Sensing processing method, integrated sensing and communication system, communication device, and storage medium

Abstract

The present invention provides a sensing processing method, an integrated sensing and communication system, a communication device, and a storage medium. The sensing processing method comprises: obtaining a trained artificial intelligence model from a sensing server; on the basis of the artificial intelligence model, processing sensing data to obtain a sensing result; and sending the artificial intelligence sensing result to a core network element.

Description

Perception processing method, synaesthesia integrated system, communication device and storage medium

[0001] This application claims priority to Chinese patent application No. 202311866134.7 filed on December 29, 2023, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present disclosure relates to the field of communication technology, and in particular to a perception processing method, a synaesthesia integration system, a communication device, and a storage medium. Background Art

[0003] With the widespread adoption of fifth-generation (5G) communication systems, 5G is not only meeting the needs of individual users, but is also gradually penetrating various industries and fields, achieving an upgrade from consumer to industrial applications. Synaesthesia, based on relevant network architectures and communication waveforms, utilizes relevant communication equipment to implement perceptual applications, driving the development of innovative application services and becoming a key research direction for 5G. Summary of the Invention

[0004] In the first aspect, a perception processing method is provided, which is applied to a perception functional entity. The perception processing method includes: obtaining a trained artificial intelligence model from a perception server; processing perception data based on the artificial intelligence model to obtain a perception result; and sending the artificial intelligence perception result to a core network element.

[0005] On the second aspect, a perception processing method is provided, which is applied to a perception server. The perception processing method includes: training an artificial intelligence model, which is used to process perception data to obtain perception results; and sending the trained artificial intelligence model to a perception function entity.

[0006] On the third aspect, a synaesthesia integration system is provided, which includes: a perception server for training artificial intelligence models; a perception functional entity for obtaining the trained artificial intelligence models from the perception server; processing perception data based on the artificial intelligence model to obtain perception results, and sending the artificial intelligence perception results to the core network network element.

[0007] In a fourth aspect, a communication device is provided, comprising: a memory and a processor. The memory is coupled to the processor; the memory is used to store a computer program; and the processor implements the aforementioned perception processing method when executing the computer program.

[0008] In a fifth aspect, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the above-mentioned perception processing method is implemented.

[0009] In a sixth aspect, a computer program product is provided, which includes computer program instructions, and when the computer program instructions are executed by a processor, the above-mentioned perception processing method is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] To more clearly illustrate the technical solutions of the present disclosure, the following briefly introduces the drawings required for use in some embodiments of the present disclosure. Obviously, the drawings described below are only drawings of some embodiments of the present disclosure, and those skilled in the art can also derive other drawings based on these drawings.

[0011] FIG1 is a schematic diagram of a perception scenario according to some embodiments of the present disclosure.

[0012] FIG2 is a schematic diagram of the architecture of a synaesthesia integration system according to some embodiments of the present disclosure.

[0013] FIG3 is a flow chart of a model training method according to some embodiments of the present disclosure.

[0014] FIG4 is a flow chart of a perception processing method according to some embodiments of the present disclosure.

[0015] FIG5 is a flowchart of another perception processing method according to some embodiments of the present disclosure.

[0016] FIG6 is a schematic diagram of a range Doppler map according to some embodiments of the present disclosure.

[0017] FIG7 is a schematic diagram of a corresponding relationship of feature data according to some embodiments of the present disclosure.

[0018] FIG8 is a schematic diagram of the architecture of a convolutional neural network according to some embodiments of the present disclosure.

[0019] FIG9 is a schematic diagram of the structure of a long short-term memory neural network model according to some embodiments of the present disclosure.

[0020] FIG10 is a schematic structural diagram of another long short-term memory neural network model according to some embodiments of the present disclosure.

[0021] FIG11 is a flow chart of another perception processing method according to some embodiments of the present disclosure.

[0022] FIG12 is a schematic structural diagram of a perception processing device according to some embodiments of the present disclosure.

[0023] FIG13 is a schematic structural diagram of another perception processing device according to some embodiments of the present disclosure.

[0024] FIG14 is a schematic structural diagram of a communication device according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions of this disclosure in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of this disclosure, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0026] It should be noted that, in this disclosure, words such as "exemplary" or "for example" are used to describe examples, illustrations, or explanations. Any embodiment or design described in this disclosure using words such as "exemplary" or "for example" should not be interpreted as being more preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0027] In the following, the terms "first," "second," etc. are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Thus, a feature defined by "first," "second," etc. may explicitly or implicitly include one or more of the features.

[0028] In the description of this disclosure, unless otherwise specified, " / " means "or". For example, A / B can mean A or B. "And / or" in this document is only used to describe the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can mean: only A, only B, and A and B. In addition, "at least one" means one or more, and "a plurality" means two or more.

[0029] The methods provided in the embodiments of the present disclosure can be applied to various communication systems. For example, the communication system can be a 5G communication system, a Wi-Fi (Wireless Fidelity) system, a 3GPP (3rd Generation Partnership Project)-related communication system, a future-evolved communication system (such as a sixth-generation (6G) communication system), or a system integrating multiple systems, and the embodiments of the present disclosure are not limited thereto.

[0030] With the widespread application of 5G, 5G is gradually penetrating into various industries and fields of society while meeting the needs of individual users, thus realizing the upgrade from consumer to industrial application. Synaesthesia integration is based on relevant network architecture and communication waveforms, and uses relevant communication equipment to realize perception applications, which promotes the development of innovative application services and is an important research direction of 5G. As shown in Figure 1, it is a schematic diagram of a perception scenario according to an embodiment of the present disclosure. The perception system in the perception base station 1 can perform base station perception of the drone 2, the human body 3, the vehicle 4 and the building 5, the perception system in the satellite 6 can perform satellite perception of the aircraft 7 (illustrated as an airplane), the perception system in the drone 2 can perform drone perception of the human body, and the perception terminal in the vehicle 4 can perform terminal perception of the human body 3.

[0031] In some embodiments, the terminal for terminal perception can also be a device with full-duplex transmission capability. The terminal can be a mobile phone, a tablet computer, a computer with wireless transceiver function, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, etc. The embodiments of the present disclosure do not limit the application scenarios. The terminal can sometimes also be referred to as a user, user equipment (UE), an access terminal, a UE unit, a UE station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a UE terminal, a wireless communication device, a UE agent or a UE device, etc., and the embodiments of the present disclosure do not limit this.

