WO2024140422A1

WO2024140422A1 - Ai unit performance monitoring method, terminal and network side device

Info

Publication number: WO2024140422A1
Application number: PCT/CN2023/140773
Authority: WO
Inventors: 任千尧
Original assignee: 维沃移动通信有限公司
Priority date: 2022-12-26
Filing date: 2023-12-21
Publication date: 2024-07-04
Also published as: CN118265043A

Abstract

The present application relates to the technical field of communications, and discloses an AI unit performance monitoring method, a terminal and a network side device. The AI unit performance monitoring method in embodiments of the present application comprises: a terminal determines monitoring information on the basis of first channel information; and the terminal sends the monitoring information to the network side device; wherein an AI unit comprises a first unit arranged on the terminal and a second unit arranged on the network side device; the first channel information is channel information input to the first unit; and the monitoring information comprises at least one of the following items: first measurement information used for indicating a first degree of association of the first channel information and a first reference base; an identifier (ID) of the first measurement information and an ID of the first reference base; and a performance monitoring result used for indicating whether the AI unit is valid or not.

Description

Performance monitoring method, terminal and network side equipment of AI unit

This application claims priority to the Chinese patent application filed with the State Intellectual Property Office on December 26, 2022, with application number 202211679082.8 and application name “Performance monitoring method, terminal and network side equipment of AI unit”, all contents of which are incorporated by reference in this application.

Technical Field

The present application belongs to the field of communication technology, and specifically relates to a performance monitoring method, terminal and network-side equipment of an AI unit.

Background technique

When using an artificial intelligence (AI) model to compress channel state information (CSI), the performance of the AI model will fluctuate as the terminal position and channel environment change, and may not match the current channel state, resulting in inaccurate CSI compression-recovered channel information. Therefore, it is necessary to regularly monitor whether the performance of the AI model meets the requirements. The main evaluation indicator of AI model performance is the correlation between the input channel information and the recovered channel information. If the two are exactly the same, it means that the model has achieved perfect compression. Usually, the model is considered to be effective if the correlation loss is within a certain degree.

CSI compression based on the AI model is divided into an encoding model and a decoding model. Usually, the encoding model is on the terminal side, and the decoding model is on the base station side. In order to verify the performance of the AI model, it is necessary to compare the recovered data with the input data. However, the input channel information is on the terminal side, and the output channel information is on the base station side, so they cannot be directly compared. If the terminal side passes the input data to the base station, it requires a high overhead and it is difficult to pass accurate channel information. If the base station passes the decoding model to the terminal, on the one hand, most of the decoding models of the base station are confidential, and on the other hand, the decoding model is usually large, and passing the decoding model requires a lot of resources and has a high overhead.

Therefore, there is a problem of high overhead in AI model performance monitoring in related technologies.

Summary of the invention

The embodiments of the present application provide a performance monitoring method, terminal and network-side device for an AI unit, which can reduce the overhead of AI model performance monitoring.

In a first aspect, a performance monitoring method for an AI unit is provided, the method comprising:

The terminal determines monitoring information based on the first channel information;

The terminal sends the monitoring information to the network side device;

The AI unit includes a first unit configured at the terminal and a second unit configured at the network side device, the first channel information is channel information input to the first unit, and the monitoring information includes at least one of the following:

first measurement information, used to indicate a first correlation between the first channel information and a first reference basis;

The first measurement information and the identification ID of the first reference base;

The performance monitoring result is used to indicate whether the AI unit is effective.

In a second aspect, a performance monitoring method for an AI unit is provided, the method comprising:

The network side device receives monitoring information from the terminal;

The network side device determines whether the AI unit is valid based on the monitoring information;

The AI unit includes a first unit configured on the terminal and a second unit configured on the network side device, and the monitoring information includes at least one of the following:

first measurement information, used to indicate a first correlation between first channel information and a first reference basis, the first channel information being channel information input to the first unit;

In a third aspect, a performance monitoring device for an AI unit is provided, the device comprising:

A determination module, configured to determine monitoring information based on the first channel information;

A first sending module, used to send the monitoring information to a network side device;

In a fourth aspect, a performance monitoring device for an AI unit is provided, the device comprising:

A fourth receiving module, used for receiving monitoring information from a terminal;

a processing module, configured to determine whether the AI unit is valid based on the monitoring information;

The AI unit includes a first unit configured on the terminal and a second unit configured on a network side device, and the monitoring information includes at least one of the following:

In a fifth aspect, a terminal is provided, comprising a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the first aspect are implemented.

In a sixth aspect, a terminal is provided, comprising a processor and a communication interface; wherein the processor is configured to determine monitoring information based on the first channel information;

The communication interface is used to send the monitoring information to the network side device;

In the seventh aspect, a network side device is provided, which includes a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of the method described in the second aspect are implemented.

In an eighth aspect, a network side device is provided, comprising a processor and a communication interface; wherein the communication interface is used to receive monitoring information from a terminal;

The processor is used to determine whether the AI unit is valid based on the monitoring information;

In the ninth aspect, a performance monitoring system for an AI unit is provided, including: a terminal and a network side device, wherein the terminal can be used to execute the steps of the method described in the first aspect, and the network side device can be used to execute the steps of the method described in the second aspect.

In the tenth aspect, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method described in the second aspect are implemented.

In the eleventh aspect, a chip is provided, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the method described in the first aspect, or to implement the method described in the second aspect.

In the twelfth aspect, a computer program/program product is provided, wherein the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the steps of the method described in the first aspect, or to implement the steps of the method described in the second aspect.

In an embodiment of the present application, the terminal determines and sends monitoring information to the network side device based on the first channel information input to the first unit. The monitoring information includes first measurement information or performance monitoring results. The first measurement information indicates the first correlation between the first channel information and the first reference base, or the performance monitoring result indicates whether the AI unit is valid. After obtaining the monitoring information, the network side device can determine whether the AI unit is valid based on the monitoring information, thereby replacing the transmission of AI unit input information/output information with the transmission of measurement information of the first channel information and the first reference base, thereby reducing information transmission overhead and achieving rapid model monitoring. In addition, since there is no need to transmit the reference base in real time, only the ID of the reference base needs to be transmitted when in use. This can also reduce information transmission overhead, reduce AI model performance monitoring overhead, and achieve rapid model monitoring.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of a wireless communication system applicable to an embodiment of the present application;

FIG2 is a flow chart of a method for monitoring the performance of an AI unit according to an embodiment of the present application;

FIG3 is a second flow chart of the performance monitoring method of the AI unit provided in an embodiment of the present application;

FIG4 is a schematic diagram of a structure of a performance monitoring device for an AI unit provided in an embodiment of the present application;

FIG5 is a second structural diagram of the performance monitoring device of the AI unit provided in an embodiment of the present application;

FIG6 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application;

FIG7 is a schematic diagram of the structure of a terminal provided in an embodiment of the present application;

FIG8 is a schematic diagram of the structure of a network side device provided in an embodiment of the present application.

Detailed ways

The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of this application.

The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, but not to describe a specific order or sequence. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first" and "second" are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and/or" in the specification and claims represents at least one of the connected objects, and the character "/" generally indicates that the objects associated with each other are in an "or" relationship. The term "indication" in the specification and claims of the present application can be either an explicit indication or an implicit indication. Among them, an explicit indication can be understood as the sender explicitly notifying the receiver of the operation or request result to be performed in the indication sent; an implicit indication can be understood as the receiver making a judgment based on the indication sent by the sender and determining the operation or request result to be performed based on the judgment result.

