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0305] The communication device provided in the embodiments of this application will be described below with reference to Figures 16 and 17. It should be understood that the description of the device embodiments corresponds to the description of the method embodiments. Therefore, for content not described in detail, please refer to the method embodiments above. For the sake of brevity, it will not be repeated here.

[0306] For example, FIG16 shows a possible exemplary block diagram of a communication device involved in an embodiment of this application. As shown in FIG16, the communication device may include modules or units for implementing the method embodiments described above. In one possible design, the communication device includes a communication unit 401 and a processing unit 402. Optionally, the communication device may further include a storage unit 403 for storing device program code and / or data.

[0307] The communication device can be the terminal device in the above embodiments, such as a terminal or a communication module in a terminal, or a circuit or chip in a terminal that is responsible for communication functions.

[0308] For example, in one embodiment, the processing unit 402 is used to: perform channel estimation and equalization processing on the first received signal based on the artificial intelligence (AI) channel estimation model to obtain symbol S1; perform channel estimation and equalization processing on the second received signal based on the non-AI channel estimation algorithm to obtain symbol S2; determine the performance index of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol; the first transmitted symbol and the second transmitted symbol have the same modulation method.

[0309] In one possible design, the processing unit 402 determines the performance index of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol, including: calculating a first parameter value based on symbol S1 and the first transmitted symbol; and calculating a second parameter value based on symbol S2 and the second transmitted symbol; the first parameter value and the second parameter value are used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol; and the performance index of the AI ​​channel estimation model is determined based on the first parameter value and the second parameter value.

[0310] In one possible design, the first parameter value is the Euclidean distance between symbol S1 and the first transmitted symbol; the second parameter value is the Euclidean distance between symbol S2 and the second transmitted symbol.

[0311] In one possible design, the processing unit 402 determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the ratio between the first parameter value and the second parameter value is less than or equal to a first threshold, it is determined that the AI ​​channel estimation model does not need to be updated; if the ratio between the first parameter value and the second parameter value is greater than the first threshold, it is determined that the AI ​​channel estimation model needs to be updated.

[0312] In one possible design, the processing unit 402 determines the performance index of the AI ​​channel estimation model based on the first parameter value and the second parameter value, including: if the difference between the first parameter value and the second parameter value is less than or equal to the second threshold, it is determined that the AI ​​channel estimation model does not need to be updated.

[0313] If the difference between the first parameter value and the second parameter value is greater than the second threshold, it indicates that the AI ​​channel estimation model needs to be updated.

[0314] In one possible design, the communication unit 401 is used to receive first configuration information, which is used to indicate the first parameter value and...

[0315] The calculation method for the second parameter value and the corresponding threshold.

[0316] In one possible design, the communication unit 401 is also used to receive second configuration information, which indicates a first resource for channel estimation based on an AI channel estimation model and a second resource for channel estimation based on a non-AI channel estimation algorithm.

[0317] In one possible design, the communication unit 401 is also used to report the performance metrics of the AI ​​channel estimation model, including whether the AI ​​channel estimation model does not need to be updated or needs to be updated.

[0318] For example, in another embodiment, the processing unit 402 is configured to: at a first moment, determine first configuration information, the first configuration information being used to indicate the measured value of a second parameter corresponding to the measured value range of a first parameter; the first parameter being used to indicate channel quality; the second parameter being determined based on an artificial intelligence (AI) channel estimation model, being used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol; based on the AI ​​channel estimation model and the first configuration information, determine the measured value range of the corresponding first parameter and the measured value of the second parameter at a second moment; based on the measured value of the second parameter at the second moment and the measured value of the second parameter at the first moment, determine the performance index of the AI ​​channel estimation model.

[0319] In one possible design, the communication unit 401 is used to report the performance metrics of the AI ​​channel estimation model, including whether the AI ​​channel estimation model does not need to be updated or needs to be updated.