[0032] Most current perception functions utilize traditional algorithms. As network scale increases, service types increase, and terminal devices become more complex and diverse, optimizing perception algorithms becomes increasingly difficult. Simultaneously, with the advancement of software and hardware resources, artificial intelligence (AI) is increasingly being applied in communication systems. However, AI-based integrated sensing architectures have yet to be effectively developed, making them unsuitable for complex and diverse scenarios and applications, resulting in limited intelligent management capabilities.

[0033] In response to the above technical problems, the embodiments of the present disclosure provide a perception processing method, the idea of ​​which is: based on the relevant core network, a new perception function entity is added, the change to the network architecture is relatively small, and the realization of the perception function is easy; at the same time, since the storage capacity of the perception server is strong and the computing power requirement for the perception function entity is relatively low, the artificial intelligence model is trained through the perception server, and the perception data is processed based on the trained artificial intelligence model to obtain the perception result, so that the method of obtaining the perception result is more accurate and less prone to errors, and it is convenient to open data to other modules in the system or third parties, so as to realize real-time communication between each module in the integrated synaesthesia system and complete the perception task.

[0034] Refer to Figure 2, which is a schematic diagram of the architecture of a synaesthesia integration system according to an embodiment of the present disclosure. As shown in Figure 2, the synaesthesia integration system includes: a core network, a perception server, a perception function entity, and a perception node.

[0035] The core network includes at least one of the following network elements: network slice selection function (NSSF), network exposure function (NEF), network repository function (NRF), policy control function (PCF), unified data management (UDM), application function (AF), authentication server function (AUSF), access and mobility management function (AMF), session management function (SMF), user plane function (UPF) or other possible network functions (NF).

[0036] In some embodiments, NSSF uses slicing technology to virtualize multiple end-to-end networks based on a common hardware. Each of the multiple end-to-end networks has different NFs to adapt to different types of service requirements.

[0037] In some embodiments, the NEF is located between the 5G core network and external third-party application functions, responsible for managing open network data. All external applications that want to access data within the 5G core network must go through the NEF.

[0038] In some embodiments, the NRF is used to register, manage, and monitor the status of NFs, enabling automated management of all NFs. Upon startup, each NF must register with the NRF before it can provide services. Registration information includes the NF's type, address, and service list.

[0039] In some embodiments, the PCF is used to manage network behavior using a unified policy framework and to execute relevant policies in conjunction with user information in a unified data repository (UDR).

[0040] In some embodiments, the UDM is used to manage user identification, subscription data, and authentication data, and is also used to manage the user's service network element registration.

[0041] In some embodiments, AF refers to various application layer services. It can be an internal application of the operator or a third-party AF (such as a video server or game server). If it is an internal AF of the operator, it is in the same trusted domain as other NFs and can directly interact with other NFs. However, third-party AFs are not in the trusted domain and need to access other NFs through the NEF.

[0042] In some embodiments, AUSF is used to receive a request from AMF to authenticate the user equipment (UE), request a key from UDM, and then forward the key issued by UDM to AMF for authentication processing.

[0043] In some embodiments, AMF is used to provide a session management message transmission channel for UE and SMF, and to provide authentication, authorization functions, terminals and wireless core network control plane access points for user access.

[0044] In some embodiments, SMF is used together with AMF to support customized mobility management solutions, such as mobile initiated connection only (MICO) or radio access network (RAN) enhancements.

[0045] In some embodiments, the UPF is mainly used to send UE service data to the data network (DN), identify data and services, execute actions and policies, etc.

[0046] In some embodiments, the perception server is used to train an artificial intelligence model. The artificial intelligence model is used to process perception data to obtain perception results.

[0047] In some embodiments, the training of the artificial intelligence model can be carried out by offline training and online prediction, or by online training and online prediction.

[0048] See Figure 3, which is a flow chart of a model training method according to an embodiment of the present disclosure. As shown in Figure 3, taking the offline training and online prediction method of the artificial intelligence model as an example, the training of the artificial intelligence model includes four stages: the initial stage, the training stage, the prediction stage, and the collection stage.

[0049] In the initial stage of AI model training, the perception server can store an initial data set using simulated data or actual scene-measured data, and use the initial data set for initial training to obtain an initial training model. During the training phase of AI model training, when the training data accumulates to a certain amount, periodic AI model training is performed. During the prediction phase of AI model training, the perception function entity performs real-time online predictions and transmits high-confidence pre-processed data and corresponding perception results to the perception server for training sample collection. During the collection phase of AI model training, the perception server continuously collects data for periodic training of the AI ​​model.

[0050] In some embodiments, the online training and online prediction methods of the artificial intelligence model refer to: the artificial intelligence function in the perception function entity can perform some lightweight online training and online prediction, and pass the high-confidence pre-processed data and corresponding perception results to the perception server for training sample collection.

[0051] In some embodiments, the perception server is also used to send the trained artificial intelligence model to the perception function entity.

[0052] In some embodiments, the perception server is further configured with at least one of the following functions: service control function, data preprocessing function, data storage function, and data opening function.

[0053] The service control functions include at least one of the following: management of artificial intelligence applications, management of model training time, management of data openness, and management of information transmission; data preprocessing function, used to extract feature data or normalize perception data; data storage function, used to store perception data with higher confidence as a data set for subsequent periodic training of artificial intelligence models; data openness function, used to provide at least one of perception data, perception results, and artificial intelligence models to the outside world.

[0054] In some embodiments, the management of artificial intelligence applications is used to determine whether the perception server performs training of the artificial intelligence model; the management of model training time is used to determine the training cycle of the artificial intelligence model; the management of data openness function is used to determine whether the perception data and perception results are open and shared; and the management of information transmission is used to determine whether the open and shared perception data and perception results are transmitted.

[0055] In some embodiments, the perception server is also used to send the trained artificial intelligence model to the perception function entity.

[0056] In some embodiments, the perception server is further used to send at least one of the following to the core network network element: perception data, perception results, and artificial intelligence models.

[0057] In some embodiments, the perception server is further configured to receive perception data sent by the perception node.