It is worth noting that the technology described in the embodiments of the present application is not limited to the Long Term Evolution (LTE)/LTE-Advanced (LTE-A) system, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency Division Multiple Access (SC-FDMA) or other systems. The terms "system" and "network" in the embodiments of the present application are often used interchangeably, and the described technology can be used for the above-mentioned systems and radio technologies as well as other systems and radio technologies. The following description describes a new radio (NR) system for example purposes, and NR terms are used in most of the following descriptions, but these technologies can also be applied to communication systems other than NR system applications, such as the ^6th Generation (6G) communication system.

FIG1 is a schematic diagram of a wireless communication system applicable to an embodiment of the present application. The wireless communication system shown in FIG1 includes a terminal 11 and a network side device 12. The terminal 11 may be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer) or a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a palm computer, a netbook, an ultra-mobile personal computer (ultra-mobile personal computer, UMPC), Mobile Internet Device (MID), augmented reality (AR)/virtual reality (VR) equipment, robots, wearable devices (Wearable Device), vehicle-mounted equipment (VUE), pedestrian terminal (PUE), smart home (home appliances with wireless communication functions, such as refrigerators, TVs, washing machines or furniture, etc.), game consoles, personal computers (PC), ATMs or self-service machines and other terminal-side devices, wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. In addition to the above-mentioned terminal devices, the terminal 11 can also be a chip in the terminal, such as a modem chip or a system-on-chip (SoC). It should be noted that the specific type of the terminal 11 is not limited in the embodiments of the present application.

The network side device 12 may include an access network device or a core network device, wherein the access network device may also be referred to as a radio access network device, a radio access network (RAN), a radio access network function or a radio access network unit. The access network device may include a base station, a WLAN access point or a WiFi node, etc. The base station may be referred to as a node B, an evolved node B (eNB), an access point, a base transceiver station (BTS), a radio base station, a radio transceiver, a basic service set (BSS), an extended service set (ESS), a home B node, a home evolved B node, a transmitting and receiving point (TRP) or some other appropriate term in the field, as long as the same technical effect is achieved, the base station is not limited to a specific technical vocabulary, it should be noted that in the embodiment of the present application, only the base station in the NR system is used as an example for introduction, and the specific type of the base station is not limited. The core network equipment may include but is not limited to at least one of the following: core network nodes, core network functions, mobility management entity (Mobility Management Entity, MME), access mobility management function (Access and Mobility Management Function, AMF), session management function (Session Management Function, SMF), user plane function (User Plane Function, UPF), policy control function (Policy Control Function, PCF), policy and charging rules function unit (Policy and Charging Rules Function, PCRF), edge application service discovery function (Edge Application Server Discovery Function, EASDF), unified data management (Unified Data Management, UDM), unified data repository (Unified Data Repos itory, UDR), Home Subscriber Server (HSS), Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (L-NEF), Binding Support Function (BSF), Application Function (AF), Location Management Function (LMF), Enhanced Serving Mobile Location Centre (E-SMLC), Network data analytics function (NWDAF), etc. It should be noted that in the embodiments of the present application, only the core network device in the NR system is taken as an example for introduction, and the specific type of the core network device is not limited.

In order to facilitate a clearer understanding of the embodiments of the present application, some relevant background knowledge is first introduced as follows.

For multi-antenna systems, the transmitter can optimize the signal transmission based on CSI to make it more compatible with the channel status. For example, the channel quality indicator (CQI) can be used to select the appropriate modulation and coding scheme (MCS) to achieve link adaptation; the precoding matrix indicator (PMI) can be used to implement eigen beamforming to maximize the received signal. The strength of the CSI is measured, or it is used to suppress interference, such as interference between cells and interference between multiple users. Therefore, since the multi-input multi-output (MIMO) technology was proposed, how to obtain CSI has always been a research hotspot.

Usually, the base station sends a CSI Reference Signal (CSI-RS) on certain time-frequency resources in a certain time slot. The terminal performs channel estimation based on the CSI-RS, calculates the channel information on this slot, and feeds back the PMI to the base station through the codebook. The base station combines the channel information based on the codebook information fed back by the terminal, and uses this to perform data precoding and multi-user scheduling before the next CSI report.

In order to further reduce the CSI feedback overhead, the terminal can change the PMI reported for each subband to reporting the PMI according to the delay. Since the channels in the delay domain are more concentrated, the PMI of all subbands can be approximately represented by PMIs with fewer delays, that is, the delay domain information is compressed before reporting.

Similarly, in order to reduce overhead, the base station can pre-code the CSI-RS in advance and send the encoded CSI-RS to the terminal. The terminal sees the channel corresponding to the encoded CSI-RS. The terminal only needs to select several ports with higher strength from the ports indicated by the network side and report the coefficients corresponding to these ports.

Furthermore, in order to better compress channel information, neural network or machine learning methods can be used.

Artificial intelligence has been widely used in various fields. There are many ways to implement AI modules, such as neural networks, decision trees, support vector machines, Bayesian classifiers, etc. This application uses neural networks as an example for illustration, but does not limit the specific type of AI modules.

The process of CSI compression recovery based on the AI model is as follows: the terminal estimates CSI-RS, calculates channel information, obtains the encoding result through the calculated channel information or the original estimated channel information through the encoding network, and sends the encoding result to the base station. The base station receives the encoded result and inputs it into the decoding network to recover the channel information.

Specifically, the neural network-based CSI compression feedback scheme is: the terminal uses the coding network to compress and encode the channel information, sends the compressed content to the base station, and uses the decoding network at the base station to decode the compressed content to restore the channel information. At this time, the decoding network of the base station and the coding network of the terminal need to be jointly trained to achieve a reasonable match. The input of the coding model is channel information, and the output is coding information. The input of the decoding model is coding information, and the output is channel information. The main evaluation indicator of AI model performance is the correlation between the input channel information and the recovered channel information.

The performance monitoring method of the AI unit provided in the embodiment of the present application is described in detail below through some embodiments and their application scenarios in combination with the accompanying drawings.

FIG. 2 is a flow chart of a method for monitoring the performance of an AI unit according to an embodiment of the present application. As shown in FIG. 2 , the method includes steps 201-202; wherein:

Step 201: The terminal determines monitoring information based on the first channel information;

Step 202: The terminal sends the monitoring information to the network side device; wherein the AI unit includes a first unit configured at the terminal and a second unit configured at the network side device, the first channel information is channel information input to the first unit, and the monitoring information includes at least one of the following:

The first measurement information and the identification (Identity, ID) of the first reference base;

It should be noted that the embodiments of the present application can be applied to the scenario of performance evaluation and monitoring of AI units in a communication network. The terminal includes but is not limited to the types of terminal 11 listed above, and the present application does not limit this; the first network side device includes but is not limited to the types of network side device 12 listed above, and the present application does not limit this.

Optionally, the AI unit may include an AI model, an AI structure, an algorithm, a formula, a process of a specific function, and the AI unit may represent a series of operations related to AI.

Optionally, the first reference base includes at least one reference base. The number of reference bases included in the first reference base is agreed upon by the protocol or configured by the network side device. The first unit includes a first AI model, such as an encoding model; the second unit includes a second AI model, such as a decoding model. After the first channel information is input into the first unit, the first unit performs a first process, such as a compression encoding process based on an AI model. The network side device receives the compressed content, and the second unit performs a second process, such as a decoding and recovery process based on an AI model. It should be noted that the compression encoding process based on the AI model and the decoding and recovery process based on the AI model in this application are completely different concepts from channel coding and channel decoding.

Optionally, the first channel information may include: channel information estimated by the terminal; or a precoding matrix obtained by the terminal after calculating the estimated channel information.

