Method, apparatus and network node for optimization of a data line

Abstract

The application provides a data line optimization method, device and network node, and relates to the technical field of communication, and the method comprises the following steps: determining a data line according to a plurality of atomic capabilities; determining the communication state of the data line according to a heartbeat mechanism; optimizing the heartbeat mechanism corresponding to the data line according to the communication state of the communication line and a nonlinear optimization model; and the nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization mode. According to the application, the activity between the atomic capabilities in the MEC node can be tracked in time, that is, whether the atomic capabilities can normally communicate is found in time, and when the abnormal communication between the atomic capabilities is found to recover to normal communication in time, and then the collaborative management between the atomic capabilities is realized.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method, apparatus and network node for optimizing data lines. Background Technology

[0002] Multi-Access Edge Computing (MEC) refers to an open platform that integrates networking, computing, storage, and applications closer to the application or data source, providing near-end services. In 2016, the European Space Technology Institute (ESTI) expanded the concept of MEC to multi-access edge computing, extending edge computing from cellular networks to other access methods. Edge computing can provide faster network service response times, meeting the basic needs of industries in areas such as real-time business, application intelligence, security, and privacy protection.

[0003] Existing MEC technologies can be broadly categorized into three types: First, those focusing on MEC data offloading methods incorporate artificial intelligence algorithms (including rule matching, neural networks, clustering, etc.) to offload different types of request messages, improving the edge UPF's data processing capabilities, reducing data transmission latency, and enhancing network service efficiency. Second, those highlighting the application of MEC+5G technology in various industries, such as vehicle-to-everything (V2X), IoT, digital industry, and smart cities, leveraging its ultra-high mobile broadband, massive machine-type communication, and ultra-reliable low latency characteristics to create integrated offloading and service solutions tailored to local conditions. Third, those focusing on analyzing the network model within MEC to optimize the network and minimize information resource allocation costs.

[0004] However, current MEC deployment schemes do not consider how to track the activity of atomic capabilities in MEC nodes in a timely manner, which makes it impossible to achieve collaborative management between atomic capabilities. Summary of the Invention

[0005] This invention provides a data line optimization method, apparatus, and network node to solve the problem in the prior art that the activity of atomic capabilities in MEC nodes cannot be tracked in a timely manner.

[0006] To address the aforementioned technical problems, the embodiments of the present invention provide the following technical solutions:

[0007] This invention provides a method for optimizing data lines, comprising:

[0008] A data line is determined based on the capabilities of multiple atoms; the data line is formed by establishing communication connections between multiple capability atoms.

[0009] The communication status of the data line is determined based on the heartbeat mechanism; the communication status includes normal communication and abnormal communication.

[0010] Based on the communication status and nonlinear optimization model of the communication line, the heartbeat mechanism corresponding to the data line is optimized;

[0011] The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method.

[0012] Optionally, the preset optimization method includes:

[0013] For data lines with normal communication, the initial heartbeat interval is increased by a first preset interval;

[0014] The initial weight value of the data line with communication failure is reduced by a first threshold, so that the initial heartbeat interval of the data line with communication failure is reduced by a second preset interval.

[0015] Optionally, before determining the data line based on multiple atomic capabilities, the method further includes:

[0016] Based on vehicle-to-everything (V2X) application data, several atomic capabilities are identified; these atomic capabilities include: data fusion atomic capability, on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0017] Optionally, the method further includes:

[0018] For application scenarios involving multiple access edge computing nodes corresponding to the data lines, the corresponding atomic capabilities are invoked.

[0019] Optionally, the method further includes:

[0020] The nonlinear optimization model is constructed based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line.

[0021] Wherein, the sum of the weight values ​​of the data sources corresponding to the data lines is 1;

[0022] The weight value of the data source is directly proportional to the heartbeat interval of the data line.