[0058] In some embodiments, the perception functional entity is used to obtain a trained artificial intelligence model from a perception server.

[0059] In some embodiments, the perception functional entity is also used to process perception data based on an artificial intelligence model to obtain perception results.

[0060] The perception results are used for at least one of the following: target detection, trajectory tracking, trajectory correction, target association, and target recognition.

[0061] In some embodiments, the perception function entity is further configured with at least one of the following functions: perception control function, perception management function, artificial intelligence function, perception processing function, and data preprocessing function.

[0062] In some embodiments, the perception control function includes at least one of the following: selection of perception nodes, management of perception resources, management of perception time, management of artificial intelligence applications, and management of information transmission.

[0063] In some embodiments, the perception management function includes at least one of the following: obtaining perception business needs, authenticating the perception business needs, and determining perception tasks based on the perception business needs.

[0064] In some embodiments, the artificial intelligence function includes at least one of the following: training an artificial intelligence model and using an artificial intelligence model.

[0065] In some embodiments, the perception processing function is used to perform target detection, false alarm suppression, target association, trajectory tracking, target identification, and trajectory correction of perception data.

[0066] In some embodiments, the data preprocessing function is used to extract feature data or normalize the perception data.

[0067] In some embodiments, the perception function entity is also used to process perception data based on an artificial intelligence model, obtain perception results, and send the perception results to the core network network element.

[0068] In some embodiments, the perception function entity may send at least one of the following to the core network element: a perception service request message, a perception service response message, and a perception result.

[0069] The awareness service request message is used to indicate awareness service requirements.

[0070] In some embodiments, the perception function entity may also send a perception control instruction to the perception node, where the perception control instruction is used to select the perception node to perform the perception task.

[0071] In some embodiments, the perception function entity may also send at least one of the following to the perception server: first indication information for indicating whether the perception server performs model training, and second indication information for indicating a training cycle.

[0072] For example, the perception control function in the perception function entity sends at least one of the following to the perception server: first indication information for indicating whether the perception server performs model training, and second indication information for indicating a training cycle.

[0073] In some embodiments, the sensing node is used to send a sensing signal and obtain sensing data.

[0074] For example, after the perception node receives the perception service request message sent by the perception function entity, the perception node sends a perception signal according to the perception mode indicated by the perception service request message, performs the perception task, and obtains perception data.

[0075] It should be noted that the perception data can be measurement data of the physical layer or perception result data.

[0076] In some embodiments, the sensing node is further configured to send sensing data to the sensing function entity.

[0077] For example, after the perception node obtains the perception data, it sends the perception data to the perception function entity.

[0078] In some embodiments, referring to FIG2 , there may be multiple sensing nodes, and communication between the sensing nodes is achieved through a Uu interface (U stands for user to network interface, u stands for Universal), and each sensing node may be a generation node B (gNB) base station or a UE.

[0079] In some embodiments, the perception data or perception results processed by the UE are transmitted to the perception function entity through the Uu interface and the gNB base station.

[0080] In some embodiments, the sensing node is also used for sensing signal processing and communication signal processing.

[0081] It should be noted that the system architecture and application scenarios described in the embodiments of the present disclosure are intended to more clearly illustrate the technical solutions of the embodiments of the present disclosure and do not constitute a limitation on the technical solutions provided by the embodiments of the present disclosure. Persons skilled in the art will appreciate that, as system architectures evolve and new business scenarios emerge, the technical solutions provided by the embodiments of the present disclosure will also be applicable to similar technical problems.

[0082] A perception processing method provided by an embodiment of the present disclosure is described below with reference to the accompanying drawings.

[0083] 4 is a flow chart of a perception processing method according to an embodiment of the present disclosure, which is applied to a perception function entity. As shown in FIG4 , the perception processing method includes the following steps S101 to S103 .

[0084] In S101, the perception function entity obtains the trained artificial intelligence model from the perception server.

[0085] In some embodiments, after the perception function entity receives the perception service request message sent by the core network network element, the perception function entity obtains the trained artificial intelligence model from the perception server.

[0086] In some embodiments, the awareness service request message may carry the awareness service type, awareness area, awareness reporting content, and awareness reporting type.

[0087] The perception service type is drone perception, vehicle perception, or perception of other targets. The perception area can be within a radius of 1,000 meters centered on the perception node or other ranges. The perception report content can include the position and speed of the target, and the perception report type can be a perception result report.

[0088] It should be noted that the target may also be a person, a car or other object with a moving track or a state change, and the present disclosure does not limit the type of the target.

[0089] In S102, the perception function entity processes the perception data based on the artificial intelligence model to obtain a perception result.

[0090] In some embodiments, the perception function entity selects a perception node to perform perception detection in response to a perception service request message, and obtains perception data obtained by the perception node during the perception detection process.

[0091] In some embodiments, the above-mentioned perception function entity selects a perception node to perform perception detection in response to a perception service request message, and obtains perception data obtained by the perception node during the perception detection process, which can be implemented as the following A1 to A3.

[0092] In A1, the sensing function entity determines the sensing task based on the sensing service request message.

[0093] For example, the perception function entity responds to the perception service request message, authenticates the perception service request message through the perception management function in the perception function entity, and after the perception service request message is authenticated, the perception management function determines the perception task based on the perception service request message (such as perceiving the drones within the perception area, mainly perceiving the position and speed of the drones).

[0094] In A2, the sensing function entity selects a sensing node to perform the sensing task and sends a sensing control instruction to the sensing node.

[0095] The sensing control instruction is used to instruct the sensing node to perform the sensing task. The sensing control instruction includes sensing configuration information, which is used to configure at least one of sensing resources, sensing time, and sensing signals.

[0096] For example, after the perception task is determined, the perception management function in the perception function entity sends a perception service request message to the perception control function; the perception control function in the perception function entity selects the perception node and perception method to perform the perception task according to the perception service request message, and sends a perception control instruction to the perception node.

[0097] In some embodiments, the sensing node may be a base station, and the sensing mode may be A-transmit-A-receive or A-transmit-B-receive.

[0098] It should be noted that A sends and A receives refers to sending a perception service request message through perception node A (or base station A) and receiving perception data through perception node A (or base station A); A sends and B receives refers to sending a perception service request message through perception node A (or base station A) and receiving perception data through perception node B (or base station B).