Specifically, the first channel information may include at least one of the following:

(a) Precoding matrix of all layers;

(b) Precoding matrix (vector) of a layer;

(c) Channel matrix;

(d) Precoding matrix of all layers after preprocessing;

(e) Precoding matrix (vector) of a layer after preprocessing;

(f) Preprocessed channel matrix;

(g) The real and imaginary parts of the precoding matrix of all layers;

(h) The real and imaginary parts of a layer’s precoding matrix (vector);

(i) the real and imaginary parts of the channel matrix;

(j) The real and imaginary parts of the precoding matrix of all layers after preprocessing;

(k) The real and imaginary parts of the precoding matrix (vector) of a preprocessed layer;

(l) Real and imaginary parts of the preprocessed channel matrix.

Optionally, the first correlation between the first channel information and the first reference basis includes: a first similarity between the first channel information and the first reference basis. For example, the first similarity is the similarity between two vectors or two matrices (the first channel information and the first reference basis).

Specifically, the first similarity may include at least one of the following:

a) correlation between the first channel information and a first reference basis;

b) cosine similarity between the first channel information and a first reference basis;

c) the inner product of the first channel information and a first reference basis.

In the performance monitoring method of the AI unit provided in the embodiment of the present application, the terminal is based on the first channel input to the first unit Information, determine and send monitoring information to the network side device, including first measurement information or performance monitoring results in the monitoring information, indicating the first correlation between the first channel information and the first reference base through the first measurement information, or indicating whether the AI unit is valid through the performance monitoring result, so that after obtaining the monitoring information, the network side device can determine whether the AI unit is valid based on the monitoring information, and replace the transmission of AI unit input information/output information with the transmission of measurement information of the first channel information and the first reference base, which can reduce information transmission overhead and thus realize fast model monitoring. In addition, since there is no need to transmit the reference base in real time, only the ID of the reference base needs to be transmitted when in use, which can also reduce information transmission overhead, reduce AI model performance monitoring overhead, and realize fast model monitoring.

Optionally, the terminal receives instruction information from the network side device, where the instruction information is used to instruct the terminal to monitor the performance of the AI unit. The terminal calculates or determines monitoring information based on the instruction information of the network side device.

Specifically, the indication information may indicate at least one of the following: 1) whether to monitor the performance of the AI unit; 2) the CSI-RS position to be monitored.

Optionally, the implementation manner in which the terminal determines the monitoring information based on the first channel information may include at least one of the following:

Mode 1: The monitoring information includes the first measurement information and the ID of the first reference base. The terminal selects one or more reference bases from the first reference base set, and the terminal uses each of the reference bases as the first reference base; the terminal calculates the first degree of association between the first channel information and the first reference base to obtain the first measurement information. The terminal sends the monitoring information, i.e., the first measurement information and the ID of the first reference base, to the network side device.

For example, the terminal selects one or more reference bases from the first reference base set, and the terminal respectively calculates the correlation between each reference base and the first channel information, selects at least one reference base with the largest corresponding correlation as the first reference base, and obtains the first measurement information based on the first correlation corresponding to the first reference base.

Optionally, when the first reference basis includes a Discrete Fourier Transform (DFT) orthogonal basis, the ID of the first reference basis includes at least one of the following:

(a) beam index and/or delay index;

(b) Type II codebook parameters.

Specifically, the network side device instructs the terminal to monitor the AI model. The first channel information is the current channel information.

The terminal selects one or more reference bases in the first reference base set, calculates the measurement value of each reference base and the first channel information, and reports the ID of the corresponding reference base and the corresponding measurement value. For example, the terminal selects the reference base with the largest measurement value as the first reference base, or the reference base with the highest correlation as the first reference base.

Alternatively, the terminal selects several DFT orthogonal bases in the spatial domain and/or frequency domain, calculates coefficients of corresponding orthogonal bases or orthogonal base pairs, and reports the coefficients and the corresponding orthogonal base index.

Mode 2: The monitoring information includes the first measurement information. The network side device selects a second reference basis from the second reference basis set, where the identifier of the second reference basis is a reference basis ID; and the network side device sends the reference basis ID to the terminal.

The terminal receives the reference base ID sent by the network side device. The terminal selects a first reference base corresponding to the reference base ID in a first reference base set based on the reference base ID; the terminal calculates the first correlation between the first channel information and the first reference base to obtain the first metric information. The terminal sends a The monitoring information is the first measurement information.

Optionally, in the case where the second reference basis includes a DFT orthogonal basis, the reference basis ID includes at least one of the following: beam index and/or delay index; Type II codebook parameters.

Mode 3: The monitoring information includes the performance monitoring result. The network side device selects a second reference base from the second reference base set, and the identifier of the second reference base is a reference base ID; the network side device calculates a second correlation between the second channel information and the second reference base to obtain second measurement information; the network side device sends the reference base ID and/or the second measurement information to the terminal.

The terminal receives the reference base ID and/or the second measurement information sent by the network side device. Based on the reference base ID, the terminal selects the first reference base corresponding to the reference base ID in the first reference base set; the terminal determines the performance monitoring result based on the first channel information, the first reference base and the second measurement information. For example, the terminal calculates the first correlation between the first channel information and the first reference base to obtain the first measurement information; the terminal determines the performance monitoring result based on the first measurement information and the second measurement information. Then, the terminal sends the monitoring information, i.e., the performance monitoring result, to the network side device.

Optionally, the second measurement information is used to indicate a second degree of association between second channel information and a second reference base, the second channel information is channel information output by the second unit, and the second reference base is a reference base corresponding to the reference base ID in a second reference base set.

Method 4: The terminal selects one or more reference bases in the first reference base set as the first reference base; the terminal calculates the first correlation between the first channel information and the first reference base, and determines the performance monitoring result based on the first correlation; wherein the monitoring information includes the performance monitoring result. Optionally, the first reference base may include the characteristics of the training data of the AI unit, such as model A. The terminal determines whether the first channel information and the training data of the AI model have similar distribution characteristics or a certain correlation by comparing the first channel information with the first reference base, and thereby determines whether the first channel information is suitable for compression using the AI model based on the correlation judgment result.

Optionally, the first reference basis set may include at least one of the following:

1) Complete orthogonal basis set;

2) Incomplete orthogonal basis sets, i.e., orthogonal basis sets with fewer numbers than dimensions;

3) A set of non-orthogonal bases;

4) A set of complete or incomplete orthogonal bases, and non-orthogonal bases;

5) The orthogonal basis set formed by each column of the unit matrix;

6) During the training process of the AI unit, a set of precoding matrices or vectors determined by the training node; the training node notifies the terminal and/or network side device;

7) A collection of historical channel information, such as a collection of historical channel information of a terminal;

8) A set of feature vectors determined by historical channel information;

9) A set of feature vectors corresponding to the AI unit;

10) Feature vectors associated with the training data distribution of the AI unit.

Optionally, the first reference basis set includes a fixed reference basis set and/or a dynamic reference basis set; wherein the dynamic The reference base set is determined by the historical channel information of the terminal. The fixed reference base set includes one or more pre-agreed or configured reference bases; the fixed reference base set satisfies at least one of the following: (a) agreed by the protocol; (b) configured by the network side device; (c) configured by the network side device and matched with the AI unit.

Optionally, the length of the reference base in the dynamic reference base set is agreed upon by a protocol or configured by the network side device.

It should be noted that the terminal uses the first reference basis set, and the network side device uses the second reference basis set.

Optionally, the reference basis in the second reference basis set includes at least one of the following:

a) Complete orthogonal basis set;

b) Incomplete orthogonal basis set;

c) a set of non-orthogonal bases;

d) A set of complete or incomplete orthogonal bases, and non-orthogonal bases;

e) The orthogonal basis set formed by each column of the unit matrix;

f) a set of precoding matrices or vectors determined by the training node;

g) a collection of historical channel information;

h) a set of feature vectors determined by historical channel information;

i) a set of feature vectors corresponding to the AI unit;

j) A feature vector associated with the training data distribution of the AI unit.

