CSI performance monitoring report triggering method based on AI / ML model and related device

Abstract

The embodiment of the invention provides a CSI (Channel State Information) performance monitoring report triggering method based on an AI / ML (Artificial Intelligence / Markup Language) model and a related device, and provides a channel environment adaptive cumulative sum CUSUM algorithm, the algorithm configures different CUSUM parameter sets for different states of UE (User Equipment) respectively, each CUSUM parameter set comprises a relaxation factor k and a decision threshold h, and the relaxation factor k and the decision threshold h correspond to each other. The values of k and h in different parameter sets are different; the UE can select the most matched CUSUM parameter according to the current state (or scene), so that a performance monitoring mechanism of the UE can intelligently balance the detection sensitivity and robustness. Moreover, according to the scheme, performance differences are accumulated through a CUSUM algorithm and then compared with a decision threshold value to judge whether a performance degradation report is triggered or not, instantaneous jitter of a channel is effectively filtered through accumulation of the performance differences, and it is ensured that the report is triggered only when the performance of the AI / ML model continuously declines. Therefore, the uplink signaling overhead caused by mistaken touch and the resource waste caused by subsequent network response (such as mode switching) are reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, and particularly relates to a CSI performance monitoring report triggering method based on an AI / ML model and related devices. BACKGROUND

[0002] In a wireless communication system, a large-scale multiple input multiple output (MIMO) technology can be used to improve system capacity and spectrum efficiency. In the MIMO technology, a terminal device needs to measure a channel state information reference signal (CSI-RS) sent by a network device and report channel state information (CSI), and the network device performs precoding according to the CSI, so as to send data after precoding to the terminal device.

[0003] Using an AI (artificial intelligence) / ML (machine learning) based compression algorithm can further enhance CSI feedback, reduce CSI feedback overhead and improve accuracy. In the AI / ML based compression algorithm, the 3GPP RAN1 working group has reached a consensus that an encoder and a decoder based on an AI / ML model are respectively deployed on a terminal device and a network device (such as a base station), which is called a bilateral AI / ML model. Specifically, the terminal device uses an AI / ML based CSI encoder to generate CSI feedback information, and the network device uses a corresponding AI / ML based CSI decoder to reconstruct the CSI according to the received CSI feedback information. Based on the bilateral AI / ML model, the model performance needs to be continuously monitored in actual deployment to ensure its reliability and stability in a dynamic environment.

[0004] Among them, one performance monitoring method is UE side performance monitoring based on precoded CSI-RS. The core idea of this monitoring method is that the network device implicitly transmits the decoded CSI using precoded CSI-RS, and the UE evaluates the performance of the AI / ML model by comparing the precoded signal with the original channel measurement value. If a single measurement result is used to trigger a report, for example, the performance of the AI / ML model is reported as long as it is lower than that of the traditional codebook method, which is easily affected by instantaneous channel fluctuations, leading to unnecessary reports and potential mode switching, and increasing signaling overhead. SUMMARY

[0005] Therefore, the present application provides an AI / ML model-based CSI performance monitoring report triggering method and related device to solve at least part of the above problems. The disclosed technical solutions are as follows.

[0006] In a first aspect, the present application provides an AI / ML model-based CSI performance monitoring report triggering method, which is executed by a terminal device. The method includes: obtaining a performance difference index between a traditional codebook scheme and an AI / ML model corresponding to a current monitoring time; determining a target CUSUM parameter set based on a current state of the terminal device and CUSUM parameter set configuration information, different states of the terminal device corresponding to different CUSUM parameter sets, each CUSUM parameter set including a relaxation factor and a decision threshold, and the relaxation factor and the decision threshold in different CUSUM parameter sets being different; updating a performance difference cumulative sum of the current monitoring time, the performance difference cumulative sum of the current monitoring time being the part of the performance difference index of the current monitoring time exceeding a preset cumulative threshold based on the performance difference cumulative sum corresponding to a previous monitoring time, the preset cumulative threshold including the sum of a target offset and the relaxation factor in the target CUSUM parameter set, and the target offset being a benchmark value for measuring the performance difference index; and if the performance difference cumulative sum is greater than the decision threshold in the target CUSUM parameter set, sending a performance degradation report to a network device, the performance degradation report being used to indicate that the performance of the AI / ML model is lower than that of the traditional codebook scheme.

[0007] It can be seen that the scheme configures different CUSUM parameter sets for different states of the UE, each CUSUM parameter set including a relaxation factor k and a decision threshold h, and the values of k and h in different parameter sets being different; the UE can select the most suitable CUSUM parameter according to the current state (or scenario), so that the performance monitoring mechanism of the UE can intelligently balance the detection sensitivity and robustness. The scheme compares the cumulative sum of the performance difference with the decision threshold to determine whether to trigger the performance degradation report, effectively filters the instantaneous jitter of the channel through the accumulation of the performance difference, and ensures that the report is triggered only when the performance of the AI / ML model continuously decreases, thereby reducing the uplink signaling overhead caused by false triggering and the resource waste caused by subsequent network response (such as mode switching).

[0008] In a possible implementation, the method further includes: if a sum of the performance difference accumulation at the last monitoring time and a first difference value is less than or equal to 0, setting the performance difference accumulation at the current monitoring time to 0, the first difference value being a difference between the performance difference indicator at the current monitoring time and the preset accumulation threshold. In this case, the performance of the AI-ML model at the current monitoring time is better than that at the last monitoring time, that is, the performance degradation at the last monitoring time is a temporary degradation, and the accumulation is only performed for a continuous degradation, and therefore the performance difference accumulation at the current monitoring time is set to 0, the accumulation of the performance difference is re-counted, and the influence of the temporary performance degradation on the performance degradation judgment is avoided.

[0009] In a possible implementation, the performance difference accumulation at the current monitoring time is updated based on the performance difference indicator and a relaxation factor in the target CUSUM parameter set, and the updating includes: updating the performance difference accumulation at the current monitoring time according to the following formula:

[0010]

[0011] wherein S t is the performance difference accumulation at the current monitoring time t, S t-1 is the performance difference accumulation at the last monitoring time t-1, y t is the performance difference indicator at the current monitoring time t, k t is the relaxation factor used at the current monitoring time, and is a target offset.

[0012] It can be seen that according to the scheme, the performance difference accumulation S t is increased only when the performance difference indicator y t at the current monitoring time is greater than the sum of the target offset and the relaxation factor k t , the instantaneous jitter of the channel is effectively filtered, and the performance of the AI / ML model is ensured to be continuously decreased before triggering the report, thereby reducing the uplink signaling overhead caused by false triggering and the resource waste caused by subsequent network responses (such as mode switching).

[0013] In another possible implementation, the method further includes: after sending the performance degradation report to the network device, setting the performance difference accumulation at the current monitoring time to 0. In this way, after triggering the performance degradation report, the performance difference accumulation at the current monitoring time is set to 0, the performance difference is accumulated again from the next monitoring time, the influence of the performance difference accumulation caused by triggering the report is avoided, and the performance degradation report is prevented from being frequently triggered, thereby saving the signaling overhead and resources.

[0014] In another possible implementation, before obtaining the performance difference index between the traditional codebook scheme and the AI / ML model corresponding to the current monitoring moment, the method further includes: receiving a first message from the network device, the first message including performance-aware configuration information, the performance-aware configuration information including CUSUM parameter set configuration information and parameter set mapping rules, the CUSUM parameter set configuration information including a plurality of CUSUM parameter sets, and the parameter set mapping rules including a mapping relationship between different state parameters of the terminal device and the CUSUM parameter sets.

[0015] In another possible implementation, the first message is an RRC reconfiguration message. In this way, the CUSUM parameter sets and the parameter set mapping rules are configured to the UE through the RRC reconfiguration message, without the need to configure by using a brand-new signaling, thereby improving signaling use efficiency, reducing signaling overhead, and saving resources.

[0016] In another possible implementation, the newly added information element in the RRC reconfiguration message is used to carry the performance-aware configuration information. As can be seen, the corresponding configuration information is carried through the new information element in the RRC reconfiguration message, without the need to configure by using a brand-new signaling, thereby improving signaling use efficiency, reducing signaling overhead, and saving resources.

[0017] In another possible implementation, the state parameters of the terminal device include at least one of a mobility state, a channel quality, a service type, and a self state of the terminal device, and the self state of the terminal device includes at least one of a remaining power, a running temperature, a beam management state, and a hardware capability. In this way, the terminal device can select the most suitable CUSUM parameter set according to the state of the current environment or the self state, so that the CUSUM algorithm can be self-adapted to the current environment state or the self state of the terminal.

[0018] In another possible implementation, the mapping relationship between the mobility state and the CUSUM parameter set includes: the first CUSUM parameter set corresponding to the stationary state, the first CUSUM parameter set including a first relaxation factor and a first decision threshold; the second CUSUM parameter set corresponding to the low mobility state, the second CUSUM parameter set including a second relaxation factor and a second decision threshold; the third CUSUM parameter set corresponding to the medium mobility state, the third CUSUM parameter set including a third relaxation factor and a third decision threshold; the fourth CUSUM parameter set corresponding to the high mobility state, the fourth CUSUM parameter set including a fourth relaxation factor and a fourth decision threshold; the values of the first relaxation factor, the second relaxation factor, the third relaxation factor, and the fourth relaxation factor increase in turn, and the values of the first decision threshold, the second decision threshold, the third decision threshold, and the fourth decision threshold increase in turn. In this way, different CUSUM parameter sets are configured for different mobility states, so that the terminal can select the most suitable parameter set for performance monitoring according to its own mobility state, thereby making the sensitivity and robustness of the performance monitoring of the AI / ML model more in line with the current mobility state.

[0019] In another possible implementation, the mapping relationship between the channel quality and the parameter set includes: the fifth CUSUM parameter set corresponding to the excellent channel quality, the fifth CUSUM parameter set including a fifth relaxation factor and a fifth decision threshold; the sixth CUSUM parameter set corresponding to the medium channel quality, the sixth CUSUM parameter set including a sixth relaxation factor and a sixth decision threshold; the seventh CUSUM parameter set corresponding to the poor channel quality, the seventh CUSUM parameter set including a seventh relaxation factor and a seventh decision threshold; the values of the fifth relaxation factor, the sixth relaxation factor, and the seventh relaxation factor increase in turn; and the values of the fifth decision threshold, the sixth decision threshold, and the seventh decision threshold increase in turn. In this way, the terminal can select a suitable CUSUM parameter set according to the current channel quality, and intelligently balance the sensitivity and robustness of performance monitoring according to the current channel quality.