[0320] In one possible design, the communication unit 401 is used to receive second configuration information, which is used to indicate the equalization algorithm for determining the second parameter.

[0321] In one possible design, the first configuration information is also used to configure the measurement value of the second parameter corresponding to the measurement range of the first parameter under different equalization algorithms.

[0322] In one possible design, when the communication device is a terminal or a communication module within a terminal, the functionality of the processing unit 402 can be implemented by one or more processors. Specifically, the processor may include a modem chip, or a system-on-a-chip (SoC) chip or a SIP chip containing a modem core. The functionality of the communication unit 401 can be implemented by transceiver circuitry.

[0323] In one possible design, when the communication device is a circuit or chip in a terminal responsible for communication functions, such as a modem chip or a system-on-a-chip (SoC) or SIP chip containing a modem core, the function of the processing unit 402 can be implemented by a circuit system in the aforementioned chip that includes one or more processors or processor cores. The function of the communication unit 401 can be implemented by an interface circuit or data transceiver circuit on the aforementioned chip.

[0324] In one possible design, when the communication device is a terminal or a processing module within a terminal, the functionality of the processing unit 402 can be implemented by one or more processors. Specifically, the processor may include a GPU, or a system-on-a-chip (SoC) or SIP chip containing a GPU. The functionality of the communication unit 401 can be implemented by transceiver circuitry.

[0325] In one possible design, when the communication device is a circuit or chip in the terminal responsible for processing functions, such as a GPU or a system-on-a-chip (SoC) or SIP chip containing a GPU, the function of the processing unit 402 can be implemented by a circuit system in the aforementioned chip that includes one or more processors or processor cores. The function of the communication unit 401 can be implemented by interface circuitry or data transceiver circuitry on the aforementioned chip.

[0326] The communication device can be a network-side device as described in the above embodiments, such as a network device or network apparatus.

[0327] For example, in one embodiment, processing unit 402 is used to determine a first transmitted symbol and a second transmitted symbol based on the same modulation scheme; communication unit 401 is used to transmit the first transmitted symbol and the second transmitted symbol. Processing unit 402 is used to receive the performance index of the AI ​​channel estimation model; the performance index is determined based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol. Symbol S1 is obtained by performing channel estimation and equalization processing on the first received signal based on the AI ​​channel estimation model; symbol S2 is obtained by performing channel estimation and equalization processing on the second received signal based on a non-AI channel estimation algorithm.

[0328] For example, in another embodiment, the communication unit 401 is used to send first configuration information, which indicates the measured value of the second parameter at a first moment corresponding to the measurement range of the first parameter; the first parameter indicates channel quality; the second parameter is determined based on an artificial intelligence (AI) channel estimation model and indicates the degree of dispersion between the equalized symbol and the corresponding transmitted symbol; the measured value of the second parameter at the first moment is used to determine the performance index of the AI ​​channel estimation model at a second moment corresponding to the measurement range of the first parameter; the communication unit 401 is used to receive the performance index of the AI ​​channel estimation model, which includes whether the AI ​​channel estimation model does not need to be updated or needs to be updated.

[0329] It is understood that the division of units in the above-described device is a logical functional division. One function can correspond to one functional unit, or two or more functions can be integrated into one functional unit. In actual implementation, all or some units can be integrated into one physical entity, or distributed across different physical entities. Furthermore, the aforementioned functional units can be implemented in hardware, software, or a combination of both. Whether a function is executed 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 specific applications, but such implementations should not be considered beyond the scope of this application.

[0330] In one example, the functional unit in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as: one or more application-specific integrated circuits (ASICs), or one or more central processing units (CPUs), one or more microcontroller units (MCUs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0331] In one example, storage unit 403 may include random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory and / or registers, etc.

[0332] For example, Figure 17 is a schematic diagram of the structure of a terminal provided in an embodiment of this application. This terminal can correspond to the terminal or terminal device shown in Figures 1 to 15 and is used to implement the operation of the terminal or terminal device in the above embodiments. As shown in Figure 17, the terminal includes: one or more antennas 510, a radio frequency processing system 520, and a processor system 530.