Optionally, the second reference basis set includes a fixed reference basis set and/or a dynamic reference basis set; wherein the dynamic reference basis set is determined by historical channel information on the network side device side. The fixed reference basis set includes one or more pre-agreed or configured reference bases; the fixed reference basis set satisfies at least one of the following: (a) agreed upon by the protocol; (b) configured by the network side device, that is, both the network side device and the terminal are known; (c) configured by the network side device and matched with the AI unit. Optionally, in the fixed reference basis set, each reference basis can be mutually orthogonal.

It should be noted that the first reference basis set used by the terminal and the second reference basis set used by the network side device may be the same. For example, the first reference basis set and the second reference basis set correspond to the same fixed reference basis set.

Alternatively, the first reference basis set used by the terminal and the second reference basis set used by the network side device may be different. For example, the dynamic reference basis set corresponding to the first reference basis set is different from the dynamic reference basis set corresponding to the second reference basis set. The set lengths of the first reference basis set and the second reference basis set may be the same. Each reference basis in the first reference basis set and each reference basis in the second reference basis set are in a one-to-one correspondence.

Taking the first reference basis in the first reference basis set corresponding to the second reference basis in the second reference basis set as an example, the first reference basis is the historical channel estimation result of the terminal, or the simulated lossy channel estimation; the second reference basis is the precoding matrix result of the historical CSI reporting corresponding to the first reference basis.

It should be noted that the first reference base set and the second reference base set can be updated in a rolling manner. For example, the length or size of the reference base set is indicated by the protocol or the network side device. The updates of the two reference base sets are performed synchronously, and a certain number of expired or earliest reference bases are discarded respectively.

Optionally, the terminal calculates the K DFT orthogonal bases with the strongest projection coefficients as the content of the first reference basis set based on the channel matrix at each moment, and the network side device uses the strongest K projected DFT orthogonal bases corresponding to the channel matrix restored by the channel characteristic information reported by the terminal at the corresponding moment as the first reference basis set. Each time the terminal reports new CSI, the terminal and the network side device update their respective reference basis sets, or each time AI model monitoring is required, the terminal and the network side device update their respective reference basis sets based on the most recent or preset number of CSI.

Optionally, the implementation manner in which the terminal sends the monitoring information to the network side device may include at least one of the following manners:

Method 1: When the terminal sends the first CSI to the network side device, the terminal sends the monitoring information to the network side device; the monitoring information is carried in the first CSI, for example, an additional information can be added to the normal CSI.

Method 2: The terminal sends a second CSI to the network side device, where the second CSI is used to report the monitoring information; the second CSI is configured by the network side device through CSI report config.

Method 3: When the terminal sends a physical uplink control channel (PUCCH) to the network side device, the terminal sends the monitoring information to the network side device. For example, the base station configures periodic reporting or DCI triggers reporting.

It should be noted that the terminal can report the model monitoring information as uplink control information (Uplink Control Information, UCI) through the physical uplink shared channel (PUSCH) or PUCCH, and can also upload the model monitoring information as data through PUSCH.

In one embodiment, the channel information is a precoding vector of a layer. The protocol stipulates a common reference basis set, including 32 orthogonal orthogonal bases, each of which is a reference basis, and the dimension of each orthogonal basis is 32. In particular, these 32 orthogonal bases can be DFT orthogonal bases or other orthogonal bases. The specific orthogonal bases can be determined during the AI model training process. The terminal selects the strongest reference basis from the set:

The terminal traverses all reference bases, calculates the squared general cosine similarity (SGCS) between the precoding vector at the current moment and each reference base, and then selects the reference base with the largest SGCS as the target reference base.

Alternatively, the terminal caches the precoding vectors during each previous CSI measurement and traverses the SGCS of the reference basis set, and selects the reference basis corresponding to the historical maximum SGCS as the target reference basis.

The terminal reports the ID corresponding to the target reference base and the calculated SGCS to the base station.

After the base station receives the CSI and the monitoring information reported by the terminal, it uses the decoding model to restore the corresponding precoding matrix according to the channel characteristic information in the CSI, finds the corresponding reference base in the reference base set according to the target reference base ID, calculates the SGCS of the restored precoding matrix and the target reference base, and compares them with the SGCS reported by the terminal to determine the status of the AI model.

Optionally, the base station uses a single result. If the base station finds that the SGCS reported by the terminal is greater than the recovered SGCS and exceeds a certain threshold, the base station considers that the model is invalid and considers switching the model or switching to a non-AI CSI feedback method.

The base station uses historical results, and the base station compares the SGCS within a period of time, for example, directly summing or summing the squares.

Base station usage trend judgment, that is, there is no threshold, the base station records the difference between the two SGCS within a period of time, if there is If there is a significant improvement, or multiple significant improvements, the judgment will be invalid.

The embodiment of the present application uses a fixed reference base set to replace the transmission of input and output information with the transmission of the measurement value of the input and output information and the reference base, thereby reducing the reporting or indication overhead, thereby achieving rapid model monitoring. Since the reference base is fixed or constructed based on historical information, no interaction of any reference base set is required, and only the corresponding reference base ID needs to be reported or indicated when in use, which greatly reduces the indication and reporting overhead.

Optionally, with respect to the construction of a reference basis set, it can be orthogonal:

1. Complete orthogonal basis, using DFT orthogonal basis;

2. Incomplete orthogonal basis, that is, the number is less than the dimension;

3. Inclusion of additional non-orthogonal bases, e.g., from oversampled DFT;

4. Other orthogonal bases, including the unit matrix.

The reference basis set can also be non-orthogonal:

1. During the training process of the AI model, the training node determines and notifies the base station and terminal;

2. It is composed of historical precoding matrices or vectors,

At this time, the reference basis set used by the base station and the reference basis set used by the terminal are different, that is, the reference basis in the reference basis set of the base station is the result of CSI compression and reporting of the reference basis in the reference basis set of the terminal.

The reference base set can be updated in a rolling manner, that is, the protocol stipulates the length of a reference base set, the terminal and the base station update synchronously, and discard expired reference bases.

Optionally, the special reference basis can be a DFT orthogonal basis, represented by a DFT index, or by a Type II parameter.

Regarding the use of DFT index, you can use the Type II codebook method and use a new parameter combination

a) L = 1, M = 1, beta = 1.0;

b) L = 2, M = 1, beta = 1.0;

c) L=1, M=2, beta=1.0.

Optionally, the DFT index can be passed directly, for example:

a) The selected airspace index, or the airspace index and the oversampling factor;

b) the selected delay index, or the delay index and the oversampling factor;

c) The protocol agrees in advance on the corresponding formulas for the index and DFT orthogonal basis.

FIG. 3 is a second flow chart of the performance monitoring method of the AI unit provided in an embodiment of the present application. As shown in FIG. 3 , the method includes steps 301-302; wherein:

Step 301: The network side device receives monitoring information from the terminal;

Step 302: The network side device determines whether the AI unit is valid based on the monitoring information; wherein the AI unit includes a first unit configured at the terminal and a second unit configured at the network side device, and the monitoring information includes at least one of the following:

It should be noted that the embodiments of the present application can be applied to the scenario of performance evaluation and monitoring of AI models in communication networks. The terminal includes but is not limited to the types of terminal 11 listed above, and the present application does not limit this; the first network side device includes but is not limited to the types of network side device 12 listed above, and the present application does not limit this.