[0023] Optionally, before optimizing the heartbeat mechanism corresponding to the data line based on the communication state and nonlinear optimization model of the communication line, the method further includes:

[0024] Set an initial heartbeat interval on each data line;

[0025] The data line that receives feedback information within the initial heartbeat interval is identified as the data line with normal communication.

[0026] The data line that did not receive feedback information within the initial heartbeat interval is identified as the data line with the communication failure.

[0027] Optionally, after the nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method, the method further includes:

[0028] The heartbeat interval of normally communicating data lines with an increased first preset interval, and the weight values ​​of normally communicating data lines with a decreased first threshold, are input into the nonlinear optimization model for the next round of calculation.

[0029] Optionally, the weight values ​​of the communication anomalies that reduce the first threshold are input into the nonlinear optimization model for the next round of calculation, including:

[0030] When the utilization rate of the CPU of the central processing unit of the multi-access edge computing node corresponding to the data line is greater than the first preset utilization rate, the initial heartbeat interval of the data line with communication abnormality is input into the nonlinear optimization model for the next round of calculation.

[0031] Optionally, for application scenarios involving multiple access edge computing nodes corresponding to the data line, the corresponding atomic capabilities can be invoked, including:

[0032] When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is less than the first duration, the accuracy requirement is greater than the second threshold, and the recall value is greater than the third threshold, the first atomic capability in the atomic capabilities is invoked based on the heartbeat feedback mechanism.

[0033] When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is greater than or equal to the first duration, the accuracy requirement is less than or equal to the second threshold, and the recall requirement is less than or equal to the third threshold, the first atomic capability and / or the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism.

[0034] When the multi-access edge computing node corresponding to the data line is used to connect with other multi-access edge computing nodes, the first atomic capability and the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism.

[0035] The first atomic capability includes at least one of the following: on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0036] Optionally, for application scenarios involving multiple access edge computing nodes corresponding to the data line, before invoking the corresponding atomic capabilities, the method further includes:

[0037] Based on the decision tree classification model, the application scenarios of the multi-access edge computing nodes corresponding to the data lines are determined.

[0038] This invention also provides a data line optimization device, comprising:

[0039] The first determining module is used to determine a data line based on multiple atomic capabilities; the data line is formed by establishing communication connections among multiple capability atoms.

[0040] The second determining module is used to determine the communication status of the data line based on the heartbeat mechanism; the communication status includes normal communication and abnormal communication.

[0041] An optimization module is used to optimize the heartbeat mechanism corresponding to the data line based on the communication status and nonlinear optimization model of the communication line.

[0042] The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method.

[0043] Optionally, the preset optimization method includes:

[0044] For data lines with normal communication, the initial heartbeat interval is increased by a first preset interval;

[0045] The initial weight value of the data line with communication failure is reduced by a first threshold, so that the initial heartbeat interval of the data line with communication failure is reduced by a second preset interval.

[0046] Optionally, the device further includes:

[0047] The third determining module is used to determine multiple atomic capabilities based on vehicle network application data; the multiple atomic capabilities include: data fusion atomic capability, on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0048] Optionally, the device further includes:

[0049] The calling module is used to invoke the corresponding atomic capabilities for application scenarios involving multiple access edge computing nodes corresponding to the data lines.

[0050] Optionally, the device further includes:

[0051] A module is established to construct the nonlinear optimization model based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line.

[0052] Wherein, the sum of the weight values ​​of the data sources corresponding to the data lines is 1;

[0053] The weight value of the data source is directly proportional to the heartbeat interval of the data line.

[0054] Optionally, the device further includes:

[0055] The settings module is used to set the initial heartbeat interval on each data line;

[0056] The fourth determining module is used to determine that the data line that receives feedback information within the initial heartbeat interval is the data line with normal communication.

[0057] The fifth determining module is used to determine that the data line that has not received feedback information within the initial heartbeat interval is the data line with communication abnormality.

[0058] Optionally, the device further includes:

[0059] The processing module is used to input the heartbeat interval of the normally communicating data lines with an increased first preset interval, and the weight values ​​of the abnormally communicating data lines with a decreased first threshold, into the nonlinear optimization model for the next round of calculation.