[0099] Furthermore, the sensing node starts to perform the sensing task, performs sensing detection on the sensing area, obtains the sensing data obtained in the sensing detection process, and sends the sensing data to the sensing function entity.

[0100] In some embodiments, the sensing node receives a sensing signal reflected from a target according to the synaesthesia-integrated sensing time-frequency resource indicated by the sensing control function, performs baseband processing on the sensing signal, and obtains range-doppler (RD) map data of the target.

[0101] In A3, the sensing function entity receives the sensing data sent by the sensing node.

[0102] In some embodiments, after the sensing node obtains the sensing data obtained during the sensing detection process, it sends the sensing data (such as RD graph data) to the sensing function entity, and the sensing function entity receives the sensing data.

[0103] Furthermore, after the perception function entity receives the perception data sent by the perception node, the perception function entity processes the perception data based on the artificial intelligence model to obtain the perception result.

[0104] In S103, the sensing function entity sends the sensing result to the core network element.

[0105] In some embodiments, after the perception function entity obtains the perception result, it sends the perception result to the core network element.

[0106] It can be understood that based on the perception processing method provided by the embodiment of the present disclosure and based on the relevant core network network architecture, by obtaining the trained artificial intelligence model from the perception server with higher storage capacity, the problem of large core network processing data volume caused by training the artificial intelligence model in the core network is avoided, and the speed of model acquisition is improved; at the same time, the perception data is processed based on the trained artificial intelligence model, so that the obtained perception results are more accurate, and the automatic processing of perception data realizes the intelligent management of perception data, reducing manpower investment; finally, the obtained perception results are sent to the core network network element, so that data of each module in the system can be shared, and the integration of perception and communication is realized.

[0107] In some embodiments, an artificial intelligence model can be trained via a perception server. After a perception function entity acquires perception data sent by a perception node, the perception processing function within the perception function entity preprocesses the perception data and then transmits it to the perception server. The perception server trains the artificial intelligence model based on the preprocessed perception data. Figure 5 is a flow diagram of another perception processing method according to an embodiment of the present disclosure, applied to a perception server. As shown in Figure 5, the perception processing method includes the following steps S201 and S202.

[0108] In S201, the perception server trains the artificial intelligence model.

[0109] Artificial intelligence models are used to process perception data to obtain perception results.

[0110] In some embodiments, the perception server trains the artificial intelligence model in response to first indication information sent by the perception control function to instruct the perception server to perform model training.

[0111] Referring to FIG6, FIG6 is a schematic diagram of a range Doppler diagram according to an embodiment of the present disclosure. As shown in FIG6, P (m，n) Represents the power of the grid at the mth row and nth column in the RD diagram, where m and n are integers greater than or equal to 0.

[0112] In some embodiments, in the initial stage of training the artificial intelligence model, the perception server uses the initial test data to perform model training on the artificial intelligence model. After the RD graph of the perception signal is determined, the units to be tested and the protection units of the RD data are extracted from the RD graph based on the reference units in the RD graph. Figure 6 shows the power data of 9 grids, which constitute the feature data of [P0, P1, P2, P3, P4, P5, P6, P7, P8, P9]. This group of feature data is taken as a sample data, and its corresponding label is 0 or 1, 0 represents no target, and 1 represents a target. The corresponding relationship is shown in Figure 7. Figure 7 is a schematic diagram of the corresponding relationship of feature data according to an embodiment of the present disclosure.

[0113] In some embodiments, the present disclosure uses a convolutional neural network (CNN) network to train the model. FIG8 is a schematic diagram of the architecture of a convolutional neural network according to an embodiment of the present disclosure. As shown in FIG8, Conv L j (P j ,P j ) in L j is the number of convolution kernels, (P j ,P j ) represents the convolution kernel size, and j is an integer greater than or equal to 1.

[0114] It should be noted that network training is not limited to CNN networks, and other networks can also be used, such as residual neural networks (ResNet).

[0115] In some embodiments, in addition to the input layer and the fully connected layer, the CNN network structure may also have J layers of networks in the middle, each layer of the network includes convolution, activation and pooling operations, and J is an integer greater than or equal to 1.

[0116] It should be noted that the CNN network model is a type of feedforward neural network that includes convolution calculations and has a deep structure. The CNN network model has representation learning capabilities and can perform translation-invariant classification of input information according to its hierarchical structure.

[0117] In some embodiments, the activation function of the CNN network adopts a linear rectification function (ReLU), as shown in formula (1): x =max(0,x) Formula (1)

[0118] It should be noted that the ReLU function, also known as the rectified linear unit, is an activation function commonly used in artificial neural networks, usually referring to nonlinear functions represented by ramp functions and their variants.

[0119] In some embodiments, the activation function of the fully connected layer uses the Sigmoid function, as shown in formula (2):

[0120] It's important to note that the sigmoid function is a common S-shaped function in biology, also known as the S-shaped growth curve. In information science, due to its monotonic increasing properties and the monotonic increasing properties of its inverse function, the sigmoid function is often used as an activation function in neural networks, mapping variables to the range (0, 1).

[0121] In some embodiments, the loss function used by the fully connected layer is a cross entropy function, as shown in formula (3):

[0122] Among them, O s is the size of the output, y i is the true value of the sample, The predicted value output by the model.

[0123] It should be noted that the cross-entropy function is mainly used to measure the difference between two probability distributions. The meaning of cross-entropy is the difficulty of using the model to recognize text, or from the perspective of compression, the average number of bits required to encode each word.

[0124] In some embodiments, the initial test data can be divided into 80% of the data as a training set and 20% of the data as a test set. The training set is used to train the artificial intelligence model, and the test set is used to test the artificial intelligence model.

[0125] In S202, the perception server sends the trained artificial intelligence model to the perception function entity.

[0126] In some embodiments, the perception server sends the trained artificial intelligence model to the perception function entity, which processes the perception data based on the artificial intelligence model and extracts the data of the RD graph. As shown in Figure 6 above, feature data [P0, P1, P2, P3, P4, P5, P6, P7, P8, P9] is extracted and used as the input of the artificial intelligence model. The online prediction output is no target (0) or target (1).

[0127] In some embodiments, for characteristic data of a target, other sensing detection is performed on the target, such as direction of arrival (DOA) estimation, signal-noise ratio (SNR) estimation, speed estimation, power calculation, etc.