[0020] In another possible implementation, the mapping relationship between the service type of the terminal device and the parameter set includes: a URLLC type corresponds to a ninth CUSUM parameter set, the ninth CUSUM parameter set includes a ninth relaxation factor and a ninth decision threshold; an eMBB service corresponds to a tenth CUSUM parameter set, the tenth CUSUM parameter set includes a tenth relaxation factor and a tenth decision threshold; a massive machine type communication service corresponds to an eleventh CUSUM parameter set, the eleventh CUSUM parameter set includes an eleventh relaxation factor and an eleventh decision threshold; the values of the ninth relaxation factor, the tenth relaxation factor, and the eleventh relaxation factor increase in turn, and the values of the ninth decision threshold, the tenth decision threshold, and the eleventh decision threshold increase in turn. In this way, the terminal can select a suitable CUSUM parameter set according to the current service type, and intelligently balance the sensitivity and robustness of performance monitoring according to the current service type.

[0021] In another possible implementation, the performance degradation report includes an AI / ML model performance degradation indication and diagnosis information, and the diagnosis information includes performance degradation information when the terminal device triggers the performance degradation report.

[0022] In another possible implementation, the diagnosis information includes at least one of a severity level and a CUSUM parameter set ID used when triggering the performance degradation report, and the severity level indicates a degree of performance degradation of the AI / ML model relative to a traditional codebook scheme.

[0023] In another possible implementation, sending the performance degradation report to the network device includes: carrying the performance degradation report through a dedicated MAC CE and sending. In this way, the network device is provided with more dimensional performance degradation information with very small additional overhead, so that the network device can obtain the severity of the degradation and the scenario in which the UE is located when the problem occurs after receiving the performance degradation report, and further make the response strategy of the network device highly differentiated and intelligent.

[0024] In a second aspect, the present application also provides an AI / ML model-based CSI performance monitoring report triggering method, executed by a network device, including: receiving a performance degradation report from a terminal device, the performance degradation report being sent when the terminal device monitors that a performance difference accumulation at a current monitoring moment is greater than a decision threshold in a target CUSUM parameter set; wherein the target CUSUM parameter set is determined by the terminal device according to a current state of the terminal device, different states of the terminal device correspond to different CUSUM parameter sets, each CUSUM parameter set includes a relaxation factor and a decision threshold, and the relaxation factor and the decision threshold in different CUSUM parameter sets are different; the performance difference accumulation at the current monitoring moment is obtained by the terminal device based on a performance difference index at the current monitoring moment and the relaxation factor in the target CUSUM parameter set, and the performance difference accumulation at the current monitoring moment is the part of the performance difference index at the current monitoring moment exceeding a preset accumulation threshold added to a performance difference accumulation at a previous monitoring moment corresponding to the performance difference accumulation.

[0025] In a third aspect, the present application also provides a communication apparatus, including a module for executing the method of any one of the first aspect, or a module for executing the method of any one of the second aspect.

[0026] In a fourth aspect, the present application also provides a communication apparatus, including: a memory for storing computer programs or computer instructions; and a processor for executing the computer programs or computer instructions stored in the memory, so that the communication apparatus executes the method of any one of the first aspect, or the method of any one of the second aspect.

[0027] In a fifth aspect, the present application also provides a computer storage medium for storing a computer program, the computer program being executed to implement the method of any one of the first aspect or the second aspect.

[0028] In a sixth aspect, the present application also provides a computer program product, whose computer program is executed to implement the method of any one of the first aspect or the second aspect. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 A structural schematic diagram of a communication system provided by the present application;

[0030] Figure 2 A structural schematic diagram of an access network device provided by the present application;

[0031] Figure 3 A flowchart of a performance monitoring method of a conventional AI / ML model provided by the present application;

[0032] Figure 4 A schematic diagram of a gNB transmitting signal in AI / ML model performance monitoring provided by the present application;

[0033] Figure 5 A flowchart of an AI / ML model performance monitoring method provided by the present application;

[0034] Figure 6 A structural schematic diagram of a communication device provided by the present application;

[0035] Figure 7 A structural schematic diagram of another communication device provided by the present application. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. The terms used in the following embodiments are only for the purpose of describing the specific embodiments and are not intended to be limiting to the present application. As used in the specification and the appended claims of the present application, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “one or more,” in the embodiments of the present application, refer to one, two, or more than two; “and / or” describes the associated relationship of the associated objects, which means that there can be three relationships; for example, A and / or B can represent the following three cases: A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. The character “ / ” generally represents an “or” relationship between the associated objects.

[0037] In the present specification, the reference to “one embodiment” or “some embodiments” or the like means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, the appearance of the phrases “in one embodiment,” “in some embodiments,” “in other embodiments,” “in additional embodiments,” and so on, in various places in the specification is not necessarily all referring to the same embodiment, unless otherwise specifically noted. The terms “comprising,” “including,” “having,” and their variants, mean “including but not limited to,” unless otherwise specifically indicated.

[0038] The plurality referred to in the embodiments of the present application means greater than or equal to two. It should be noted that in the description of the embodiments of the present application, the terms “first,” “second,” and the like are used only for the purpose of distinguishing the described objects, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying an order.

[0039] The technical solutions provided by the embodiments of the present application can be applied to a communication system, which can include but is not limited to the following systems, for example: a second generation (2G) communication system, a third generation (3G) communication system, a long term evolution (LTE) system, a universal mobile telecommunications system (UMTS), a worldwide interoperability for microwave access (WiMAX) communication system, a 5th generation (5G) system or new radio (NR), a 5.5G system or a 6th generation (6G) system and future mobile communication systems, vehicle to X (V2X); V2X can include vehicle to network (V2N), vehicle to vehicle (V2V), vehicle to infrastructure (V2I), vehicle to pedestrian (V2P), etc., long term evolution-vehicle (LTE-V), Internet of Vehicles, machine type communication (MTC), Internet of Things (IOT), ambient Internet of Things (AIOT), long term evolution-machine (LTE-M), machine to machine (M2M), etc.

[0040] The scenarios to which the communication system is applicable can include: terrestrial cellular communication, non-terrestrial network (NTN), satellite communication, high altitude platform station (HAPS) communication, vehicle to everything (V2X) communication, integrated access and backhaul (IAB) communication, reconfigurable intelligent surface (RIS) communication, etc.

[0041] For example,Figure 1 A schematic diagram of an architecture of a communication system is shown.

[0042] As shown in Figure 1 , the communication system includes a network device 101 and a terminal device 102.

[0043] The network device 101 can be a device used for providing network communication function on the network side, and is also called network element in some cases. The network device can be generally a base station (including a functional unit of the base station or a combination of functional units of the base station) or a core network unit. The core network unit can be a functional unit in the core network, including but not limited to an access and mobility management function (AMF) unit or a session management function (SMF) unit.

[0044] In the embodiments of the present application, the base station can be any device with wireless transceiving function, including but not limited to: an evolved Node B (eNB or e-NodeB) in long term evolution (LTE), a base station (gNodeB or gNB) or a transmission receiving point (TRP) in new radio (NR), a base station in subsequent evolution of 3GPP, an access node in a Wi-Fi system, a wireless relay node, a wireless backhaul node, etc. The base station can be: a macro base station, a micro base station, a pico base station, a small station, a relay station, or a balloon station, etc. The base station can include one or more co-sited or non-co-sited transmission reception points (TRPs). The base station can also be a wireless controller, a centralized unit (CU), and / or a distributed unit (DU) in a cloud radio access network (CRAN) scenario. The base station can communicate with the terminal device 102, or communicate with the terminal device 102 through a relay station. The terminal device can communicate with multiple base stations of different technologies, for example, the terminal device can communicate with a base station supporting an LTE network, and can also communicate with a base station supporting a 5G network, and can also communicate with a base station supporting an LTE network and a base station supporting a 5G network in dual connectivity.

[0045] In actual application, the network device, as an access network device, can be cooperated by multiple network devices to assist terminal devices to implement wireless access, and different network devices respectively implement part of functions of a base station. For example, the network device can be a central unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU), etc. The CU and the DU can be separately arranged or can be included in the same network element, such as a baseband unit (BBU). The RU can be included in a radio frequency device or a radio frequency unit, such as a remote radio unit (RRU), an active antenna processing unit (AAU), or a remote radio head (RRH).

[0046] In different systems, the CU (or CU-CP and CU-UP), DU or RU can also have different names, but those skilled in the art can understand their meanings. For example, in an ORAN system, the CU can also be referred to as an O-CU (open CU), the DU can also be referred to as an O-DU, the CU-CP can also be referred to as an O-CU-CP, the CU-UP can also be referred to as an O-CU-UP, and the RU can also be referred to as an O-RU. Any one of the CU (or CU-CP, CU-UP), DU and RU in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module. The CU (or CU-CP and CU-UP), DU and RU can implement different protocol layer functions.

[0047] Figure 2 is a structural schematic diagram of an access network device. As an implementation example, as shown in Figure 2As shown, an access network device may include at least one CU and at least one DU. This design can be referred to as CU and DU separation. A CU can be connected to one or more DUs. CU and DU can be separated according to the protocol layer of the wireless network: for example, the functions of the Packet Data Convergence Protocol (PDCP) layer and above (such as the Radio Resource Control (RRC) layer and the Service Data Adaptation Protocol (SDAP) layer) are set in the CU, while the functions of the protocol layers below the PDCP layer (such as the Radio Link Control (RLC) layer, the Media Access Control (MAC) layer, and the Physical (PHY) layer) are set in the DU; or, for example, the functions of the protocol layers above the PDCP layer are set in the CU, while the functions of the protocol layers below the PDCP layer are set in the DU, without restriction. When the CU includes CU-CP and CU-UP, CU-CP is used to implement the control plane functions of the CU, and CU-UP is used to implement the user plane functions of the CU. For example, when a CU is configured to implement the functions of the PDCP, RRC, and SDAP layers, CU-CP is used to implement the RRC layer functions and the control plane functions of the PDCP layer, while CU-UP is used to implement the SDAP layer functions and the user plane functions of the PDCP layer. This application does not limit the names of CU and DU. The above division of CU and DU processing functions according to protocol layers is merely an example; other division methods are also possible.