[0060] Optionally, the processing module includes:

[0061] The processing unit is configured to input the initial heartbeat interval of the communication-abnormal data line into the nonlinear optimization model for the next round of calculation when the utilization rate of the CPU of the central processing unit of the multi-access edge computing node corresponding to the data line is greater than a first preset utilization rate.

[0062] Optionally, the calling module includes:

[0063] The first invocation unit is used to invoke the first atomic capability among the atomic capabilities based on the heartbeat feedback mechanism when the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is less than the first duration, the accuracy requirement is greater than the second threshold, and the recall value is greater than the third threshold.

[0064] The second calling unit is used to call the first atomic capability and / or data fusion atomic capability in the atomic capabilities based on the heartbeat feedback mechanism when the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is greater than or equal to the first duration, the accuracy requirement is less than or equal to the second threshold, and the recall value is less than or equal to the third threshold.

[0065] The third calling unit is used to call the first atomic capability and the data fusion atomic capability in the atomic capabilities based on the heartbeat feedback mechanism when the multi-access edge computing node corresponding to the data line is used to connect with other multi-access edge computing nodes.

[0066] The first atomic capability includes at least one of the following: on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0067] Optionally, the device further includes:

[0068] The sixth determination module is used to determine the application scenario of the multi-access edge computing node corresponding to the data line based on the decision tree classification model.

[0069] This invention also provides a network node, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the data line optimization method as described in any one of the above.

[0070] This invention also provides a readable storage medium storing a program that, when executed by a processor, implements the data line optimization method described in any of the above embodiments.

[0071] The beneficial effects of this invention are:

[0072] The present invention configures a heartbeat mechanism on the data lines between multiple atomic capabilities. Based on the heartbeat mechanism, the communication status of the data lines is determined. Furthermore, based on the communication status of the communication lines and a nonlinear optimization model, the heartbeat mechanism corresponding to the data lines is optimized. This enables timely tracking of the activity between atomic capabilities in MEC nodes, that is, timely detection of whether atomic capabilities can communicate normally, and timely detection of when atomic capabilities with abnormal communication will resume normal communication, thereby achieving collaborative management between atomic capabilities. Attached Figure Description

[0073] Figure 1 A flowchart illustrating the data line optimization method provided in an embodiment of the present invention;

[0074] Figure 2This diagram illustrates the architecture of the MEC node provided in an embodiment of the present invention.

[0075] Figure 3 This describes the specific optimization process of the heartbeat mechanism on the data line provided in the embodiments of the present invention;

[0076] Figure 4 This is a schematic diagram illustrating the structure of the decision tree classification model provided in an embodiment of the present invention;

[0077] Figure 5 This is a schematic diagram illustrating the structure of the data line optimization device provided in an embodiment of the present invention;

[0078] Figure 6 This is a schematic diagram illustrating the structure of a network node provided in an embodiment of the present invention. Detailed Implementation

[0079] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0080] This invention addresses the problem in existing technologies that cannot track the activity of atomic capabilities in MEC nodes in a timely manner, by providing a method, apparatus, and network node for optimizing data lines.

[0081] First, let's explain some concepts.

[0082] Heartbeat mechanism

[0083] In a communication network, nodes need to check each other's "liveness status" to ensure normal communication. This mechanism for checking "liveness status" is called the heartbeat mechanism. Like a heartbeat, it sends a heartbeat packet periodically. First, node A sends a heartbeat packet to another network node B. When node B receives the heartbeat packet from node A, it knows that node A is still alive and ready to communicate. After receiving the heartbeat packet, node B sends a corresponding heartbeat packet back to node A. When node A receives the reply heartbeat packet, it can also confirm that node B is currently "alive" and network communication is normal.