[0128] It should be noted that DOA estimation is a positioning technology that obtains the distance and direction information of the target by processing the received echo signal. The SNR refers to the ratio of the strength of the received useful signal to the strength of the received interference signal (noise and interference).

[0129] Furthermore, based on the perception data of the above-mentioned perception detection, target detection, trajectory tracking, trajectory correction, target association, and target recognition are performed on the target.

[0130] It should be noted that if it is multi-sensory node perception, data fusion between the multiple sensing nodes is performed to obtain more complete and accurate perception results of target detection, trajectory tracking, trajectory correction, target association, and target recognition.

[0131] In some embodiments, the fusion of perception results among multiple perception nodes can be achieved by merging, optimizing, or the like, or by using video.

[0132] In some embodiments, the perception function entity transmits the perception results with higher confidence and the corresponding labels to the perception server for data collection, and the perception server performs periodic model training on the artificial intelligence model according to the second indication information sent by the perception control function.

[0133] In some embodiments, the perception server sends at least one of the perception data, perception results, or artificial intelligence model to the core network element.

[0134] In some embodiments, the perception processing function in the perception function entity sends response information of the perception service request message to the perception management function; the perception management function sends response information of the perception service request message to the core network, and reports the perception result according to the perception service request message.

[0135] In some embodiments, the perception function entity performs target detection, trajectory tracking, trajectory correction, target association, and target identification based on the DOA estimation, SNR estimation, speed estimation, power calculation and other information reported by the perception node to obtain the target's movement trajectory.

[0136] In some embodiments, wireless communications, especially those in high-frequency bands, are prone to signal obstruction, resulting in signal blind spots, which can lead to missing target tracks. In this case, the missing track can be predicted based on known historical track information.

[0137] In some embodiments, in the initial stage of training the artificial intelligence model, the perception server uses initial test data to train the artificial intelligence model, and the initial test data is actual test trajectory data that can also be the target.

[0138] For example, the trajectory point sequence of the target can be expressed as [(Lo1, La1, H1, V1), (Lo2, La2, H2, V2), ..., (Lo n , La n , H n , V n )].

[0139] Lo t La represents the longitude of the target trajectory point at time t, t represents the latitude of the target trajectory point perceived at time t, H t represents the height of the target trajectory point perceived at time t, V t Indicates that the target is perceived at the trajectory point (Lo t , La t , H t ), t=1,2,…,n, where n is an integer greater than or equal to 1.

[0140] In some embodiments, after the actual test trajectory data of the target is obtained, data preprocessing is required for the actual test trajectory data, and the actual test trajectory data is normalized between the minimum and maximum values ​​within a certain range to achieve standardization of the training data.

[0141] In some embodiments, the trajectory data of adjacent moments in the trajectory point sequence are differentiated, that is, x t =Lo t+1 -Lo t ,y t =La t+1 -La t , z t =H t+1 -H t , v t =Vt+1 -V t .

[0142] Furthermore, we get (n-1) trajectory sequences, namely [(x1, y1, z1, v1), (x2, y2, z2, v2), ..., (x n-1 ,y n-1 , z n-1 , v n-1 )].

[0143] In some embodiments, the sklearn tool in the machine learning library is used to normalize the differential trajectory sequence using the MinMaxScaler() function, scale it to the range of [-1, 1], and save the output parameter scaler of the function. The normalized data is used to construct a supervised sequence. Table 1 is a schematic diagram of a construction method of a supervised sequence provided in some embodiments of the present disclosure, taking n=13 as an example, as shown in Table 1:

[0144] Table 1

[0145] For example, if the constructed supervision sequence is [(x1, y1, z1, v1), (x2, y2, z2, v2), ..., (x 12 ,y 12 , z 12 , v 12 )], then the label of the supervised sequence is (x 13 ,y 13 , z 13 , v 13 ); If the constructed supervision sequence is [(x2, y2, z2, v2), (x3, y3, z3, v3), ..., (x 13 ,y 13 , z 13 , v 13 )], then the label of the supervised sequence is (x 14 ,y 14 , z 14 , v 14 ); If the constructed supervision sequence is [(x3, y3, z3, v3), (x4, y4, z4, v4), ..., (x 14 ,y 14 , z 14 , v 14 )], then the label of the supervised sequence is (x 15 ,y 15 , z 15 , v 15 ); If the constructed supervision sequence is [(x4, y4, z4, v4), (x5, y5, z5, v5), ..., (x15 ,y 15 , z 15 , v 15 )], then the label of the supervised sequence is (x 16 ,y 16 , z 16 , v 16 ); If the constructed supervision sequence is [(x t-12 ,y t-12 , z t-12 , v t-12 ), (x t-11 ,y t-11 , z t-11 , v t-11 ),...,(x t-1 ,y t-1 , z t-1 , v t-1 )], then the label of the supervised sequence is (x t ,y t , z t , v t ).

[0146] In some embodiments, the processed data set is split into a training data set and a test data set in a ratio of 80% and 20%. The training data set is used to train the artificial intelligence model, and the test data set is used to evaluate the artificial intelligence model.

[0147] In some embodiments, a long short term memory (LSTM) neural network model can be constructed, and the network model parameters can be configured to train the artificial intelligence model.

[0148] In some embodiments, the LSTM network model includes three gate information, namely the forget gate, the input gate, and the output gate. FIG9 is a schematic diagram of the structure of a long short-term memory neural network model according to an embodiment of the present disclosure. As shown in FIG9, the forget gate mainly controls the proportion of the data features of the previous layer of the network that are forgotten, that is, the input data features are weighted and summed with the previous sequence state, and input to the forget gate to obtain the data feature output of the forget gate. The calculation method is shown in formula (4): f t =σ(W xf ·x t +W hf ·h t-1 +b f ) Formula (4)

[0149] Among them, f t is the data feature output of the forget gate, W xf is the weighted coefficient of the forget gate for the current moment data, xt is the input data feature, W hf Represents the weighted coefficient of the forget gate to the state of the previous moment, h t-1 is the previous sequence state, b f is the forget gate data bias constant, and σ(x) represents the sigmoid activation function.