Optionally, the first reference base includes at least one reference base. The number of reference bases included in the first reference base is agreed upon by the protocol or configured by the network side device. The AI unit includes an AI model. The first unit includes a first AI model, such as an encoding model; the second unit includes a second AI model, such as a decoding model. After the first channel information is input into the first unit, the first unit performs a first process, such as a compression encoding process based on an AI model. The network side device receives the compressed content, and the second unit performs a second process, such as a decoding and recovery process based on an AI model. It should be noted that the compression encoding process based on the AI model and the decoding and recovery process based on the AI model in this application are completely different concepts from channel coding and channel decoding.

Optionally, the first correlation degree includes: a first similarity between the first channel information and a first reference basis.

Optionally, the second reference base includes at least one reference base.

Optionally, the number of reference bases included in the second reference base is agreed upon by a protocol or configured by the network side device.

Optionally, the AI unit includes an AI model.

In the performance monitoring method of the AI unit provided in the embodiment of the present application, the network side device receives monitoring information from the terminal, and the monitoring information includes first measurement information or performance monitoring results. The first measurement information indicates the first correlation between the first channel information and the first reference base, or the performance monitoring result indicates whether the AI unit is valid. After obtaining the monitoring information, the network side device can determine whether the AI unit is valid based on the monitoring information, thereby replacing the transmission of the AI unit input information/output information with the transmission of the measurement information of the first channel information and the first reference base, thereby reducing the information transmission overhead and realizing rapid model monitoring. In addition, since there is no need to transmit the reference base in real time, only the ID of the reference base needs to be transmitted when in use, which can also reduce the information transmission overhead, reduce the AI model performance monitoring overhead, and realize rapid model monitoring.

Optionally, the network side device sends indication information to the terminal; wherein the indication information is used to instruct the terminal to monitor the performance of the AI unit.

Optionally, the monitoring information includes the first measurement information and the ID of the first reference base; the network side device selects a second reference base corresponding to the ID of the first reference base in the second reference base set based on the ID of the first reference base; the network side device calculates the second correlation degree between the second channel information and the second reference base to obtain the second measurement information; the network side device determines whether the AI unit is valid based on the first measurement information and the second measurement information.

Optionally, the monitoring information includes the first measurement information. The network side device selects a second reference base from the second reference base set, and the identifier of the second reference base is a reference base ID; the network side device sends the reference base ID to the terminal. Based on the monitoring information, the network side device calculates a second correlation between the second channel information and the second reference base to obtain second measurement information; the network side device calculates a second correlation between the second channel information and the second reference base based on the first measurement information and the reference base ID. The second measurement information is used to determine whether the AI unit is valid.

Optionally, the monitoring information includes the performance monitoring result. The network side device selects a second reference base from the second reference base set, and the identifier of the second reference base is a reference base ID; the network side device calculates a second correlation between the second channel information and the second reference base to obtain second measurement information; the network side device sends the reference base ID and/or the second measurement information to the terminal. The network side device determines whether the AI unit is valid based on the performance monitoring result.

Optionally, the second reference basis set includes at least one of the following:

A fixed reference basis set, wherein the fixed reference basis set includes one or more pre-agreed or configured reference bases; the fixed reference basis set satisfies at least one of the following: agreed upon by a protocol; configured by the network side device; configured by the network side device and matched with the AI unit;

A dynamic reference basis set is determined by historical channel information.

Complete orthogonal basis set;

Incomplete orthogonal basis sets;

A set of non-orthogonal bases;

Complete or incomplete orthogonal bases, and sets of non-orthogonal bases;

The orthogonal basis set formed by each column of the unit matrix;

A set of precoding matrices or vectors determined by a training node;

A collection of historical channel information;

A set of feature vectors determined by historical channel information;

A set of feature vectors corresponding to the AI unit;

A feature vector associated with the training data distribution of the AI unit.

Optionally, when the first reference basis includes a discrete Fourier transform DFT orthogonal basis, the ID of the first reference basis includes at least one of the following:

beam index beam index and/or delay index delay index;

Type II TypeII codebook parameters.

Optionally, the implementation manner in which the network side device receives the monitoring information from the terminal may include at least one of the following:

1) The network side device receives the monitoring information sent by the terminal during the process of receiving the first channel state information CSI sent by the terminal; optionally, the monitoring information is carried in the first CSI.

2) The network side device receives the second CSI sent by the terminal, where the second CSI is used to report the monitoring information;

3) The network side device receives a physical uplink control channel PUCCH sent by the terminal, where the PUCCH carries the monitoring information.

The performance monitoring method of the AI unit provided in the embodiment of the present application may be performed by a performance monitoring device of the AI unit. In the embodiment of the present application, the performance monitoring device of the AI unit executing the performance monitoring method of the AI unit is taken as an example to illustrate the performance monitoring device of the AI unit provided in the embodiment of the present application.

FIG. 4 is a schematic diagram of a performance monitoring device for an AI unit provided in an embodiment of the present application. As shown in FIG. 4 , the performance monitoring device 400 for an AI unit is applied to a terminal and includes:

A determination module 401, configured to determine monitoring information based on the first channel information;

A first sending module 402, configured to send the monitoring information to a network side device;

In the performance monitoring device of the AI unit provided in the embodiment of the present application, based on the first channel information input to the first unit, monitoring information is determined and sent to the network side device, and the monitoring information includes first measurement information or performance monitoring results. The first measurement information indicates the first correlation between the first channel information and the first reference base, or the performance monitoring result indicates whether the AI unit is valid, so that after obtaining the monitoring information, the network side device can determine whether the AI unit is valid based on the monitoring information, thereby replacing the transmission of AI unit input information/output information with the transmission of measurement information of the first channel information and the first reference base, thereby reducing information transmission overhead and realizing rapid model monitoring. In addition, since there is no need to transmit the reference base in real time, only the ID of the reference base needs to be transmitted when in use, which can also reduce information transmission overhead, reduce AI model performance monitoring overhead, and realize rapid model monitoring.

Optionally, the device further comprises:

The first receiving module is used to receive indication information from the network side device, where the indication information is used to instruct the terminal to monitor the performance of the AI unit.

Optionally, the determination module 401 is specifically configured to select one or more reference bases from a first reference base set, and the terminal uses each of the reference bases as the first reference base, calculates the first degree of association between the first channel information and the first reference base, and obtains the first measurement information;

The monitoring information includes the first measurement information and the ID of the first reference base.

Optionally, the determination module 401 is specifically configured to:

Based on the reference base ID, selecting a first reference base corresponding to the reference base ID in the first reference base set;

The first correlation degree between the first channel information and the first reference basis is calculated to obtain the first measurement information; wherein the monitoring information includes the first measurement information.

Optionally, the device further comprises:

The second receiving module is used to receive the reference base ID sent by the network side device.

Optionally, the determination module 401 is specifically configured to:

Determining the performance monitoring result based on the first channel information, the first reference basis, and the second measurement information; Wherein, the monitoring information includes the performance monitoring result.

Optionally, the second measurement information is used to indicate a second correlation between second channel information and a second reference base, the second channel information is channel information output by the second unit, and the second reference base is a reference base corresponding to the reference base ID in a second reference base set.

Optionally, the device further comprises:

The third receiving module is used to receive the reference base ID and/or the second measurement information sent by the network side device.

Optionally, the determination module 401 is specifically configured to:

The terminal selects one or more reference bases from the first reference base set as the first reference base;

The terminal calculates the first correlation between the first channel information and the first reference base, and determines the performance monitoring result based on the first correlation; wherein the monitoring information includes the performance monitoring result; and the first reference base includes features of the training data of the AI unit.

Optionally, the first reference basis set includes at least one of the following:

A dynamic reference basis set is determined by historical channel information.