[0048] The CU can be connected to the core network. Optionally, the CU can have some of the functions of the core network.

[0049] Furthermore, some functions of the DU can be separated and configured. For example... Figure 2As shown, the part of the functions can be implemented by a radio unit (RU). The RU can have radio frequency functions. The name of the RU is not limited in the present application. The DU and the RU can be split or separated at the PHY layer. For example, the DU can implement high layer functions in the PHY layer, and the RU can implement low layer functions in the PHY layer or implement the low layer functions and radio frequency functions. The high layer functions in the PHY layer include functions closer to the MAC layer, and the low layer functions in the PHY layer include functions closer to the radio frequency. For example, the high layer functions of the PHY layer include one or more of the following: forward error correction (FEC) encoding / decoding, scrambling, or modulation / demodulation. The low layer functions of the PHY layer include one or more of the following: fast Fourier transform (FFT) / inverse fast Fourier transform (IFFT), beamforming, or extraction and filtering of a physical random access channel (PRACH), etc. The RU can communicate radio frequency signals with the terminal device through an air interface. The pre-coding function of the PHY layer code can be located in the DU or in the RU. The split manner between the DU and the RU can be various possible manners and is not limited. There is an interface between the DU and the RU. For example, according to different split manners, the interface between the DU and the RU can be a common public radio interface (CPRI) interface or an enhanced common public radio interface (eCPRI) interface.

[0050] Optionally, any of the above CU, CU-CP, CU-UP, DU and RU can be a software module, a hardware structure, or a software module plus a hardware structure, which is not limited. The forms of existence of different entities can be the same or different. For example, the CU, the CU-CP, the CU-UP and the DU are software modules, and the RU is a hardware structure. For the sake of brevity of description, all possible combination forms are not listed one by one here. These modules and the methods performed by them are also within the protection scope of the embodiments of the present application. For example, when the method of the embodiments of the present application is performed by an access network device, it can be specifically performed by at least one of the CU, the CU-CP, the CU-UP, the DU or the RU.

[0051] For example, the receiving circuit of the terminal device 102 can include a main radio (MR) and a low power radio (LR).

[0052] In the embodiments of the present application, the terminal device can be various forms, for example, a mobile phone, a tablet computer, a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a vehicle-mounted terminal device, a wireless terminal in self driving, a wireless terminal in remote medical, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, a wearable terminal device, and the like. The terminal can also be referred to as a terminal device, a user equipment (UE), an access terminal device, a vehicle-mounted terminal, an industrial control terminal, a UE unit, a UE station, a mobile station, a mobile station, a remote station, a remote terminal device, a mobile device, a UE terminal device, a wireless communication device, a UE agent, or a UE apparatus, and the like. The terminal can also be a fixed terminal or a mobile terminal.

[0053] In some embodiments, the communication system can further include other devices in communication with the network device and / or the interrupt device, which is not limited in the present application.

[0054] As shown in Figure 3 The basic flow of the UE-side performance monitoring scheme based on the precoded CSI-RS is as follows:

[0055] S1, the gNB (base station) sends a downlink CSI-RS signal.

[0056] S2, the UE measures the downlink CSI-RS signal and obtains CSI feedback information based on the AI / ML model.

[0057] S3, the UE sends the CSI feedback information to the gNB.

[0058] The UE measures the downlink CSI-RS signal sent by the gNB at T0 to obtain the original channel matrix H, and obtains the precoding matrix V after singular value decomposition (SVD) of the channel matrix H. The precoding matrix V is subjected to feature extraction, compression coding and quantization by the encoder in the UE-side AI / ML model to obtain a bit stream, which is the CSI feedback information. The UE reports the CSI feedback information model to the gNB.

[0059] Compressed precoding matrix (also known as implicit CSI feedback) can save feedback overhead, and is consistent with the traditional codebook-based CSI feedback framework, so the AI / ML model-based CSI feedback is of great concern in 3GPP research.

[0060] S4, the gNB sends the standard CSI-RS and the precoded CSI-RS to the UE.

[0061] The standard CSI-RS is not precoded, and the UE uses it to measure the complete channel information.

[0062] The precoded CSI-RS is obtained by precoding based on the CSI feedback information.

[0063] Precoding is a processing performed on the signal at the signal transmitting end (gNB). The gNB uses the CSI information it obtains to weight and phase adjust the signal to be transmitted, with the purpose of making the signal energy more accurately aligned with the UE to overcome the fading and interference in the wireless channel transmission process. Mathematically, precoding is equivalent to multiplying the data symbol vector to be sent by a precoding matrix, which is calculated based on the CSI.

[0064] The gNB uses the precoding matrix indicated by the first CSI feedback information reported by the UE at T0 to perform precoding to obtain the first precoded CSI-RS. At the same time, the gNB uses the decoder in the AI / ML model to decode the bit stream (i.e. the second CSI feedback information) reported by the UE at T0 to obtain the reconstructed precoding matrix , and uses the reconstructed precoding matrix to perform precoding on the CSI-RS signal to obtain the second precoded CSI-RS.

[0065] The gNB transmits the first precoded CSI-RS and the second precoded CSI-RS at T1.

[0066] As shown in Figure 4 , at T0, the gNB transmits a set of CSI-RS resources for measurement to the UE, and the UE measures the CSI-RS signal to obtain the CSI and feeds back two kinds of CSI to the gNB, one is the CSI feedback information based on the traditional codebook scheme, and the other is the compressed CSI feedback information based on the AI / ML model. At T1, the gNB transmits two sets of CSI-RS resources to the UE, one is the CSI-RS for channel inference, which focuses on the real-time information of the current channel; the other is the CSI-RS for performance monitoring, which uses the precoding matrix obtained by the gNB decoding the second CSI feedback information to precoding the reference signal, which focuses on the quality after AI / ML model processing at T0 in the past, and performs performance evaluation at the current time T1.

[0067] S5, the UE calculates a first performance indicator of a legacy codebook scheme based on the measurement result of the standard CSI-RS, and a second performance indicator of the AI / ML model based on the precoded CSI-RS.

[0068] The UE receives two sets of CSI-RS resources at T1, i.e., a set of standard CSI-RS resources and a set of precoded CSI-RS.

[0069] In the performance monitoring scenario, the UE measures the full channel information using the standard CSI-RS, and calculates a first performance indicator of the legacy codebook scheme, such as the equivalent signal to interference plus noise ratio (SINR) or the hypothetical block error rate (BLER). The equivalent SINR or the hypothetical BLER represents the performance of the legacy codebook scheme. The first performance indicator of the legacy codebook scheme is used as a benchmark for comparing the actual performance of the AI / ML model scheme.

[0070] The legacy codebook is a set of pre-defined precoding matrices in the 3GPP standard. The process of calculating the performance indicator of the legacy codebook scheme is as follows: the UE obtains the channel information, i.e., the channel matrix H, by measuring the downlink CSI-RS. Further, the UE will traverse all candidate precoding matrices in the legacy codebook set. For each candidate precoding matrix W, the UE will calculate the corresponding performance metric in combination with the channel matrix H. The performance metric is usually the effective SINR after precoding, and the goal is to find the precoding matrix W_legacy that maximizes the effective SINR. The UE uses the selected optimal precoding matrix W_legacy and the channel matrix H to calculate the effective SINR, which represents the performance of the legacy codebook scheme.

[0071] In addition, the UE can further obtain the hypothetical BLER based on the calculated effective SINR. The UE can be configured by the network to predefine a mapping table (or curve) that describes the mapping relationship between the effective SINR and the hypothetical BLER under a specific modulation and coding scheme. The UE obtains the BLER value corresponding to the previously obtained effective SINR by querying the SINR-BLER mapping table. This BLER is not obtained by actual data transmission and decoding failure statistics, but is theoretically calculated based on the channel quality, so it is also called hypothetical BLER. This parameter represents the expected performance of the legacy codebook scheme under the channel condition.

[0072] Meanwhile, the UE can calculate a second performance indicator of the AI / ML model scheme, such as the equivalent signal-to-interference-and-noise ratio or the hypothetical block error rate, using the second precoded CSI-RS. By comparing the second performance indicator with the first performance indicator, the UE can determine whether the AI / ML model-based scheme is superior to the conventional codebook-based scheme. The UE can obtain the precoding matrix output by the gNB based on the second precoded CSI-RS and the original channel matrix H measured based on the standard CSI-RS received at T1, and calculate the equivalent signal-to-interference-and-noise ratio of the AI / ML model. The hypothetical block error rate of the AI / ML model is calculated in the same way as the hypothetical block error rate of the conventional codebook-based scheme, which will not be described here.

[0073] S6, the UE determines whether the second performance indicator is lower than the first performance indicator; if yes, S7 is performed; if no, the UE continues to return to perform S1-S5.

[0074] If the second performance indicator is superior to the first performance indicator, such as the equivalent signal-to-interference-and-noise ratio of the conventional codebook-based scheme is less than the equivalent signal-to-interference-and-noise ratio of the AI / ML model-based scheme, or the hypothetical block error rate of the conventional codebook-based scheme is greater than the hypothetical block error rate of the AI / ML model-based scheme, it indicates that the performance of the AI / ML model is superior to the performance of the conventional codebook-based scheme. Conversely, it indicates that the performance of the AI / ML model is inferior to the performance of the conventional codebook-based scheme.

[0075] If the performance of the AI / ML model is superior to the performance of the conventional codebook-based scheme, the UE continues to monitor the performance of the AI / ML model.

[0076] S7, the UE sends an AI-CSI performance degradation report (hereinafter referred to as a performance degradation report) to the gNB.

[0077] If the performance of the AI / ML model is inferior to the performance of the conventional codebook-based scheme, the UE triggers the performance degradation report to the gNB, which indicates that the performance of the AI / ML model is lower than that of the conventional codebook-based scheme.