[0333] In the downlink or sidelink direction, the RF processing system 520 receives RF signals through the antenna 510 and sends the RF-processed signals to the processor system 530 for further processing. In the uplink or sidelink direction, the processor system 530 processes the terminal-side information and sends it to the RF processing system 520, which then processes the signal and transmits it through the antenna 510.

[0334] In one example, the radio frequency (RF) processing system 520 serves as the communication interface for external communication of the terminal and may include a radio frequency front end (RFFE) 521 and a radio frequency transceiver 522. The RFFE 521 is primarily used for one or more processing operations, such as shaping, passband selection, or gain adjustment, on the RF signals received by the antenna or the RF signals to be transmitted through the antenna. It may include one or more components such as RF switches, duplexers, filters, power amplifiers, antenna tuners, and low-noise amplifiers. The RFFE 521 can be a circuit system composed of multiple discrete devices or integrated into one or more chips. The RF transceiver 522 processes the RF signals received by the RFFE 521 into baseband / IF signals for further processing by the processor system 530, and processes the baseband / IF signals provided by the processor system 530 into RF signals for transmission to the RFFE 521. The baseband / IF signals transmitted between the RF transceiver 522 and the processor system 530 can be digital or analog signals. The radio frequency transceiver 522 can be implemented by one or more chips, which are usually referred to as radio frequency integrated circuits (RFICs).

[0335] In one example, processor system 530 may include one or more processors for processing signals and executing one or more communication protocols. Optionally, processor system 530 may also include memory 536. In one example, the one or more processors include at least one baseband processor 531 (also known as a modem processor). Memory 536 is used to store data and / or computer program instructions. Optionally, processor system 530 may also include one or more application processors 532 for implementing processing of the terminal operating system and application layer. Application processor 532 may include, for example, a GPU. Optionally, processor system 530 may also include one or more of a voice subsystem 533, a multimedia subsystem 534, or an interface circuit 535. The voice subsystem 533 is used to process voice signals, the multimedia subsystem 534 is used to handle multimedia-related operations, such as video encoding / decoding, image processing, etc., and the interface circuit 535 is used to implement communication with other terminal components, such as a display 540, an input device 550, memory 560, etc. The above-mentioned components in processor system 530 can communicate with each other via a bus or communication interface circuit.

[0336] In one example, the processor system 530 can be packaged as a single processor chip, such as a SoC chip or a SIP chip. In another example, the processor system 530 can be a system of multiple chips, for example, the baseband processor 531 can be packaged as a single chip, or packaged with part or all of the circuitry of the radio frequency processing system into a single chip.

[0337] In one example, memory 536 can be on-chip memory, i.e., located on the processor system 530 chip. In another example, memory 560 can be off-chip memory, i.e. located outside the processor system 530 chip.

[0338] In one example, the baseband processor 531 may include one or more processor cores 5311 and interface circuitry 5314. The one or more processor cores 5311 are used to process signals and execute one or more communication protocols. Optionally, the baseband processor 531 may also include a memory 5312 for storing at least a portion of the corresponding computer program instructions and / or data. In one example, the one or more processor cores 5311 execute the computer program instructions stored in the memory 5312 to implement the relevant operations (such as generating and sending first information) in the above method embodiments. In this application, the memory 5312 storing the corresponding computer program instructions and / or data may mean that the memory 5312 stores all the corresponding computer program instructions and / or data for the processor core 5311 to execute; or it may mean that the memory 5312 stores a portion of the corresponding computer program instructions and / or data, which includes the computer program instructions and / or data that the processor core 5311 currently needs to execute. The memory 5312 can store different portions of computer program instructions and / or data multiple times for the processor core 5311 to execute in order to implement the relevant operations in the above method embodiments. Interface circuit 5314 serves as a communication interface for communication with other components, such as transmitting signals with RF processing system 520, communicating with other subsystems and related components of processor system 530 via bus, such as transmitting data control signals with application processor 532, and transmitting data or computer program instructions with memory 536 or memory 560. Optionally, to reduce the load on the processor core, baseband signal processing circuit 5313 can also be provided to perform at least some baseband signal processing, including one or more of signal demodulation, modulation, encoding, or decoding.