Complete orthogonal basis set;

Incomplete orthogonal basis sets;

A set of non-orthogonal bases;

Complete or incomplete orthogonal bases, and sets of non-orthogonal bases;

The orthogonal basis set formed by each column of the unit matrix;

A set of precoding matrices or vectors determined by a training node;

A collection of historical channel information;

A set of feature vectors determined by historical channel information;

A set of feature vectors corresponding to the AI unit;

A feature vector associated with the training data distribution of the AI unit.

beam index beam index and/or delay index delay index;

Type II TypeII codebook parameters.

Optionally, the first sending module is specifically used for at least one of the following:

In the process of sending the first CSI to the network side device, sending the monitoring information to the network side device;

Sending a second CSI to a network side device, where the second CSI is used to report the monitoring information;

In the process of sending a physical uplink control channel PUCCH to a network side device, the monitoring information is sent to the network side device.

Optionally, the monitoring information is carried in the first CSI.

Optionally, the first reference base includes at least one reference base.

Optionally, the number of reference bases included in the first reference base is agreed upon by a protocol or configured by the network side device.

Optionally, the AI unit includes an AI model.

FIG. 5 is a second structural diagram of a performance monitoring device of an AI unit provided in an embodiment of the present application. As shown in FIG. 5 , the performance monitoring device 500 of the AI unit is applied to a network side device, and includes:

The fourth receiving module 501 is used to receive monitoring information from the terminal;

A processing module 502, configured to determine whether the AI unit is valid based on the monitoring information;

In the performance monitoring device of the AI unit provided in the embodiment of the present application, monitoring information is received from the terminal, and the monitoring information includes first measurement information or performance monitoring results. The first measurement information indicates the first correlation between the first channel information and the first reference base, or the performance monitoring result indicates whether the AI unit is valid, so that after obtaining the monitoring information, the network side device can determine whether the AI unit is valid based on the monitoring information, thereby replacing the transmission of the AI unit input information/output information with the transmission of the measurement information of the first channel information and the first reference base, thereby reducing the information transmission overhead and realizing fast model monitoring. In addition, since there is no need to transmit the reference base in real time, only the ID of the reference base needs to be transmitted when in use, which can also reduce the information transmission overhead, reduce the AI model performance monitoring overhead, and realize fast model monitoring.

Optionally, the device further comprises:

The second sending module is used to send indication information to the terminal; wherein the indication information is used to instruct the terminal to monitor the performance of the AI unit.

Optionally, the processing module 502 is specifically configured to:

Based on the ID of the first reference base, selecting a second reference base corresponding to the ID of the first reference base in the second reference base set;

Calculating the second degree of association between the second channel information and the second reference basis to obtain the second measurement information;

Based on the first metric information and the second metric information, it is determined whether the AI unit is valid.

Optionally, the device further comprises:

A first selection module, configured to select a second reference basis from the second reference basis set, wherein the identifier of the second reference basis is a reference basis ID;

The third sending module is used to send the reference base ID to the terminal.

Optionally, the processing module 502 is specifically configured to:

Calculating a second correlation degree between the second channel information and the second reference basis to obtain second measurement information;

Optionally, the device further comprises:

A second selection module, configured to select a second reference base from the second reference base set, wherein the identifier of the second reference base is a reference base ID;

A calculation module, used for calculating a second correlation degree between the second channel information and the second reference basis to obtain second measurement information;

A fourth sending module is used to send the reference base ID and/or the second measurement information to the terminal.

Optionally, the processing module 502 is specifically configured to determine whether the AI unit is valid based on the performance monitoring result.

A dynamic reference basis set is determined by historical channel information.

Complete orthogonal basis set;

Incomplete orthogonal basis sets;

A set of non-orthogonal bases;

Complete or incomplete orthogonal bases, and sets of non-orthogonal bases;

The orthogonal basis set formed by each column of the unit matrix;

A set of precoding matrices or vectors determined by a training node;

A collection of historical channel information;

A set of feature vectors determined by historical channel information;

A set of feature vectors corresponding to the AI unit;

A feature vector associated with the training data distribution of the AI unit.

beam index beam index and/or delay index delay index;

Type II TypeII codebook parameters.

Optionally, the fourth receiving module 501 is specifically used for at least one of the following:

In a process of receiving the first channel state information CSI sent by the terminal, receiving the monitoring information sent by the terminal;

receiving a second CSI sent by the terminal, where the second CSI is used to report the monitoring information;

A physical uplink control channel PUCCH sent by the terminal is received, where the PUCCH carries the monitoring information.

Optionally, the monitoring information is carried in the first CSI.

Optionally, the second reference base includes at least one reference base.

Optionally, the AI unit includes an AI model.

The performance monitoring device of the AI unit in the embodiment of the present application can be an electronic device, such as an electronic device with an operating system, or a component in an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal, or it can be other devices other than a terminal. Exemplarily, the terminal can include but is not limited to the types of terminal 11 listed above, and other devices can be servers, network attached storage (NAS), etc., which are not specifically limited in the embodiment of the present application.

The performance monitoring device of the AI unit provided in the embodiment of the present application can implement the various processes implemented by the method embodiments of Figures 2 to 5 and achieve the same technical effect. To avoid repetition, it will not be repeated here.

FIG6 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application. As shown in FIG6, the communication device 600 includes a processor 601 and a memory 602. The memory 602 stores programs or instructions that can be run on the processor 601. For example, when the communication device 600 is a terminal, the program or instruction is executed by the processor 601 to implement the various steps of the embodiment of the performance monitoring method of the above-mentioned AI unit, and can achieve the same technical effect. When the communication device 600 is a network-side device, the program or instruction is executed by the processor 601 to implement the various steps of the embodiment of the performance monitoring method of the above-mentioned AI unit, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

The embodiment of the present application further provides a terminal, comprising a processor and a communication interface; wherein the processor is used to determine monitoring information based on the first channel information;

This terminal embodiment corresponds to the above-mentioned terminal side method embodiment. Each implementation process and implementation method of the above-mentioned method embodiment can be applied to this terminal embodiment and can achieve the same technical effect.

FIG. 7 is a schematic diagram of the structure of a terminal provided in an embodiment of the present application. As shown in FIG. 7 , the terminal 700 includes but is not limited to In: at least some of the components in the radio frequency unit 701, the network module 702, the audio output unit 703, the input unit 704, the sensor 705, the display unit 706, the user input unit 707, the interface unit 708, the memory 709 and the processor 710.

Those skilled in the art will appreciate that the terminal 700 may also include a power source (such as a battery) for supplying power to each component, and the power source may be logically connected to the processor 710 through a power management system, so as to implement functions such as managing charging, discharging, and power consumption management through the power management system. The terminal structure shown in FIG7 does not constitute a limitation on the terminal, and the terminal may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently, which will not be described in detail here.

It should be understood that in the embodiment of the present application, the input unit 704 may include a graphics processing unit (GPU) 7041 and a microphone 7042, and the graphics processor 7041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The display unit 706 may include a display panel 7061, and the display panel 7061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 707 includes a touch panel 7071 and at least one of other input devices 7072. The touch panel 7071 is also called a touch screen. The touch panel 7071 may include two parts: a touch detection device and a touch controller. Other input devices 7072 may include, but are not limited to, a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.

In the embodiment of the present application, after receiving downlink data from the network side device, the RF unit 701 can transmit the data to the processor 710 for processing; in addition, the RF unit 701 can send uplink data to the network side device. Generally, the RF unit 701 includes but is not limited to an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.

The memory 709 can be used to store software programs or instructions and various data. The memory 709 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instruction required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 709 may include a volatile memory or a non-volatile memory, or the memory 709 may include a transient and non-transient memory. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM). The memory 709 in the embodiment of the present application includes but is not limited to these and any other suitable types of memories.