[0078] This monitoring scheme triggers a report based on a single measurement result, that is, as long as the performance of the AI / ML model is lower than that of the conventional codebook-based scheme, the AI / ML model performance degradation is reported to the gNB. This performance degradation report triggering mechanism is too sensitive and is easily affected by instantaneous channel fluctuations, resulting in frequent reporting and potential mode switching, thereby increasing signaling overhead.

[0079] Although the scheme based on the precoded CSI-RS has low overhead itself, it lacks a clear and effective triggering criterion. Simply reporting low (i.e., initiating reporting when the detected performance indicator is lower than the threshold) can result in large and unstable system signaling overhead.

[0080] At the 3GPP RAN1#120 meeting, Apple proposed that for UE-side performance monitoring, a mechanism similar to Radio Link Failure (RLF) or Beam Failure Detection (BFD) can be introduced, initiated by the UE to report. Although the proposal points out the direction of "similar RLF / BFD", it does not provide a specific technical solution for how to implement the mechanism.

[0081] In addition, at the 3GPP RAN1#120 meeting, ViVo proposed two trigger logics for performance monitoring based on precoded CSI-RS:

[0082] One trigger logic is to use the probability of actual KPI being lower than the threshold as the UE-side monitoring KPI. When the probability exceeds the set threshold, two types of operations can be triggered: one operation is to trigger reporting directly, and the other operation is to trigger network-side monitoring / monitoring decision;

[0083] The other trigger logic is to use the actual KPI being lower than the threshold as the trigger condition of the UE-side monitoring event, and the report of the event can trigger the network-side monitoring and monitoring decision.

[0084] Event triggering is essentially based on single measurement instantaneous result judgment, which cannot effectively solve the false triggering problem caused by channel instantaneous fluctuation, and cannot respond to persistent performance decline.

[0085] To solve the above problems, the present application provides a performance monitoring report triggering method based on an AI / ML model. Based on the performance monitoring using precoded CSI-RS, an adaptive Cumulative Sum (CUSUM) algorithm for channel environment is designed, which can capture the performance decline trend and is not too sensitive to abnormal values, resulting in frequent reporting. Among them, the gNB can configure different CUSUM parameter sets according to the different states of the UE, and each CUSUM parameter set includes different relaxation factors k and decision thresholds h. The UE can dynamically select the CUSUM parameter set that best matches the current state according to the current state, i.e. a CUSUM algorithm for CSI performance monitoring is realized.

[0086] The performance monitoring report triggering method based on the AI / ML model provided by the present application will be described in detail below with reference to the accompanying drawings.

[0087] Please refer to Figure 5Fig. 1 shows a flowchart of a method for triggering a CSI performance monitoring report based on an AI / ML model according to an embodiment of the present application. In the embodiment, the network device is a gNB and the terminal device is a UE. In actual applications, the network device can be any network-side device capable of providing network communication functions, and the terminal device can be of various forms. The present application does not limit this. As shown in Fig. 1, the method can include the following steps: Figure 5

[0088] S101, the gNB sends a first message to the UE. The first message carries performance awareness configuration information.

[0089] For example, the first message can be an RRC reconfiguration message (RRCReconfiguration). For example, the performance awareness configuration information can be configured to the UE through a new information element (IE) in the RRC reconfiguration message.

[0090] In other embodiments, the performance awareness configuration information can also be carried by an existing IE in the RRC reconfiguration message. In this scenario, it is necessary to identify whether the IE currently carries the information originally carried by the IE or the newly added performance awareness configuration information.

[0091] The performance awareness configuration information is used to configure the parameter configuration information required by the UE for AI / ML model-based performance monitoring, such as precoding CSI-RS configuration information, traditional codebook reference configuration information, CUSUM parameter set list configuration information, and parameter mapping rule configuration information.

[0092] The configuration information of the precoding CSI-RS is used to configure the precoding CSI-RS resource used to carry the precoding matrix output by the decoder using the AI / ML model to precode the CSI-RS.

[0093] The reference configuration information of the traditional codebook is used to specify the traditional codebook configuration (such as a specific eType2 codebook subset or parameter) that the UE should refer to when calculating the performance index of the traditional codebook scheme.

[0094] The CUSUM parameter set list configuration information is used to configure a list (cusumParameterSetList) containing multiple parameter sets. Each CUSUM parameter set includes a CUSUM parameter set ID (parameterSetId), a relaxation factor k, and a decision threshold h.

[0095] The CUSUM parameter set ID is used to uniquely identify the CUSUM parameter set.

[0096] ​Relaxation factor k: used to absorb normal fluctuations in performance differences between AI / ML models and traditional codebook schemes, avoiding the accumulation of minor and short-lived performance differences.

[0097] Decision threshold h: represents the threshold for UE to decide whether to trigger a performance degradation report, and if the performance difference accumulation corresponding to the current monitoring moment is greater than the decision threshold, a performance degradation report is triggered.

[0098] The values of the different CUSUM parameter sets k and h are different, and correspond to different UE states. Here, the UE state can include state information of the environment in which the UE is located, and / or the state of the UE itself. In other words, the UE can select the CUSUM parameter set that best matches the current state according to the current state, thereby implementing a CUSUM parameter set mechanism that is self-adaptive to the channel environment. For example, in a service scenario that requires low latency and high reliability, the UE can select a parameter set with smaller values of k and h to achieve rapid and sensitive detection of minor performance degradation; while in a scenario of high-speed movement or other rapidly changing channels, the UE can select a parameter set with larger values of k and h, thereby filtering out noise caused by rapid channel fading and avoiding frequent false positives.

[0099] Configuration information of parameter set mapping rules: used to inform the UE how to select which parameter set according to the state of the UE. The configuration information includes mapping parameters and a mapping relationship table;

[0100] Mapping parameters (mappingCriteria): used to specify the basis for UE parameter mapping, such as mobility state (mobilityState), average channel quality (averageCQILevel), service type, or UE state, etc. The gNB can configure all UEs within a service range to use the same mapping parameters, or configure UEs within a service range to use different mapping parameters, which is not limited by the present application.

[0101] Mapping relationship table (mappingTable): used to provide the mapping relationship between the mapping parameters and the CUSUM parameter sets. For example, the relaxation factor k and the decision threshold h in the CUSUM parameter set used by the UE in different mobility states are different; the average channel quality of the channel in which the UE is located uses different CUSUM parameter sets; the UE uses different CUSUM parameter sets for different service types. The CUSUM parameter sets used by the UE for different states of the UE itself are also different. Of course, the mapping relationship between at least two combinations of mapping parameters and different CUSUM parameter sets can also be configured.

[0102] Optionally, the gNB can determine whether to send the first message to the UE according to the type of the UE and the performance of the UE. For example, terminals in the Internet of Things do not have the ability to deploy AI / ML models, so the gNB will not send the first message to such terminals.

[0103] S102, at each monitoring time t, the gNB sends the resources of the standard CSI-RS and the precoded CSI-RS resources.

[0104] The standard CSI-RS resource refers to the physical resources (such as time-frequency resources) of the CSI-RS required for the UE to provide channel inference; the precoded CSI-RS resource refers to the periodic or semi-persistent CSI-RS resource provided by the UE for performance monitoring.

[0105] The standard CSI-RS is a conventional reference signal sent by the base station, and the UE calculates the completed channel information, i.e., the channel matrix H, by receiving and processing the standard CSI-RS. In order for the UE to accurately measure the true characteristics of the channel (such as multipath, fading, etc.), the standard CSI-RS itself must be sent in a pre-defined manner that does not carry additional dynamic information, i.e., the standard CSI-RS is an unprecoded reference signal.

[0106] The precoded CSI-RS is a special reference signal specifically used for AI / ML model performance monitoring.

[0107] The gNB decodes the encoded and compressed CSI reported by the UE at the last monitoring time, and the output CSI after reconstruction is used as the precoding matrix to obtain the precoded CSI-RS, in other words, the gNB implicitly transmits the decoded CSI (i.e., the reconstructed precoding matrix) through the precoded CSI-RS. After receiving the precoded CSI-RS, the UE can obtain the quality of the CSI on the gNB side by analysis, thereby evaluating the performance of the entire AI / ML model.

[0108] S103, the UE calculates a first performance indicator of the codebook scheme based on the standard CSI-RS, calculates a second performance indicator of the AI / ML model scheme based on the precoded CSI-RS, and obtains a performance difference indicator y t .

[0109] At each monitoring time t, the UE obtains channel information (i.e., channel matrix H) by measuring the standard CSI-RS, and further calculates a first performance indicator of the traditional codebook scheme using the configured traditional codebook benchmark under this channel, such as equivalent SINR or hypothetical BLER, the calculation process is described in S5 of Figure 3 . This embodiment takes the equivalent SINR as an example for the first performance indicator, denoted as SINR Legacy,t .

[0110] At the same time, at each monitoring time t, the UE can obtain the output CSI (i.e., the reconstructed precoding matrix ), the second performance indicator of the AI / ML model scheme is calculated according to the reconstructed precoding matrix and the original channel matrix H, such as the equivalent SINR or the hypothetical BLER, and the calculation process is similar to that of the performance indicator of the traditional codebook scheme, which will not be described here. In this embodiment, the second performance indicator takes the equivalent SINR as an example, denoted as SINR AI,t .

[0111] The performance difference indicator of a single monitoring at monitoring time t is the difference between the performance indicator of the traditional codebook scheme and the performance indicator of the AI / ML model scheme, and its calculation formula is as follows:

[0112] (1)

[0113] y t > 0 indicates that at the monitoring time t, the performance of the traditional codebook scheme is better than that of the AI / ML model scheme. The purpose is to detect whether there is a persistent positive deviation, that is, the performance of the AI / ML model scheme is continuously inferior to that of the traditional codebook scheme.

[0114] Formula (1) takes the equivalent SINR as an example of the performance indicator. In other embodiments of the present application, the performance difference indicator can also be the difference between the hypothetical BLER of the traditional codebook scheme and the hypothetical BLER of the AI / ML model scheme, that is, .