[0339] The aforementioned processors, processor systems, application processors, baseband processors, processor circuits, or processor cores can be collectively referred to as processors. These processors may include one or more combinations of a central processing unit (CPU), a digital signal processor (DSP), a microprocessor unit (MPU), a microcontroller unit (MCU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), an artificial intelligence processor (AI processor), or a neural network processing unit (NPU). Some or all steps of the communication method in the embodiments of this application can be implemented by a GPU or NPU, or by a GPU or NPU in conjunction with other processors.

[0340] The aforementioned memory may include one or more of the following storage media: random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), phase-change memory (PCM), resistive random access memory (RERAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), cache, register, read-only memory (ROM), flash memory, erasable programmable read-only memory (EPROM), hard disk, etc. In one example, computer program instructions for executing the above embodiments may be stored in non-volatile memory, such as at least a portion of the aforementioned memory 560 (e.g., one or more of ROM, flash memory, EPROM, or hard disk). When the terminal is running, the corresponding computer program instructions may be partially or wholly loaded onto a memory with a faster transfer speed than the processor, such as at least a portion of memory 536 and / or memory 5312 (e.g., one or more of RAM, SRAM, DRAM, PCM, RERAM, MRAM, FRAM, cache, or register), for the processor to execute in order to implement the steps in the above method embodiments.

[0341] In one example, the RF transceiver 522 and the RF front-end 521 can also be packaged in a single chip. In another example, the RF transceiver 522, the RF front-end 521, and the baseband processor 531 can also be packaged in a single chip.

[0342] The terms "system" and "network" in the embodiments of this application may be used interchangeably. "At least one" means one or more, and "multiple" means two or more.

[0343] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0344] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0345] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0346] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0347] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A communication method, characterized in that, The method includes: Based on the artificial intelligence (AI) channel estimation model, channel estimation and equalization processing are performed on the first received signal to obtain symbol S1; The second received signal is channel estimated and equalized based on a non-AI channel estimation algorithm to obtain symbol S2. The performance index of the AI ​​channel estimation model is determined based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol; the modulation methods of the first transmitted symbol and the second transmitted symbol are the same.

2. The method according to claim 1, characterized in that, The determination of the performance metrics of the AI ​​channel estimation model based on the degree of dispersion between symbol S1 and the first transmitted symbol, and the degree of dispersion between symbol S2 and the second transmitted symbol, includes: A first parameter value is calculated based on symbol S1 and the first transmitted symbol; and a second parameter value is calculated based on symbol S2 and the second transmitted symbol; the first parameter value and the second parameter value are used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol; Based on the first parameter value and the second parameter value, the performance index of the AI ​​channel estimation model is determined.

3. The method according to claim 2, characterized in that, The first parameter value is the Euclidean distance between the symbol S1 and the first transmitted symbol; The second parameter value is the Euclidean distance between symbol S2 and the second transmitted symbol.

4. The method according to claim 2 or 3, characterized in that, The step of determining the performance metrics of the AI ​​channel estimation model based on the first parameter value and the second parameter value includes: If the ratio between the first parameter value and the second parameter value is less than or equal to the first threshold, it is determined that the AI ​​channel estimation model does not need to be updated. If the ratio between the first parameter value and the second parameter value is greater than the first threshold, it is determined that the AI ​​channel estimation model needs to be updated.