The processor 710 may include one or more processing units; optionally, the processor 710 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to the operating system, user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the above-mentioned modem processor The processor may not be integrated into the processor 710.

The processor 710 is configured to determine monitoring information based on the first channel information;

The radio frequency unit 701 is used to send the monitoring information to the network side device;

The terminal provided in the embodiment of the present application determines and sends monitoring information to the network side device based on the first channel information input to the first unit. The monitoring information includes first measurement information or performance monitoring results. The first measurement information indicates the first correlation between the first channel information and the first reference base, or the performance monitoring result indicates whether the AI unit is valid, so that after obtaining the monitoring information, the network side device can determine whether the AI unit is valid based on the monitoring information, thereby replacing the transmission of AI unit input information/output information with the transmission of measurement information of the first channel information and the first reference base, thereby reducing information transmission overhead and achieving rapid model monitoring. In addition, since there is no need to transmit the reference base in real time, only the ID of the reference base needs to be transmitted when in use, which can also reduce information transmission overhead, reduce AI model performance monitoring overhead, and achieve rapid model monitoring.

The embodiment of the present application also provides a network side device, including a processor and a communication interface; wherein the communication interface is used to receive monitoring information from a terminal;

This network side device embodiment corresponds to the above-mentioned network side device method embodiment. Each implementation process and implementation method of the above-mentioned method embodiment can be applied to this network side device embodiment and can achieve the same technical effect.

FIG8 is a schematic diagram of the structure of a network side device provided in an embodiment of the present application. As shown in FIG8 , the network side device 800 includes: an antenna 801, a radio frequency device 802, a baseband device 803, a processor 804, and a memory 805. The antenna 801 is connected to the radio frequency device 802. In the uplink direction, the radio frequency device 802 receives information through the antenna 801 and sends the received information to the baseband device 803 for processing. In the downlink direction, the baseband device 803 processes the information to be sent and sends it to the radio frequency device 802. The radio frequency device 802 processes the received information and sends it out through the antenna 801.

The method executed by the network-side device in the above embodiment may be implemented in the baseband device 803, which includes a baseband processor.

The baseband device 803 may include, for example, at least one baseband board, on which a plurality of chips are arranged, as shown in FIG. As shown, one of the chips is, for example, a baseband processor, which is connected to the memory 805 through a bus interface to call the program in the memory 805 to execute the network device operations shown in the above method embodiment.

The network side device may also include a network interface 806, which is, for example, a common public radio interface (CPRI).

Specifically, the network side device 800 of the embodiment of the present application also includes: instructions or programs stored in the memory 805 and executable on the processor 804. The processor 804 calls the instructions or programs in the memory 805 to execute the method described above... and achieve the same technical effect. To avoid repetition, it will not be repeated here.

An embodiment of the present application also provides a performance monitoring system for an AI unit, including: a terminal and a network side device, wherein the terminal can be used to execute the steps of the performance monitoring method for the AI unit on the terminal side as described above, and the network side device can be used to execute the steps of the performance monitoring method for the AI unit on the network side device side as described above.

An embodiment of the present application also provides a readable storage medium, which may be volatile or non-volatile, and stores a program or instruction. When the program or instruction is executed by a processor, the various processes of the performance monitoring method embodiment of the above-mentioned AI unit are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

The processor is the processor in the terminal described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the performance monitoring method embodiment of the above-mentioned AI unit, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

The embodiment of the present application further provides a computer program/program product, which is stored in a storage medium, and is executed by at least one processor to implement the various processes of the above-mentioned AI unit performance monitoring method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises one..." does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be noted that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

Through the above description of the implementation mode, those skilled in the art can clearly understand that the above embodiment method can be It can be implemented by means of software plus a necessary general hardware platform, or by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM/RAM, disk, CD), and includes several instructions for enabling a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

A performance monitoring method for an artificial intelligence (AI) unit, comprising:

The terminal determines monitoring information based on the first channel information;

The terminal sends the monitoring information to the network side device;

The AI unit includes a first unit configured at the terminal and a second unit configured at the network side device, the first channel information is channel information input to the first unit, and the monitoring information includes at least one of the following:

first measurement information, used to indicate a first correlation between the first channel information and a first reference basis;

The first measurement information and the identification ID of the first reference base;

The performance monitoring result is used to indicate whether the AI unit is effective.
The performance monitoring method of an AI unit according to claim 1, wherein the method further comprises:

The terminal receives indication information from the network side device, where the indication information is used to instruct the terminal to monitor the performance of the AI unit.
The performance monitoring method of the AI unit according to claim 1 or 2, wherein the terminal determines the monitoring information based on the first channel information, comprising:

The terminal selects one or more reference bases from a first reference base set, uses each of the reference bases as the first reference base, calculates the first degree of association between the first channel information and the first reference base, and obtains the first metric information;

The monitoring information includes the first measurement information and the ID of the first reference base.
The performance monitoring method of the AI unit according to claim 1 or 2, wherein the terminal determines the monitoring information based on the first channel information, comprising:

The terminal selects, based on the reference base ID, a first reference base corresponding to the reference base ID in the first reference base set;

The terminal calculates the first degree of association between the first channel information and the first reference basis to obtain the first measurement information; wherein the monitoring information includes the first measurement information.
The performance monitoring method of the AI unit according to claim 4, wherein the method further comprises:

The terminal receives the reference base ID sent by the network side device.
The performance monitoring method of the AI unit according to claim 1 or 2, wherein the terminal determines the monitoring information based on the first channel information, comprising:

The terminal selects, based on the reference base ID, a first reference base corresponding to the reference base ID in the first reference base set;

The terminal determines the performance monitoring result based on the first channel information, the first reference basis and the second measurement information; wherein the monitoring information includes the performance monitoring result.
The performance monitoring method of an AI unit according to claim 6, wherein the second measurement information is used to indicate a second degree of association between second channel information and a second reference base, the second channel information is channel information output by the second unit, and the second reference base is a reference base corresponding to the reference base ID in a second reference base set.
The performance monitoring method of the AI unit according to claim 6 or 7, wherein the method further comprises:

The terminal receives the reference base ID and/or the second measurement information sent by the network side device.
The performance monitoring method of the AI unit according to claim 1 or 2, wherein the terminal determines the monitoring information based on the first channel information, comprising:

The terminal selects one or more reference bases from the first reference base set as the first reference base;

The terminal calculates the first correlation between the first channel information and the first reference base, and determines the performance monitoring result based on the first correlation; wherein the monitoring information includes the performance monitoring result; and the first reference base includes features of the training data of the AI unit.
The performance monitoring method of an AI unit according to any one of claims 3 to 9, wherein the first reference basis set includes at least one of the following:

A fixed reference basis set, wherein the fixed reference basis set includes one or more pre-agreed or configured reference bases; the fixed reference basis set satisfies at least one of the following: agreed upon by a protocol; configured by the network side device; configured by the network side device and matched with the AI unit;

A dynamic reference basis set is determined by historical channel information.
The performance monitoring method of an AI unit according to claim 10, wherein the length of the reference base in the dynamic reference base set is agreed upon by a protocol or configured by the network side device.
The performance monitoring method of an AI unit according to any one of claims 3 to 9, wherein the first reference basis set includes at least one of the following:

Complete orthogonal basis set;

Incomplete orthogonal basis sets;

A set of non-orthogonal bases;

Complete or incomplete orthogonal bases, and sets of non-orthogonal bases;

The orthogonal basis set formed by each column of the unit matrix;

A set of precoding matrices or vectors determined by a training node;

A collection of historical channel information;

A set of feature vectors determined by historical channel information;

A set of feature vectors corresponding to the AI unit;