[0115] In theory, it is expected that the performance of the AI / ML model scheme is not inferior to that of the traditional codebook scheme, that is, y t < 0, so the target deviation of the performance difference indicator can be set. If it is expected that the AI / ML model scheme has certain advantages, such as the AI / ML model scheme must be superior to the traditional codebook scheme by G dB, then can be set. The present application takes as an example for illustration. The target deviation is used as a reference value.

[0116] S104, the UE determines the target CUSUM parameter set matched with the current state.

[0117] The UE queries the state information corresponding to the mapping parameter according to the mapping parameter and the mapping relationship table configured by the gNB, and further queries the CUSUM parameter set corresponding to the state information, that is, the target CUSUM parameter set, including the relaxation factor k t and the decision threshold h t .

[0118] S105, the UE uses the target CUSUM parameter set and the performance difference indicator y t, the performance difference accumulation sum S t .

[0119] Exemplarily, at each monitoring time t, the performance difference accumulation sum S t is updated according to the following formula:

[0120] (2)

[0121] wherein S0=0, the formula indicates that the performance difference accumulation sum S t is increased only when the performance difference indicator y at the current monitoring time is greater than the target offset t and the relaxation factor k t . When y , it indicates that the performance of the AI-ML model at the current monitoring time is better than that at the previous monitoring time, i.e., the performance degradation at the previous monitoring time is a temporary degradation, and such temporary performance degradation is not accumulated but only for the case of continuous degradation, so the performance difference accumulation sum at the current monitoring time is set to 0, and the accumulation sum of the performance difference is recalculated, so as to avoid the influence of temporary performance degradation on the performance degradation judgment.

[0122] According to relevant statistical theory and practice, when the relaxation factor k is set to half of the target offset, the CUSUM chart has the highest efficiency for detecting the offset of this specific size, i.e., the average detection time is the shortest. Therefore, in the present application, the conventional value (the value of k in the conventional case) of the relaxation factor k is usually set to half of the target offset , for example, if the target offset is =3dB, then k=1.5dB.

[0123] S106, the UE compares whether the performance difference accumulation sum S t is greater than the decision threshold h t corresponding to the current monitoring time t; if yes, S107 is executed.

[0124] When S t >h t , it is determined that the triggering reporting condition is met, and the performance degradation report is reported to the gNB. If S t ≤h t , the performance difference at the next monitoring time is continuously monitored.

[0125] S107, the UE sends the performance degradation report to the gNB.

[0126] The UE can construct a dedicated media access control (MAC) layer control element (CE) to carry the performance degradation report.​

[0127] Exemplarily, the performance degradation report not only contains the trigger indication, but also includes key diagnostic information fields, such as severity level and CUSUM parameter set ID, which can enable the gNB to obtain more diagnostic information.

[0128] Trigger indication: used to inform the gNB that the performance degradation event has been triggered, which can occupy 1 bit.

[0129] Severity level: quantifies the degree of degradation of the performance of the triggered AI / ML model scheme relative to the traditional codebook scheme, which occupies 2 bits ~ 3 bits. For example, using 2 bits to encode 4 levels: 00: mild warning; 01: moderate warning; 10: severe warning; 11: reserved or other special cases.

[0130] CUSUM parameter set ID: optional field, which can occupy 1 ~ 2 bits. If the gNB configures multiple CUSUM parameter sets, the UE can report the CUSUM parameter set ID used at the time of triggering, so that the gNB knows the state of the UE.

[0131] Exemplarily, the UE sends the dedicated MAC CE constructed when obtaining the uplink transmission opportunity to the gNB, for example, the dedicated MAC CE can be transmitted to the gNB together with other data or control information, or the dedicated MAC CE can be transmitted through a specific allocated PUSCH resource.

[0132] S108, after the UE successfully sends the performance degradation report, the UE resets S t = 0.

[0133] After the performance degradation report is successfully reported, S t = 0, that is, the cumulative sum of the performance difference is re-counted.

[0134] S109, the gNB parses the performance degradation report and performs corresponding response processes according to the parsing result and response strategy.

[0135] After the gNB receives the performance degradation report, the trigger event, severity level, and optionally the CUSUM parameter set ID used at the time of triggering are parsed. According to the information obtained by parsing, the information known on the gNB side (such as the historical performance of the AI / ML model, the location of the UE, the configuration, etc.) and the response strategy, the gNB performs corresponding responses.

[0136] Exemplarily, the response strategy on the gNB side can include:

[0137] Low severity: record and observe. The accumulated data for recording is used for long-term performance analysis and model evaluation.

[0138] Medium severity: send the UE a handover instruction (e.g., downlink DCI or MAC CE) to temporarily switch to legacy codebook CSI feedback to ensure link stability. Further, consider adjusting the CUSUM parameters of this UE. If it is determined that the parameters are too sensitive (e.g., frequently triggering low-severity performance degradation reports, which in turn trigger medium-severity reports), the CUSUM parameters corresponding to this context in the mapping relationship table can be updated, such as increasing h or k.

[0139] High severity: force a switch to legacy codebook CSI feedback and trigger AI / ML model lifecycle management, such as online updating, retraining, or version rollback of AI / ML models in related areas or scenarios.

[0140] The AI / ML model-based CSI performance monitoring report triggering method provided by the embodiment provides a CUSUM algorithm that is adaptive to channel environments. The algorithm configures different CUSUM parameter sets for different states of the UE, and each CUSUM parameter set includes a relaxation factor k and a decision threshold h, and the values of k and h in different parameter sets are different. The UE can select the most suitable CUSUM parameters according to the current state (or scenario), so that the performance monitoring mechanism of the UE can intelligently balance detection sensitivity and robustness. This approach solves the inherent contradiction that fixed thresholds or counting schemes cannot meet different needs in a changing wireless environment, significantly improving the reliability and adaptability of AI / ML models. Moreover, the scheme accumulates performance differences through the CUSUM algorithm and then compares them with the decision threshold to trigger performance degradation reports. Through the accumulation and effective filtering of channel transient jitter, it ensures that reports are only triggered when the performance of the AI / ML model has a persistent decline, thereby reducing the uplink signaling overhead caused by false triggering and the resource waste caused by subsequent network responses (such as mode switching).

[0141] Moreover, the performance degradation report reported by the scheme can also carry the severity level and the CUSUM parameter set ID used when triggered. In this way, the gNB is provided with more dimensional performance degradation information with minimal additional overhead, so that the gNB can obtain the severity of the degradation and the scenario in which the UE is located when the problem occurs after receiving the performance degradation report, further enabling the response strategy of the gNB to be highly differentiated and intelligent.

[0142] The mapping parameters and corresponding mapping relationships will be introduced below. As mentioned earlier, the mapping parameters can include mobility state, average channel quality, traffic type, and UE state.

[0143] 1) Mobility state

[0144] In 3GPP communication standards, the mobility state of a UE is usually determined according to the number of cell reselections / handovers experienced by the UE within a certain time. For example, four mobility states can be defined:

[0145] Stationary: The UE has not experienced cell reselection or handover within a certain time T_cr_max (e.g., 10s), indicating that the UE is stationary. In this state, the channel state is relatively stable, and small, continuous degradation of AI / ML model performance is most sensitive. A small k value can be used to accumulate small differences, and a small h value can be used to ensure fast triggering.

[0146] Low Mobility: The UE has experienced a small number of cell reselections / handovers within a certain time (e.g., 10s), which is less than the moderate mobility threshold N_cr_M (e.g., 2 times). For example, a UE in a walking or slow-moving vehicle. In this state, the channel changes slowly and may have some fluctuations. A suitable k value can be increased to absorb normal channel fluctuation noise, and an h value can be increased to avoid false positives due to slow movement.

[0147] Medium Mobility: The UE has experienced a number of cell reselections / handovers within a certain time that is greater than or equal to the moderate mobility threshold N_cr_M and less than the high-speed mobility threshold N_cr_H (e.g., 5 times). For example, a UE in a vehicle driving on an ordinary road. In this state, the channel changes quickly, and a larger k value is needed to smooth the performance difference indicator y t , and a higher h value is used to ensure that only when there is a significant and continuous downward trend in performance is the report sent.

[0148] High Mobility: The UE has experienced a number of cell reselections / handovers within a certain time that is greater than or equal to the high-speed mobility threshold N_cr_H. For example, a UE on a highway or high-speed rail. In this state, the channel changes dramatically and decays quickly, with the largest k value and h value to provide the strongest robustness, filtering out false performance degradation due to rapid decay or frequent handover, and only focusing on the true serious failure of the AI / ML model.

[0149] For example, the gNB can configure the UE with the medium counting threshold (e.g., N_cr_M, N_cr_H) and the timing duration (T_cr_max) through RRC signaling. The gNB can adjust the judgment criteria of the mobility state by updating the technical threshold and timing duration flexibly according to the actual scene (e.g., dense urban center, open highway, etc.).

[0150] In one example, the time duration T_cr_max = 10s, N_cr_M = 2, N_cr_H = 5, and the UE determines its mobility state in the following cases:

[0151] If the number of cell switching within 10s is 0, the UE is determined to be in the stationary state; if the number of cell switching within 10s is 1, the UE is determined to be in the low mobility state; if the number of cell switching within 10s is 2-4, the UE is determined to be in the medium mobility state; if the number of cell switching within 10s is greater than or equal to 5, the UE is determined to be in the high mobility state.

[0152] The mapping relationship between the mobility state of the UE and the CUSUM parameter set is shown in the following table:

[0153] Table 1

[0154]

[0155] 2) Average channel quality

[0156] The UE can calculate the average channel quality indicator (CQI) in a period of time as a measure of channel quality. For example, the following three average CQI levels can be defined:

[0157] High: 12≤average CQI≤15, indicating good channel quality. The performance fluctuation in this state is mainly due to the AIML model itself, and a high sensitivity configuration, i.e., small k value and small h value, should be used to quickly detect model problems.

[0158] Medium: 7≤average CQI≤11, indicating that the channel quality has certain fluctuations. This state is a common scenario, and a medium k value and h value can be used to balance the detection sensitivity and false alarm rate.

[0159] Low: 1≤average CQI≤6, the channel itself is of poor quality, and the performance jitter is large. In this state, a robust parameter configuration should be used, such as a large k value and h value, to avoid misjudging channel degradation as AI / ML model performance degradation and reduce false alarms.