5. The method according to claim 2 or 3, characterized in that, The step of determining the performance metrics of the AI ​​channel estimation model based on the first parameter value and the second parameter value includes: If the difference between the first parameter value and the second parameter value is less than or equal to the second threshold, it is determined that the AI ​​channel estimation model does not need to be updated. If the difference between the first parameter value and the second parameter value is greater than the second threshold, it is determined that the AI ​​channel estimation model needs to be updated.

6. The method according to claim 4 or 5, characterized in that, The method further includes: Receive first configuration information, which is used to indicate the calculation method between the first parameter value and the second parameter value and the corresponding threshold.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Receive second configuration information, which is used to indicate a first resource for channel estimation based on the AI ​​channel estimation model and a second resource for channel estimation based on the non-AI channel estimation algorithm.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: Report the performance metrics of the AI ​​channel estimation model, including whether the AI ​​channel estimation model does not need to be updated or needs to be updated.

9. A communication method, characterized in that, The method includes: At the first moment, first configuration information is determined, which is used to indicate the measurement value of the second parameter corresponding to the measurement range of the first parameter; the first parameter is used to indicate the channel quality; the second parameter is determined based on the artificial intelligence (AI) channel estimation model and is used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. Based on the AI ​​channel estimation model and the first configuration information, the range of measurement values ​​corresponding to the first parameter is determined, and the measurement value of the second parameter at the second time point is determined. The performance index of the AI ​​channel estimation model is determined based on the measured value of the second parameter at the second time point and the measured value of the second parameter at the first time point.

10. The method according to claim 9, characterized in that, The method further includes: Report the performance metrics of the AI ​​channel estimation model, including whether the AI ​​channel estimation model does not need to be updated or needs to be updated.

11. The method according to claim 9 or 10, characterized in that, The method further includes: Receive second configuration information, which is used to indicate the equalization algorithm for determining the second parameter.

12. The method according to any one of claims 9 to 11, characterized in that, The first configuration information is also used to configure the measurement value of the second parameter corresponding to the measurement range of the first parameter under different equalization algorithms.

13. A communication method, characterized in that, The method includes: Send first configuration information, which is used to indicate the measurement value of the second parameter at the first moment corresponding to the measurement range of the first parameter; the first parameter is used to indicate channel quality; the second parameter is determined based on an artificial intelligence (AI) channel estimation model and is used to indicate the degree of dispersion between the equalized symbol and the corresponding transmitted symbol. The measured value of the second parameter at the first moment is used to determine the performance index of the AI ​​channel estimation model at the second moment corresponding to the range of measured values ​​of the first parameter; Receive the performance metrics of the AI ​​channel estimation model, wherein the performance metrics include whether the AI ​​channel estimation model does not need to be updated or needs to be updated.

14. The method according to claim 13, characterized in that, The method further includes: Send second configuration information, which is used to indicate the equalization algorithm for determining the second parameter.

15. The method according to claim 13 or 14, characterized in that, The first configuration information is also used to configure the measurement value of the second parameter corresponding to the measurement range of the first parameter under different equalization algorithms.

16. A communication device, characterized in that, Includes units or modules for implementing the method as described in any one of claims 1 to 15.

17. A communication device, characterized in that, The communication device includes at least one processor; the at least one processor is configured to enable the communication device to implement the method as described in any one of claims 1 to 15.

18. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed, cause the computer to perform the method as described in any one of claims 1 to 15.

19. A computer program product, characterized in that, The computer program product includes: computer program code, which, when executed by a computer, causes the computer to perform the method as described in any one of claims 1 to 15.

20. A communication device, characterized in that, The communication device includes logic circuitry and an interface, the interface being used for inputting and / or outputting information, and the logic circuitry being used to cause the communication device to perform the method as described in any one of claims 1 to 15.

21. A chip, characterized in that, It includes at least one processor, the processor being configured to execute instructions to cause a communication device including the chip to perform the method as described in any one of claims 1 to 15.

22. The chip according to claim 21, characterized in that, The chip also includes an interface circuit, which is used to receive the executed instructions and transmit them to the processor, or to output information from the processor.

Images