A feature vector associated with the training data distribution of the AI unit.
The performance monitoring method of the AI model according to any one of claims 1 to 12, wherein, when the first reference basis includes a discrete Fourier transform DFT orthogonal basis, the ID of the first reference basis includes at least one of the following:

beam index beam index and/or delay index delay index;

Type II TypeII codebook parameters.
The performance monitoring method of the AI unit according to any one of claims 1 to 13, wherein the terminal sends the monitoring information to the network side device, including at least one of the following:

The terminal sends the monitoring information to the network side device during the process of sending the first CSI to the network side device;

The terminal sends a second CSI to the network side device, where the second CSI is used to report the monitoring information;

The terminal sends the monitoring information to the network side device during the process of sending a physical uplink control channel PUCCH to the network side device.
The performance monitoring method of an AI unit according to claim 14, wherein the monitoring information is carried in the first CSI.
The performance monitoring method of an AI unit according to any one of claims 1 to 15, wherein the first correlation degree comprises: a first similarity between the first channel information and a first reference basis.
The performance monitoring method of the AI model according to any one of claims 1 to 16, wherein the first reference base includes at least one reference base.
The performance monitoring method of an AI unit according to claim 17, wherein the number of reference bases included in the first reference base is agreed upon by a protocol or configured by the network side device.
The performance monitoring method of an AI unit according to any one of claims 1 to 18, wherein the AI unit comprises an AI model.
A performance monitoring method for an artificial intelligence (AI) unit, comprising:

The network side device receives monitoring information from the terminal;

The network side device determines whether the AI unit is valid based on the monitoring information;

The AI unit includes a first unit configured on the terminal and a second unit configured on the network side device, and the monitoring information includes at least one of the following:

first measurement information, used to indicate a first correlation between first channel information and a first reference basis, the first channel information being channel information input to the first unit;

The first measurement information and the identification ID of the first reference base;

The performance monitoring result is used to indicate whether the AI unit is effective.
The performance monitoring method of an AI unit according to claim 20, wherein the method further comprises:

The network side device sends indication information to the terminal; wherein the indication information is used to instruct the terminal to monitor the performance of the AI unit.
The performance monitoring method of the AI unit according to claim 20 or 21, wherein the network side device determines whether the AI unit is valid based on the monitoring information, comprising:

The network-side device selects, based on the ID of the first reference base, a second reference base corresponding to the ID of the first reference base from the second reference base set;

The network side device calculates the second correlation degree between the second channel information and the second reference basis to obtain the second measurement information;

The network-side device determines whether the AI unit is valid based on the first metric information and the second metric information.
The performance monitoring method of an AI unit according to claim 20 or 21, wherein the method further comprises:

The network side device selects a second reference base from the second reference base set, where the identifier of the second reference base is a reference Base ID;

The network side device sends the reference base ID to the terminal.
The performance monitoring method of the AI unit according to claim 23, wherein the network side device determines whether the AI unit is valid based on the monitoring information, comprising:

The network side device calculates a second correlation degree between the second channel information and the second reference basis to obtain second measurement information;

The network-side device determines whether the AI unit is valid based on the first metric information and the second metric information.
The performance monitoring method of an AI unit according to claim 20 or 21, wherein the method further comprises:

The network side device selects a second reference base from the second reference base set, where the identifier of the second reference base is a reference base ID;

The network side device calculates a second correlation degree between the second channel information and the second reference basis to obtain second measurement information;

The network side device sends the reference base ID and/or the second metric information to the terminal.
The performance monitoring method of the AI unit according to claim 25, wherein the network side device determines whether the AI unit is valid based on the monitoring information, comprising:

The network side device determines whether the AI unit is valid based on the performance monitoring result.
The performance monitoring method of an AI unit according to any one of claims 22 to 26, wherein the second reference basis set includes at least one of the following:

A fixed reference basis set, wherein the fixed reference basis set includes one or more pre-agreed or configured reference bases; the fixed reference basis set satisfies at least one of the following: agreed upon by a protocol; configured by the network side device; configured by the network side device and matched with the AI unit;

A dynamic reference basis set is determined by historical channel information.
The performance monitoring method of an AI unit according to claim 27, wherein the length of the reference base in the dynamic reference base set is agreed upon by a protocol or configured by the network side device.
The performance monitoring method of an AI unit according to any one of claims 22 to 26, wherein the reference basis in the second reference basis set includes at least one of the following:

Complete orthogonal basis set;

Incomplete orthogonal basis sets;

A set of non-orthogonal bases;

Complete or incomplete orthogonal bases, and sets of non-orthogonal bases;

The orthogonal basis set formed by each column of the unit matrix;

A set of precoding matrices or vectors determined by a training node;

A collection of historical channel information;

A set of feature vectors determined by historical channel information;

A set of feature vectors corresponding to the AI unit;

A feature vector associated with the training data distribution of the AI unit.
The performance monitoring method of an AI unit according to any one of claims 20 to 29, wherein, in the case where the first reference basis includes a discrete Fourier transform DFT orthogonal basis, the ID of the first reference basis includes at least one of the following:

beam index beam index and/or delay index delay index;

Type II TypeII codebook parameters.
The performance monitoring method of the AI unit according to any one of claims 20 to 30, wherein the monitoring information received by the network side device from the terminal includes at least one of the following:

The network side device receives the monitoring information sent by the terminal during the process of receiving the first channel state information CSI sent by the terminal;

The network side device receives the second CSI sent by the terminal, where the second CSI is used to report the monitoring information;

The network side device receives a physical uplink control channel PUCCH sent by the terminal, where the PUCCH carries the monitoring information.
The performance monitoring method of an AI unit according to claim 31, wherein the monitoring information is carried in the first CSI.
The performance monitoring method of an AI unit according to any one of claims 20 to 32, wherein the first correlation degree comprises: a first similarity between the first channel information and a first reference base.
The performance monitoring method of the AI model according to any one of claims 20 to 33, wherein the second reference base includes at least one reference base.
The performance monitoring method of an AI unit according to claim 34, wherein the number of reference bases included in the second reference base is agreed upon by a protocol or configured by the network side device.
The performance monitoring method of an AI unit according to any one of claims 20 to 35, wherein the AI unit comprises an AI model.
A performance monitoring device for an artificial intelligence (AI) unit, comprising:

A determination module, configured to determine monitoring information based on the first channel information;

A first sending module, used to send the monitoring information to a network side device;

The AI unit includes a first unit configured at the terminal and a second unit configured at the network side device, the first channel information is channel information input to the first unit, and the monitoring information includes at least one of the following:

first measurement information, used to indicate a first correlation between the first channel information and a first reference basis;

The first measurement information and the identification ID of the first reference base;

The performance monitoring result is used to indicate whether the AI unit is effective.
A performance monitoring device for an artificial intelligence (AI) unit, comprising:

A fourth receiving module, used for receiving monitoring information from a terminal;

a processing module, configured to determine whether the AI unit is valid based on the monitoring information;

The AI unit includes a first unit configured on the terminal and a second unit configured on a network side device. The monitoring information includes at least one of the following:

first measurement information, used to indicate a first correlation between first channel information and a first reference basis, the first channel information being channel information input to the first unit;

The first measurement information and the identification ID of the first reference base;

The performance monitoring result is used to indicate whether the AI unit is effective.
A terminal, comprising a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of the performance monitoring method of the AI unit as described in any one of claims 1 to 19 are implemented.
A network side device, comprising a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of the performance monitoring method of the AI unit as described in any one of claims 20 to 36 are implemented.
A readable storage medium, wherein the readable storage medium stores a program or instruction, and when the program or instruction is executed by a processor, it implements the performance monitoring method of an AI unit as described in any one of claims 1 to 19, or implements the steps of the performance monitoring method of an AI unit as described in any one of claims 19 to 36.