[0160] The mapping relationship between the average CQI level of the UE and the CUSUM parameter set is shown in Table 2:

[0161] Table 2

[0162]

[0163] 3) Service type

[0164] The gNB can learn the current service type and its quality of service (QoS) requirement served for the UE, and can configure the corresponding CUSUM parameter set according to the quality of service requirement corresponding to different service types, for example, for the service with extremely high reliability and latency requirement (such as ultra-reliability and ultra low-latency communication (URLLC) service), a sensitive parameter configuration (such as small k value and h value) should be adopted to realize real-time performance degradation awareness and immediate reporting. For the service that pays attention to long-term throughput (such as the service in the enhanced mobile broadband (eMBB) scenario), there is a certain tolerance to temporary performance fluctuations, and a standard parameter configuration (such as a medium k value and h value) can be adopted. For the service with small data packet and insensitive to latency (such as the service in the massive machine type communication (mMTC) scenario), in order to avoid excessive performance report storm caused by massive UEs, a non-sensitive parameter configuration (such as a large k value and h value) should be adopted to report only serious and continuous performance degradation.

[0165] The mapping relationship table corresponding to the service type of the UE is shown in Table 3:

[0166] Table 3

[0167]

[0168] The above Tables 1-3 are only examples, and the relaxation factor k and the decision threshold h in the CUSUM parameter set can be configured according to actual application, and the present application does not specially limit the k value and the h value in the CUSUM parameter set.

[0169] 4) UE's own state

[0170] The UE's own state covers all possible internal states of the UE itself that can affect the performance of the AI / ML model or the UE reporting strategy, so that the performance monitoring mechanism is more intelligent and more suitable for actual scenarios. Some possible UE's own states can include:

[0171] 1) Power / Battery State

[0172] The remaining power of the UE or the current power consumption mode. When the UE power is too low, the UE can enter the power saving mode and actively reduce the performance of its processor, which can affect the real-time performance and accuracy of the AI / ML model inference.

[0173] For example, in the case of sufficient UE battery level, the performance monitoring can be performed with the regular parameter set. When the UE battery level is below 20%, a less sensitive CUSUM parameter set (e.g., larger relaxation factor k and decision threshold h) can be selected to reduce the false positive triggered by performance fluctuation (possibly caused by power consumption optimization), thus saving the power consumption for reporting.

[0174] ②Thermal State

[0175] The UE can generate a large amount of heat when performing complex calculations (AI / ML model inference). When the temperature is too high, the UE will trigger thermal protection and force the processor to reduce the processing frequency (i.e., frequency reduction), which will directly affect the performance of the AI / ML model.

[0176] For example, the UE can use the regular parameter set when it is running in the normal temperature range. If the UE detects that the processor temperature is too high and enters the frequency reduction mode, the UE can select a parameter set with higher tolerance, i.e., larger relaxation factor k and decision threshold h, to avoid misjudging temporary performance degradation caused by frequency reduction as continuous degradation of the AI / ML model performance, thus avoiding unnecessary fallback, i.e., fallback to the traditional codebook scheme.

[0177] ③Beam Management State

[0178] In 5G (especially in millimeter wave) systems, beam stability is the key to ensuring communication quality. The beam management state of the UE directly reflects the stability of the connection, such as whether frequent beam switching occurs or whether beam failure (BFD) has been experienced.

[0179] For example, in the case of stable beam, the UE is locked on one or a few stable and strong beams, and the channel is relatively stable at this time. A more sensitive parameter set can be used to quickly detect the subtle performance degradation of the AI / ML model.

[0180] In the case of beam switching / beam instability, the UE is experiencing frequent beam switching or has just recovered from beam failure. At this time, the channel changes dramatically, and a more robust (less sensitive) parameter set (e.g., larger relaxation factor k and decision threshold h) should be selected to filter out the dramatic performance jitter caused by beam changes.

[0181] ④Capability / Processing Load State

[0182] This state includes the capability level of the UE's own AI processing unit (such as a neural network processing unit (NPU)) and the current load of the processor.

[0183] For example, in a low-load state, the UE is not currently running other applications that consume a large amount of computing resources (such as large games, video rendering), the NPU resources are sufficient, the AI / ML model performance is stable, and a regular parameter set can be used.

[0184] In a high-load state, the UE is running multiple high-load applications, and the computing resources allocated to CSI processing can be preempted, causing the AI / ML model performance to fluctuate. At this time, the UE can switch to a parameter set with higher tolerance (such as a larger relaxation factor k and decision threshold h), avoiding the performance fluctuations caused by resource competition from triggering the report.

[0185] In summary, the UE's own state is a multi-dimensional concept that enables the UE to intelligently diagnose itself before reporting performance degradation to the network side, and to determine whether the performance degradation is caused by the AI / ML model or by special circumstances caused by the UE's own state. In this way, the accuracy of the report can be greatly improved, the signaling overhead can be reduced, and more refined network operation and maintenance can be achieved.

[0186] In addition, the mapping relationship between multiple state combinations and corresponding CUSUM parameter sets can also be configured, for example, the gNB can configure the mapping relationship between the mobility state of the UE and the UE's own state and the CUSUM parameter set, and for another example, the gNB can also configure the mapping relationship between the combination of the UE's service type and the average channel quality and the CUSUM parameter set, which will not be described in detail here.

[0187] Figure 6 is a schematic block diagram of a communication device provided by an embodiment of the present application.

[0188] As shown in Figure 6 , the communication device can include a processing module 201 and a communication module 202. The processing module 201 can implement corresponding processing functions. Alternatively, the processing module 201 can also be referred to as a processing unit. The communication module 202 can implement corresponding communication functions. The communication functions can be internal communication functions of the communication device, or communication functions of the communication device and other devices. Alternatively, the communication module 202 can also be referred to as a communication interface or a transceiver module, a transceiver unit.

[0189] Alternatively, the communication device further includes a storage module, which can be used to store instructions and / or data; the processing module 201 can read the instructions and / or data in the storage module, so that the communication device implements the foregoing method embodiments.

[0190] In a possible design, the communication apparatus can correspond to a terminal device in the method embodiments above, or a component (such as a circuit, a chip, or a chip system, etc.) configured in the terminal device. The communication apparatus can be used to perform the steps or procedures performed by the terminal device in any of the method embodiments above.

[0191] In a possible implementation, the processing module 201 is configured to obtain a performance difference indicator between a traditional codebook scheme and an AI / ML model corresponding to a current monitoring moment; and determine a target CUSUM parameter set based on a current state of the terminal device and CUSUM parameter set configuration information, where different states of the terminal device correspond to different CUSUM parameter sets, each CUSUM parameter set includes a relaxation factor and a decision threshold, and the relaxation factor and the decision threshold in different CUSUM parameter sets are different; further, update a performance difference cumulative sum at the current monitoring moment based on the performance difference indicator and the relaxation factor in the target CUSUM parameter set, where the performance difference cumulative sum at the current monitoring moment is the part of the performance difference indicator at the current monitoring moment that exceeds a preset cumulative threshold added to the performance difference cumulative sum at a previous monitoring moment, and the preset cumulative threshold includes the relaxation factor in the target CUSUM parameter set.

[0192] The communication module 202 is configured to send a performance degradation report to the network device in a case where the processing module 201 determines that the performance difference cumulative sum is greater than the decision threshold in the target CUSUM parameter set, where the performance degradation report is used to indicate that the performance of the AI / ML model is lower than that of the traditional codebook scheme.

[0193] In another possible implementation, the performance difference cumulative sum corresponding to the current monitoring moment is updated according to the following formula:

[0194]

[0195] where S t is the performance difference cumulative sum corresponding to the current monitoring moment t, S t-1 is the performance difference cumulative sum corresponding to the previous monitoring moment t-1, y t is the performance difference indicator corresponding to the current monitoring moment t, k t is the relaxation factor used at the current monitoring moment, and k t is the target offset. t t-1 t t

[0196] In another possible implementation, the processing module 201 is further configured to set the performance difference cumulative sum corresponding to the current monitoring moment to 0 after sending the performance degradation report to the network device.

[0197] ​​​​​In a possible implementation, the communication module 202 is further configured to receive a first message from the network device, the first message comprising performance awareness configuration information, the performance awareness configuration information comprising CUSUM parameter set list information and parameter set mapping rules, the CUSUM parameter set list information comprising a CUSUM parameter set, and the parameter set mapping rules comprising a mapping relationship between different state parameters of the terminal device and the CUSUM parameter set.

[0198] In a possible implementation, the first message is an RRC reconfiguration message.

[0199] In a possible implementation, the performance awareness configuration information is carried in a newly added information element in the RRC reconfiguration message.

[0200] In a possible implementation, the state parameters of the terminal device comprise at least one of a mobility state, a channel quality, a service type, and a self state of the terminal device, and the self state of the terminal device comprises at least one of a remaining power, a running temperature, a beam management state, and a hardware capability.

[0201] In a possible implementation, the mapping relationship between the mobility state and the CUSUM parameter set comprises:

[0202] a first CUSUM parameter set corresponding to a stationary state, the first CUSUM parameter set comprising a first relaxation factor and a first decision threshold;

[0203] a second CUSUM parameter set corresponding to a low mobility state, the second CUSUM parameter set comprising a second relaxation factor and a second decision threshold;

[0204] a third CUSUM parameter set corresponding to a medium mobility state, the third CUSUM parameter set comprising a third relaxation factor and a third decision threshold;

[0205] a fourth CUSUM parameter set corresponding to a high mobility state, the fourth CUSUM parameter set comprising a fourth relaxation factor and a fourth decision threshold;

[0206] the first relaxation factor, the second relaxation factor, the third relaxation factor, and the fourth relaxation factor increase in value in turn, and the first decision threshold, the second decision threshold, the third decision threshold, and the fourth decision threshold increase in value in turn.