[0150] Furthermore, the main memory data at the current moment is obtained by combining the memory system data at the previous moment. The calculation method is shown in formula (5): t1 =c t-1 ⊙f t Formula (5)

[0151] Among them, c t1 is the main memory data at the current moment, c t-1 It is the memory system data of the previous moment.

[0152] In some embodiments, the input gate mainly filters the effective features of the current input data and compensates and updates the control data features of the forget gate in combination with the current data. The data feature extraction method of the input gate is shown in formula (6): t =σ(W xi ·x t +W hi ·h t-1 +b i ) Formula (6)

[0153] Among them, i t is the data feature output of the input gate, W xi is the weighted coefficient of the input gate for the current moment data, W hi is the weighted coefficient of the input gate to the state at the previous moment, b i is the input gate data bias constant.

[0154] Furthermore, the input data characteristics will compensate for the memory data characteristics to a certain extent, and the input gate characteristic parameters need to be screened. The calculation method of the screening control coefficient is shown in formula (7): c' t =tan h(W xc ·x t +W hc ·h t-1 +b c ) Formula (7)

[0155] Among them, c' t is the characteristic screening control coefficient of the input gate, W xc Filter the control weight coefficient for the current input data, Whc Input the filter weight coefficient for the data state at the previous moment, b c To filter the control data bias constant, tan h(x) is the activation function.

[0156] Furthermore, the input gate data is filtered and controlled to obtain the memory system data at the next moment. The calculation method is shown in formula (8): t2 =c' t ⊙i t Formula (8)

[0157] Among them, c t2 Memory system data for the next moment.

[0158] In some embodiments, the output value of the memory parameter at the current moment can be determined based on the main memory data at the current moment and the memory system data at the next moment. The calculation method is shown in formula (9): t =c t1 +c t2 Formula (9)

[0159] Among them, c t The output value of the memory parameter at the current moment.

[0160] In some embodiments, the output gate mainly learns the model and predicts the training set based on the forgotten data features and the current data features. The training method is shown in formula (10): t =σ(W xo ·x t +W ho ·h t-1 +b o ⊙tan h(c t )) Formula (10)

[0161] Among them, h t Output data for the output gate, W xo is the weight coefficient of the data at the current moment in the output gate, W ho is the weight coefficient of the state data at the previous moment in the output gate, b o is the data bias constant in the output gate.

[0162] In some embodiments, the present disclosure may employ a two-layer LSTM plus fully connected architecture. FIG10 is a schematic diagram of the architecture of a long short-term memory neural network model according to an embodiment of the present disclosure. Table 2 shows parameter configurations for a two-layer LSTM plus fully connected architecture, as shown in Table 2.

[0163] Table 2

[0164] Exemplarily, in combination with FIG10 and Table 2, the number of hidden layers of the LSTM may be 64 layers, the activation function may be ReLU, and the loss function may be a mean square error (MSE) function.

[0165] In some embodiments, after the artificial intelligence model is trained, the perception server sends the trained artificial intelligence model to the perception function entity.

[0166] It should be noted that based on the perception processing method provided by the embodiment of the present disclosure, the artificial intelligence model is trained through a perception server with high storage capacity and low computing power requirements, and the training efficiency is high; then, the perception data is processed based on the trained artificial intelligence model to make the obtained perception results more accurate, and the automatic processing of the perception data realizes the intelligent management of the perception data, reducing manpower investment; finally, the trained artificial intelligence model is sent to the perception function entity, so that the perception function entity can predict the perception data based on the trained artificial intelligence model, thereby realizing data sharing between various modules in the system.

[0167] In some embodiments, after the perception function entity receives the trained artificial intelligence model, the perception function entity preprocesses the actual test trajectory data (the preprocessing method is consistent with the preprocessing method when training the model), inputs the preprocessed data set into the artificial intelligence model, and obtains the predicted value (x t ,y t , z t , v t ), according to the scaler saved during data preprocessing, perform denormalization, and then add (Lo n , La n , H n , V n ), and get the final predicted value (Lo n+1 , La n+1 , H n+1 , V n+1 ).

[0168] In some embodiments, the perception function entity transmits the actual test trajectory data with higher confidence and the corresponding labels to the perception server for data collection.

[0169] In some embodiments, the perception function entity transmits the perception results with higher confidence and the corresponding labels to the perception server for data collection, and the perception server performs periodic model training on the artificial intelligence model according to the second indication information sent by the perception control function to indicate the training cycle.

[0170] In some embodiments, the perception processing function in the perception function entity sends response information of the perception service request message to the perception management function, and the perception management function sends response information of the perception service request message to the core network, and reports the perception results according to the perception service request message, thereby displaying a complete operation trajectory of the target.

[0171] In some embodiments, the above method can also be implemented by the process shown in Figure 11. Referring to Figure 11, a schematic flow diagram of another perception processing method according to an embodiment of the present disclosure is shown. As shown in Figure 11, taking the target as a drone as an example, the method includes the following steps S1 to S15.

[0172] In S1, the core network element sends the drone's perception service request message to the perception management function in the perception function entity.

[0173] The perception service request message is used to indicate the perception service demand, and the perception service request message may carry the perception service type, perception area, perception reporting content, and perception reporting type.

[0174] The perception service type is drone perception, the perception area can be within a radius of 1000 meters centered on the perception node or other ranges, the perception reporting content can include the position and speed of the drone, and the perception reporting type can be perception result reporting.

[0175] In S2, after the perception management function obtains the perception service request message, it authenticates the perception service request message.

[0176] In S3, the perception management function sends the perception service request message that passes authentication to the perception control function in the perception function entity.

[0177] In S4, after the perception control function obtains the perception service request message, it selects the perception method, perception node and perception data processing method for the drone according to the perception service request message.

[0178] In S5, the perception control function sends a perception control instruction to the perception node.

[0179] The sensing control instruction is used to instruct the sensing node to perform the sensing task. The sensing control instruction includes sensing configuration information, which is used to configure at least one of sensing resources, sensing time, and sensing signals.

[0180] In S6, the perception control function sends first instruction information for performing model training to the perception server.

[0181] In S7, after receiving the perception control instruction, the perception node sends a perception signal to the UAV according to the perception control instruction, performs the perception task, and obtains the perception data of the UAV.

[0182] In S8, the perception node sends the perception data of the UAV to the perception processing function of the perception function entity.