[0207] In a possible implementation, the mapping relationship between the channel quality and the parameter set comprises:

[0208] a fifth CUSUM parameter set corresponding to a good channel quality, the fifth CUSUM parameter set comprising a fifth relaxation factor and a fifth decision threshold;

[0209] The medium channel quality corresponds to a sixth CUSUM parameter set, and the sixth CUSUM parameter set includes a sixth relaxation factor and a sixth decision threshold;

[0210] The poor channel quality corresponds to a seventh CUSUM parameter set, and the seventh CUSUM parameter set includes a seventh relaxation factor and a seventh decision threshold;

[0211] The numerical values of the fifth relaxation factor, the sixth relaxation factor, and the seventh relaxation factor increase in turn; and the numerical values of the fifth decision threshold, the sixth decision threshold, and the seventh decision threshold increase in turn.

[0212] In another possible implementation, the mapping relationship between the service type of the terminal device and the parameter set includes:

[0213] The URLLC type corresponds to a ninth CUSUM parameter set, and the ninth CUSUM parameter set includes a ninth relaxation factor and a ninth decision threshold;

[0214] The eMBB service corresponds to a tenth CUSUM parameter set, and the tenth CUSUM parameter set includes a tenth relaxation factor and a tenth decision threshold;

[0215] The mMTC service corresponds to an eleventh CUSUM parameter set, and the eleventh CUSUM parameter set includes an eleventh relaxation factor and an eleventh decision threshold;

[0216] The numerical values of the ninth relaxation factor, the tenth relaxation factor, and the eleventh relaxation factor increase in turn, and the numerical values of the ninth decision threshold, the tenth decision threshold, and the eleventh decision threshold increase in turn.

[0217] In another possible implementation, the processing module 201 is specifically configured to: determine a current state of the terminal device; determine a target CUSUM parameter set ID matched with the current state based on a parameter set mapping rule; and read a CUSUM parameter set corresponding to the target CUSUM parameter set ID.

[0218] In another possible implementation, the performance degradation report includes AI / ML model performance degradation indication and diagnosis information, and the diagnosis information includes performance degradation information when the terminal device triggers the performance degradation report.

[0219] In another possible implementation, the diagnosis information includes at least one of a severity level and a CUSUM parameter set ID used when triggering the performance degradation report, and the severity level indicates a degradation degree of the performance of the AI / ML model relative to a traditional codebook scheme.

[0220] In another possible implementation, the communication module 202 is specifically configured to carry the performance degradation report and send the performance degradation report through a dedicated MAC CE.

[0221] In another possible implementation, the communication module 202 is specifically configured to transmit the performance degradation report through an uplink shared channel resource or an uplink control channel resource.

[0222] The above is only an example, and detailed steps or processes can refer to the descriptions of the foregoing embodiments.

[0223] In a possible design, the communication apparatus can correspond to a network device (which can be a RAN, or a core network device or a functional unit in a core network device) in the foregoing method embodiments, or a component (such as a circuit, a chip or a chip system, etc.) configured in the network device. The communication apparatus can be configured to perform the steps or processes performed by the network device in any of the foregoing method embodiments.

[0224] In a possible implementation, the communication module 202 is configured to: receive a performance degradation report from a terminal device, the performance degradation report being sent when the terminal device monitors that a performance difference cumulative sum at a current monitoring moment is greater than a decision threshold in a target CUSUM parameter set; wherein the target CUSUM parameter set is determined by the terminal device according to a current state of the terminal device, different states of the terminal device correspond to different CUSUM parameter sets, each CUSUM parameter set includes a relaxation factor and a decision threshold, and the relaxation factor and the decision threshold in different CUSUM parameter sets are different; the performance difference cumulative sum at the current monitoring moment is updated by the terminal device based on a performance difference index at the current monitoring moment and the relaxation factor in the target CUSUM parameter set, and the performance difference cumulative sum at the current monitoring moment is the part of the performance difference index at the current monitoring moment that exceeds a preset cumulative threshold added to the performance difference cumulative sum at a previous monitoring moment; and the preset cumulative threshold includes the relaxation factor in the target CUSUM parameter set.

[0225] In another possible implementation, the communication module 202 is further configured to: send a first message, the first message including performance awareness configuration information, the performance awareness configuration information including CUSUM parameter set list information and parameter set mapping rules, the CUSUM parameter set list information including the CUSUM parameter set, and the parameter set mapping rules including a mapping relationship between different state parameters of the terminal device and the CUSUM parameter set.

[0226] In another possible implementation, the first message is an RRC reconfiguration message.

[0227] In another possible implementation, a newly added information element in the RRC reconfiguration message carries the performance awareness configuration information.

[0228] The above is only an example, and detailed steps or processes can refer to the descriptions of the foregoing embodiments.

[0229] Figure 7 is another schematic block diagram of a communication apparatus provided in embodiments of the present application.

[0230] The communication apparatus can be a terminal device, a network device, a chip, a chip system, a processor, or the like, which implements the above method. The communication apparatus can be used to implement the method described in the above method embodiments, and details can be referred to the description in the above method embodiments.

[0231] As shown in Figure 7 , the communication apparatus can include one or more processors 301, which can also be referred to as processing units or processing modules, and can implement certain control functions. The processor 301 can be a general-purpose processor or a special-purpose processor, for example, a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, and the central processing unit can be used to control the communication apparatus (e.g., a base station, a baseband chip, a user, and a user chip), execute software programs, and process data of the software programs.

[0232] In an optional design, the processor 301 can also store instructions and / or data, which can be executed by the processor 301, so that the communication apparatus performs the method described in the above method embodiments.

[0233] In another optional design, the communication apparatus can include a communication interface 302 for implementing receiving and sending functions. For example, the communication interface 302 can be a transceiver circuit, an interface, an interface circuit, or a transceiver, etc. The transceiver circuit, the interface, the interface circuit, or the transceiver for implementing the receiving and sending functions can be separate or integrated together. The above transceiver circuit, interface, interface circuit, or transceiver can be used for reading and writing of codes / data, or the above transceiver circuit, interface, interface circuit, or transceiver can be used for transmission or transfer of signals.

[0234] Optionally, the communication apparatus can include one or more memories 303, which can store instructions that can be executed on the processor 301, so that the communication apparatus performs the method described in the above method embodiments. Optionally, the memory 303 can also store data. Optionally, the processor 301 can also store instructions and / or data. The processor 301 and the memory 303 can be separately arranged or integrated together.

[0235] It should be understood that, in a possible design, each step in the method embodiments provided in the present application can be completed by integrated logic circuits of hardware in a processor or instructions in the form of software. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as execution completed by a hardware processor, or execution completed by a combination of hardware and software modules in the processor. The software modules can be located in storage media which are mature in the art, such as random storage, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, register, etc. The storage media are located in the storage, and the processor reads information in the storage, and combines hardware to complete the steps of the above method. To avoid repetition, no longer detailed description is made here.

[0236] In an implementation, the communication apparatus can correspond to the UE in the above method embodiments, and can be used to execute each step and / or procedure executed by the UE in the above method embodiments. The processor 301 can be used to execute instructions stored in the memory 303, and when the processor 301 executes the instructions stored in the memory, the processor 301 is used to execute each step and / or procedure of the above method embodiments corresponding to the terminal device.

[0237] In another implementation, the communication apparatus can correspond to the network device in the above method embodiments, and can be used to execute each step and / or procedure executed by the network device in the above method embodiments. The processor 301 can be used to execute instructions stored in the memory 303, and when the processor 301 executes the instructions stored in the memory, the processor 301 is used to execute each step and / or procedure of the above method embodiments corresponding to the network device.

[0238] It can be understood that the above processor can be one or more chips. For example, the processor can be a field programmable gate array (FPGA), can be an application specific integrated circuit (ASIC), can be a system on chip (SoC), can be a central processor unit (CPU), can be a network processor (NP), can be a digital signal processing circuit (digital signal processor, DSP), can be a micro controller unit (MCU), can be a programmable logic device (programmable logic device, PLD) or other integrated chip.

[0239] It is to be appreciated that the memory in the embodiments of the application can be a volatile memory or a nonvolatile memory, or can include both volatile and nonvolatile memory. Where the nonvolatile memory is, for example, read-only memory (ROM), programmable ROM (PROM), erasable PROM (EPROM), electrically EPROM (EEPROM), or flash memory. The volatile memory, which can be used as external cache, can be, for example, random access memory (RAM), static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct Rambus RAM (DR RAM). It is to be appreciated that the memory described herein is intended to include, among other things, these and any other suitable types of memory.

[0240] The embodiments of the application further provide a computer readable storage medium, which stores instructions, and when the instructions are run on one or more computing devices, the one or more computing devices perform the data transmission method described in the above embodiments.

[0241] The computer readable storage medium can be a non-transitory computer readable storage medium, for example, the non-transitory computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0242] The embodiment of the present application further provides a computer program product. When executed by one or more computing devices, the one or more computing devices execute any of the aforementioned data transmission methods. The computer program product can be a software package. When any of the aforementioned data transmission methods is needed, the computer program product can be downloaded and executed on a computer.

[0243] The embodiment of the present application further provides a processor. The processor comprises an input circuit, an output circuit and a processing circuit. The processing circuit is configured to receive a signal through the input circuit and transmit the signal through the output circuit, so that the processor executes the data transmission method described in the above embodiment.

[0244] In the implementation process, the processor can be one or more chips, the input circuit can be an input pin, the output circuit can be an output pin, and the processing circuit can be a transistor, a gate circuit, a flip-flop, various logic circuits, etc. The input signal received by the input circuit can be received and input by, for example but not limited to, a receiver, the signal output by the output circuit can be output to and transmitted by, for example but not limited to, a transmitter, and the input circuit and the output circuit can be the same circuit which is used as the input circuit and the output circuit at different times. The embodiment of the present application does not limit the specific implementation of the processor and various circuits.

[0245] The embodiment of the present application further provides a chip system. The chip system comprises one or more processors configured to call and run instructions stored in a memory, so that the data transmission method described in the above embodiment is executed. The chip system can be composed of a chip, or can comprise a chip and other discrete devices. The chip system can comprise an input circuit or an interface for transmitting information or data, and an output circuit or an interface for receiving information or data.

[0246] In the embodiment of the present application, each term and English abbreviation is an exemplary example given for convenience of description, and should not constitute any limitation on the present application. The present application does not exclude the possibility of defining other terms capable of achieving the same or similar functions in existing or future protocols.

[0247] In the above embodiment, all or part of the embodiment can be realized by software, hardware, firmware or any combination thereof. When realized by software, all or part of the embodiment can be realized in the form of a computer program product. The computer program product comprises one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiment of the present application are generated.

[0248] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other manners. For example, the division of the above-described device embodiment is only a logical function division, and there can be another division manner for actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units, or the among different units, can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0249] It should be understood that, in various embodiments of the present application, the sequence of the processes does not mean the execution sequence, and the execution sequence of the processes should be determined according to the functions and the inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0250] In summary, the above description is only the preferred embodiment of the technical scheme of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An AI / ML model based CSI performance monitoring report triggering method, characterized in that, The method is performed by a terminal device, and the method comprises: obtaining a performance difference index between a traditional codebook scheme and an AI / ML model corresponding to a current monitoring moment; determining a target CUSUM parameter set based on a current state of the terminal device and CUSUM parameter set configuration information, different states of the terminal device corresponding to different CUSUM parameter sets, each CUSUM parameter set comprising a relaxation factor and a decision threshold, the relaxation factor and the decision threshold in different CUSUM parameter sets being different; updating a performance difference accumulation and of the current monitoring moment, the performance difference accumulation and of the current monitoring moment being a part of the performance difference index of the current monitoring moment exceeding a preset accumulation threshold based on a performance difference accumulation and of a previous monitoring moment, the preset accumulation threshold comprising a sum of a target offset and the relaxation factor in the target CUSUM parameter set, the target offset being a benchmark value for measuring the performance difference index; if the performance difference accumulation and is greater than the decision threshold in the target CUSUM parameter set, sending a performance degradation report to a network device, the performance degradation report being used to indicate that the performance of the AI / ML model is lower than that of the traditional codebook scheme.

2. The method of claim 1, wherein, The method further comprises: if a sum of the performance difference accumulation and of the previous monitoring moment and a first difference value is less than or equal to 0, setting the performance difference accumulation and of the current monitoring moment to 0, the first difference value being a difference value between the performance difference index of the current monitoring moment and the preset accumulation threshold.

3. The method according to claim 1 or 2, characterized in that, The updating of the performance difference accumulation and of the current monitoring moment based on the performance difference index and the relaxation factor in the target CUSUM parameter set comprises: updating the performance difference accumulation and corresponding to the current monitoring moment according to the following formula: wherein S t is the performance difference cumulative sum corresponding to the current monitoring moment t, S t-1 is the performance difference cumulative sum corresponding to the previous monitoring moment t-1, y t is the performance difference index corresponding to the current monitoring moment t, k t is the relaxation factor used at the current monitoring moment, is the target offset.

4. The method according to claim 1 or 2, characterized in that, The obtaining of the performance difference index between the traditional codebook scheme and the AI / ML model corresponding to the current monitoring moment comprises: obtaining a first performance index of the traditional codebook scheme corresponding to the current monitoring moment, the first performance index being calculated based on a standard CSI-RS received at the current monitoring moment; obtaining a second performance index of the AI / ML model corresponding to the current monitoring moment, the second performance index being calculated based on a precoded CSI-RS received at the current monitoring moment, the first performance index and the second performance index both being an equivalent signal-to-interference-and-noise ratio or a hypothetical block error rate; calculating a difference value between the first performance index and the second performance index to obtain the performance difference index.

5. The method according to claim 1 or 2, characterized in that, The method further comprises: after sending the performance degradation report to the network device, setting the performance difference accumulation and corresponding to the current monitoring moment to 0.

6. The method of claim 1 or 2, wherein, Before the obtaining of the performance difference index between the traditional codebook scheme and the AI / ML model corresponding to the current monitoring moment, the method further comprises: receiving a first message from a network device, the first message comprising performance awareness configuration information, the performance awareness configuration information comprising CUSUM parameter set configuration information and parameter set mapping rule, the CUSUM parameter set configuration information comprising a plurality of CUSUM parameter sets, the parameter set mapping rule comprising a mapping relationship between different state parameters of the terminal device and the CUSUM parameter sets.

7. The method of claim 6, wherein, The first message is an RRC reconfiguration message.

8. The method of claim 7, wherein, The newly added information element in the RRC reconfiguration message is used to carry the performance awareness configuration information.

9. The method of claim 6, wherein, The state parameters of the terminal device comprise at least one of mobility state, channel quality, service type and self-state of the terminal device, and the self-state of the terminal device comprises at least one of remaining power, running temperature, beam management state and hardware capability.

10. The method of claim 9, wherein, The mapping relationship between the mobility state and the CUSUM parameter set comprises: a first CUSUM parameter set corresponding to a stationary state, the first CUSUM parameter set comprising a first relaxation factor and a first decision threshold; a second CUSUM parameter set corresponding to a low mobility state, the second CUSUM parameter set comprising a second relaxation factor and a second decision threshold; a third CUSUM parameter set corresponding to a medium mobility state, the third CUSUM parameter set comprising a third relaxation factor and a third decision threshold; a fourth CUSUM parameter set corresponding to a high mobility state, the fourth CUSUM parameter set comprising a fourth relaxation factor and a fourth decision threshold; The values of the first relaxation factor, the second relaxation factor, the third relaxation factor and the fourth relaxation factor increase in turn, and the values of the first decision threshold, the second decision threshold, the third decision threshold and the fourth decision threshold increase in turn.

11. The method of claim 9, wherein, The mapping relationship between the channel quality and the parameter set comprises: a fifth CUSUM parameter set corresponding to excellent channel quality, the fifth CUSUM parameter set comprising a fifth relaxation factor and a fifth decision threshold; a sixth CUSUM parameter set corresponding to medium channel quality, the sixth CUSUM parameter set comprising a sixth relaxation factor and a sixth decision threshold; a seventh CUSUM parameter set corresponding to poor channel quality, the seventh CUSUM parameter set comprising a seventh relaxation factor and a seventh decision threshold; The values of the fifth relaxation factor, the sixth relaxation factor and the seventh relaxation factor increase in turn, and the values of the fifth decision threshold, the sixth decision threshold and the seventh decision threshold increase in turn.

12. The method of claim 9, wherein, The mapping relationship between the service type of the terminal device and the parameter set comprises: a ninth CUSUM parameter set corresponding to ultra-reliable low-latency communication type, the ninth CUSUM parameter set comprising a ninth relaxation factor and a ninth decision threshold; a tenth CUSUM parameter set corresponding to enhanced mobile broadband service, the tenth CUSUM parameter set comprising a tenth relaxation factor and a tenth decision threshold; an eleventh CUSUM parameter set corresponding to massive machine type communication service, the eleventh CUSUM parameter set comprising an eleventh relaxation factor and an eleventh decision threshold; The values of the ninth relaxation factor, the tenth relaxation factor and the eleventh relaxation factor increase in turn, and the values of the ninth decision threshold, the tenth decision threshold and the eleventh decision threshold increase in turn.

13. The method of claim 1 or 2, wherein, The determining the target CUSUM parameter set based on the current state of the terminal device and the CUSUM parameter set configuration information comprises: determining the current state of the terminal device; determining a target CUSUM parameter set ID matched with the current state based on a parameter set mapping rule; reading a CUSUM parameter set corresponding to the target CUSUM parameter set ID.

14. The method of claim 1 or 2, wherein, The performance degradation report comprises an AI / ML model performance degradation indication and diagnostic information, and the diagnostic information comprises performance degradation information when the terminal device triggers the performance degradation report.

15. The method of claim 14, wherein, The diagnostic information comprises at least one of a severity level and a CUSUM parameter set ID used when triggering the performance degradation report, and the severity level indicates a degree of degradation of AI / ML model performance relative to a traditional codebook scheme.

16. The method of claim 1 or 2, wherein, The sending the performance degradation report to the network device comprises: carrying the performance degradation report through a dedicated MAC CE and sending.

17. The method of claim 1 or 2, wherein, The sending the performance degradation report to the network device comprises: transmitting the performance degradation report through uplink shared channel resources or uplink control channel resources. 18.A method for triggering an AI / ML model based CSI performance monitoring report, the method comprising: The method is performed by a network device, and the method comprises: receiving a performance degradation report from a terminal device, the performance degradation report being sent when the terminal device monitors that a performance difference cumulative sum at a current monitoring time is greater than a decision threshold in a target CUSUM parameter set; wherein the target CUSUM parameter set is determined by the terminal device according to a current state of the terminal device, different states of the terminal device correspond to different CUSUM parameter sets, each CUSUM parameter set comprises a relaxation factor and a decision threshold, and the relaxation factor and the decision threshold in different CUSUM parameter sets are different; the performance difference cumulative sum at the current monitoring time is updated by the terminal device based on a performance difference indicator at the current monitoring time and the relaxation factor in the target CUSUM parameter set, and the performance difference cumulative sum at the current monitoring time is obtained by adding, to a performance difference cumulative sum at a previous monitoring time, a part of the performance difference indicator at the current monitoring time that exceeds a preset cumulative threshold, and the preset cumulative threshold comprises a sum of a target offset and the relaxation factor in the target CUSUM parameter set, and the target offset is a benchmark value for measuring the performance difference indicator.

19. The method of claim 18, wherein, Before the receiving the performance degradation report from the terminal device, the method further comprises: sending a first message, the first message comprising performance awareness configuration information, the performance awareness configuration information comprising CUSUM parameter set list information and a parameter set mapping rule, the CUSUM parameter set list information comprising the CUSUM parameter sets, and the parameter set mapping rule comprising a mapping relationship between different state parameters of the terminal device and the CUSUM parameter sets.

20. The method of claim 19, wherein, The first message is an RRC reconfiguration message.

21. The method of claim 20, wherein, The newly-added information element in the RRC reconfiguration message carries the performance-aware configuration information.

22. A communications device, characterized by comprising means for performing the method of any one of claims 1 to 17, or means for performing the method of any one of claims 18 to 21.

23. A communications device, characterized by comprising: a memory for storing computer programs or computer instructions; a processor for executing the computer programs or computer instructions stored in the memory, so that the communication device performs the method of any one of claims 1 to 17, or the method of any one of claims 18 to 21.

Images