[0183] In S9, the perception processing function performs perception data preprocessing on the perception data.

[0184] Perception data preprocessing includes extracting feature data or normalizing the perception data.

[0185] In S10, after receiving the first instruction information for model training sent by the perception control function, the perception server trains the artificial intelligence model and sends the trained artificial intelligence model to the perception processing function.

[0186] It should be noted that for the training of the artificial intelligence model, please refer to S201 above and will not be repeated here.

[0187] In S11, the perception processing function performs artificial intelligence prediction on the artificial intelligence model to obtain predicted perception results.

[0188] In S12, the perception processing function sends the perception results to the perception server for storage.

[0189] In S13, the perception server periodically trains the artificial intelligence model based on the second indication information sent by the perception control function for indicating the training cycle and the perception result.

[0190] In S14, after obtaining the perception result, the perception processing function in the perception function entity sends a perception service request message response to the perception management function.

[0191] In S15, the perception management function sends a perception service request message response to the core network element, and the core network element obtains the perception result of the drone.

[0192] The perception results can include the position and speed of the drone, showing a complete motion trajectory of the drone.

[0193] The above mainly introduces the scheme of the embodiment of the present disclosure from the perspective of method. It can be understood that in order to realize the above functions, the perception processing device includes at least one of the hardware structure and software modules corresponding to the execution of each function. It should be easy for those skilled in the art to realize that, in combination with the units and algorithm steps of each example described in the embodiment disclosed herein, the embodiment of the present disclosure can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the embodiment of the present disclosure.

[0194] It is understandable that, in order to realize the above functions, the perception processing device includes hardware structures and / or software modules corresponding to the execution of each function. It should be easily appreciated by those skilled in the art that, in combination with the algorithm steps of each example described in the embodiments of the present disclosure, the present disclosure can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present disclosure.

[0195] The embodiment of the present disclosure can divide the functional modules of the perception processing device according to the above-mentioned method embodiment. For example, each functional module can be divided corresponding to each function, or two or more functions can be integrated into one functional module. The above-mentioned integrated module can be implemented in the form of hardware or software. It should be noted that the division of modules in the embodiment of the present disclosure is schematic and is only a logical function division. There may be other division methods in actual implementation. The following is an example of dividing each functional module corresponding to each function.

[0196] Figure 12 is a schematic diagram of the structure of a perception processing device according to an embodiment of the present disclosure. The perception processing device is applied to a perception functional entity and can execute the perception processing method provided in the above method embodiment. As shown in Figure 12, the perception processing device 200 includes an acquisition module 201, a processing module 202, and a sending module 203.

[0197] In some embodiments, the acquisition module 201 is used to obtain a trained artificial intelligence model from a perception server.

[0198] In some embodiments, the processing module 202 is used to process the perception data based on the artificial intelligence model to obtain a perception result.

[0199] In some embodiments, the sending module 203 is configured to send the sensing result to a core network element.

[0200] In some embodiments, the acquisition module 201 is further configured to receive a perception service request message sent by a core network element.

[0201] In some embodiments, the acquisition module 201 is further configured to select a sensing node to perform sensing detection in response to a sensing service request message, and to obtain sensing data obtained by the sensing node during the sensing detection process.

[0202] In some embodiments, the processing module 202 is further configured to determine a sensing task based on the sensing service request message.

[0203] In some embodiments, the sending module 203 is further configured to select a sensing node to perform a sensing task and send a sensing control instruction to the sensing node. The sensing control instruction is configured to instruct the sensing node to perform the sensing task.

[0204] In some embodiments, the sensing control instruction includes sensing configuration information, which is used to configure at least one of a sensing resource, a sensing time, and a sensing signal.

[0205] In some embodiments, the acquisition module 201 is further configured to receive perception data sent by a perception node.

[0206] In some embodiments, the processing module 202 is further configured to authenticate the perception service request message.

[0207] In some embodiments, the processing module 202 is further configured to determine a sensing task based on the sensing service request message after the sensing service request message is authenticated.

[0208] In some embodiments, the perception function entity is configured with at least one of the following functions: perception control function, perception management function, artificial intelligence function, perception processing function, and data preprocessing function.

[0209] In some embodiments, the perception control function includes at least one of the following: selection of perception nodes, management of perception resources, management of perception time, management of artificial intelligence applications, and management of information transmission.

[0210] In some embodiments, the perception management function includes at least one of the following: obtaining perception business needs, authenticating the perception business needs, and determining perception tasks based on the perception business needs.

[0211] In some embodiments, the artificial intelligence function includes at least one of the following: training an artificial intelligence model, and using an artificial intelligence model.

[0212] In some embodiments, the sending module 203 is further used to send at least one of the following to the perception server: first indication information for indicating whether the perception server performs model training, and second indication information for indicating a training cycle.

[0213] Figure 13 is a schematic diagram of the structure of another perception processing device according to an embodiment of the present disclosure. The perception processing device is applied to a perception server and can execute the perception processing method provided in the above method embodiment. As shown in Figure 13, the perception processing device 300 includes a training module 301, a sending module 302, and an acquisition module 303.

[0214] In some embodiments, the training module 301 is used to train an artificial intelligence model. The artificial intelligence model is used to process perception data to obtain perception results.

[0215] In some embodiments, the sending module 302 is used to send the trained artificial intelligence model to the perception function entity.

[0216] In some embodiments, the perception server is configured with at least one of the following functions: service control function, data preprocessing function, data storage function, and data opening function.

[0217] In some embodiments, the service control function includes at least one of the following: management of artificial intelligence applications, management of model training time, management of data openness functions, and management of information transmission.

[0218] In some embodiments, the data opening function is used to provide at least one of perception data, perception results, and artificial intelligence models to the outside world.

[0219] In some embodiments, the sending module 302 is further used to send at least one of the following to the core network element: perception data, artificial intelligence model.

[0220] In some embodiments, the acquisition module 303 is configured to receive perception data sent by a perception node.

[0221] In the case of implementing the functions of the above-mentioned integrated modules in hardware, the embodiments of the present disclosure provide a structure of the communication device involved in the above-mentioned embodiments. As shown in Figure 14, the communication device 400 includes: a processor 402 and a bus 404. In some embodiments, the communication device 400 may also include a memory 401. In some embodiments, the communication device 400 may also include a communication interface 403.

[0222] In some embodiments, the processor 402 may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the embodiments of the present disclosure. The processor 402 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array, or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof, which may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the embodiments of the present disclosure. The processor 402 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP (digital signal processor) and a microprocessor, and the like.

[0223] In some embodiments, the communication interface 403 is used to connect to other devices via a communication network, which may be Ethernet, wireless access network, wireless local area network (WLAN), etc.

[0224] In some embodiments, the memory 401 can be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.

[0225] As an implementation, memory 401 may exist independently of processor 402. Memory 401 may be connected to processor 402 via bus 404 and used to store instructions or program code. When processor 402 calls and executes the instructions or program code stored in memory 401, the perception processing method provided in the embodiments of the present disclosure can be implemented.

[0226] In another implementation, the memory 401 may also be integrated with the processor 402 .

[0227] In some embodiments, bus 404 may be an Extended Industry Standard Architecture (EISA) bus, etc. Bus 404 may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, FIG14 shows only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0228] Some embodiments of the present disclosure provide a computer-readable storage medium (e.g., a non-transitory computer-readable storage medium). The computer-readable storage medium stores computer program instructions, which, when executed on a computer, cause the computer to execute the perception processing method of any of the above embodiments.

[0229] Exemplarily, the above-mentioned computer-readable storage media may include, but are not limited to: magnetic storage devices (e.g., hard disks, floppy disks, or magnetic tapes), optical disks (e.g., compact disks (CDs), digital versatile disks (DVDs), etc.), smart cards, and flash memory devices (e.g., erasable programmable read-only memories (EPROMs), cards, sticks, or key drives, etc.). The various computer-readable storage media described in the present disclosure may represent one or more devices and / or other machine-readable storage media for storing information. The term "machine-readable storage medium" may include, but is not limited to, wireless channels and various other media capable of storing, containing, and / or carrying instructions and / or data.

[0230] An embodiment of the present disclosure provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the perception processing method of any one of the above embodiments.

[0231] The above are only specific embodiments of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any changes or replacements within the technical scope disclosed in the present disclosure should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure should be based on the scope of protection of the claims.

Claims

1. A perception processing method, applied to a perception functional entity, includes: Obtaining a trained artificial intelligence model from a perception server; Processing perception data based on the artificial intelligence model to obtain a perception result; Sending the perception result to a core network element.

2. The method according to claim 1, wherein, Before processing the perception data based on the artificial intelligence model to obtain the perception result, the method further includes: Receiving a perception service request message sent by the core network element; In response to the perception service request message, selecting a perception node for perception detection, and obtaining perception data obtained by the perception node during the perception detection process.

3. The method according to claim 2, wherein The responding to the perception service request message, selecting the perception node for the perception detection, and obtaining the perception data obtained by the perception node during the perception detection process includes: Determining a perception task based on the perception service request message; Selecting a perception node to execute the perception task, and sending a perception control instruction to the perception node, where the perception control instruction is used to instruct the perception node to execute the perception task; Receiving the perception data sent by the perception node.

4. The method according to claim 3, wherein The perception control instruction includes perception configuration information, and the perception configuration information is used to configure at least one of perception resources, perception time, and perception signals.

5. The method according to claim 3, wherein The determining the perception task based on the perception service request message includes: Authenticating the perception service request message; After the authentication of the perception service request message passes, determining a perception task based on the perception service request message.

6. The method according to claim 1, wherein, The perception functional entity is configured with at least one of the following functions: perception control function, perception management function, artificial intelligence function, perception processing function, data preprocessing function.

7. The method according to claim 6, wherein, The perception control function includes at least one of the following: selection of perception nodes, management of perception resources, management of perception time, management of artificial intelligence applications, management of information transmission.

8. The method according to claim 6, wherein The perception management function includes at least one of the following: obtaining perception service requirements, authenticating the perception service requirements, determining a perception task based on the perception service requirements.

9. The method according to claim 6, wherein The artificial intelligence function includes at least one of the following: training the artificial intelligence model, using the artificial intelligence model.

10. According to the method described in claim 1, it further includes: Sending at least one of the following to the core network element: a perception service request message, a perception service response message, the perception result; where the perception service request message is used to indicate perception service requirements.

11. The method according to claim 1 further comprises: Sending a perception control instruction to a perception node; the perception control instruction is used to select the perception node to execute a perception task.

12. The method according to claim 1 further comprises: Sending at least one of the following to the perception server: first indication information for indicating whether the perception server performs model training, second indication information for indicating a training period.

13. A perception processing method, applied to a perception server, includes: Training an artificial intelligence model, where the artificial intelligence model is used to process perception data to obtain a perception result; Sending the trained artificial intelligence model to a perception functional entity.

14. The method according to claim 13, wherein, The sensing server is configured with at least one of the following functions: service control function, data preprocessing function, data storage function, and data opening function.

15. The method according to claim 14, wherein, The service control function includes at least one of the following: management of artificial intelligence applications, management of model training time, management of the data opening function, and management of information transmission.

16. The method according to claim 14, wherein, The data opening function is used to externally provide at least one of the sensing data, the sensing results, and the artificial intelligence model.

17. The method according to claim 13 further includes: Sending at least one of the following to the core network element: the sensing data, the sensing results, and the artificial intelligence model.

18. The method according to claim 13 further includes: Receiving the sensing data sent by the sensing node.

19. A communication and sensing integrated system includes: A sensing server for training an artificial intelligence model; A sensing function entity for obtaining the trained artificial intelligence model from the sensing server; Processing the sensing data based on the artificial intelligence model to obtain sensing results, and sending the sensing results to the core network element.

20. The system according to claim 19 further includes: A sensing node for sending sensing signals and obtaining the sensing data; Sending the sensing data to the sensing function entity.

21. A communication device, comprising: A memory and a processor; The memory is coupled to the processor; The memory is used to store instructions executable by the processor; When the processor executes the instructions, it executes the method according to any one of claims 1-12, or the method according to any one of claims 13-18.

22. A computer-readable storage medium, wherein, Computer instructions are stored on the computer-readable storage medium, and when the computer instructions run on the computer, the computer is caused to execute the method according to any one of claims 1-12, or the method according to any one of claims 13-18.

Images