[0084] Vehicle-to-Everything (V2X) communication technology standards

[0085] V2X refers to the use of sensors, onboard terminals, and electronic tags mounted on vehicles to provide vehicle data and information. It leverages advanced communication technologies to acquire information about the vehicle's surrounding environment, thereby enabling vehicle control and service provision. Specifically, V2X encompasses communication scenarios such as vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), vehicle-to-person (V2P), and vehicle-to-network (V2N).

[0086] Currently, V2X communication technology standards mainly include Dedicated Short Range Communication (DSRC) technology and Cellular Vehicle to Everything (C-V2X) based on cellular mobile communication systems. C-V2X includes Long Term Evolution (LTE-V2X) and 5G New Radio (5G NR-V2X) technologies. LTE-V2X, as a comprehensive communication solution for vehicle-to-everything (V2X) communication, can provide low latency, high reliability, and high speed communication capabilities in high-speed mobile environments, meeting the needs of various V2X applications.

[0087] like Figure 1 As shown, this embodiment of the invention provides a method for optimizing data lines, characterized by comprising:

[0088] Step 101: Determine the data line based on the capabilities of multiple atoms; the data line is formed by establishing communication connections among multiple capability atoms.

[0089] It should be noted that, as Figure 2 As shown, this embodiment of the invention provides an MEC node, which is deployed based on multi-source traffic data. The MEC node includes a Layer I for accessing various applications of the Internet of Vehicles, such as traffic incident early warning, real-time adjustment of navigation routes, and vehicle information early warning; a Layer II for data fusion atomic capabilities; and a Layer III for Onboard Unit (OBU) data parsing atomic capabilities, camera recognition algorithm atomic capabilities, radar data parsing atomic capabilities, and data communication atomic capabilities.

[0090] Layer I, Layer II, and Layer III constitute the main components of the aggregation layer MEC server. Layer I adds most of the capabilities of the access layer servers. Of course, it is also possible to have access layer MEC server capabilities in addition to Layer I and Layer II. Whether to integrate at this layer depends mainly on resource configuration and business needs. The deployment and separation of atomic capabilities are quite flexible. The combination of N+ access edge cloud and the combination of access layer MEC edge cloud + aggregation MEC edge cloud can both be managed uniformly through clustering.

[0091] Please continue reading. Figure 2 This establishes connections between atomic capabilities, forming multiple data lines.

[0092] Step 102: Determine the communication status of the data line according to the heartbeat mechanism; the communication status includes normal communication and abnormal communication.

[0093] In this embodiment of the invention, a heartbeat mechanism is configured on each data line. According to the heartbeat mechanism, if an atomic capability A can send a heartbeat packet to another atomic capability B, and atomic capability B receives the heartbeat packet and returns a corresponding heartbeat packet to atomic capability A, then it is determined that atomic capability A and atomic capability B can communicate normally. Otherwise, the data line communication between atomic capability A and atomic capability B is abnormal.

[0094] Step 103: Optimize the heartbeat mechanism corresponding to the data line based on the communication status and nonlinear optimization model of the communication line;

[0095] The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method.

[0096] In this embodiment of the invention, based on the communication status of the data line, the heartbeat mechanism on the data line is optimized by constructing a nonlinear optimization model. This enables timely tracking of the activity between atomic capabilities in the MEC node, that is, timely detection of whether atomic capabilities can communicate normally, and timely detection of when atomic capabilities with abnormal communication will resume normal communication, thereby achieving collaborative management between atomic capabilities.

[0097] Optionally, the preset optimization method includes:

[0098] For data lines with normal communication, the initial heartbeat interval is increased by a first preset interval;

[0099] The initial weight value of the data line with communication failure is reduced by a first threshold, so that the initial heartbeat interval of the data line with communication failure is reduced by a second preset interval.

[0100] In this embodiment of the invention, for data lines with normal communication, increasing the original heartbeat interval by a first preset interval can minimize resource consumption. For data lines with abnormal communication, decreasing the initial weight value by a first threshold can reduce the initial heartbeat interval of the abnormal data lines by a second preset interval. This enables timely tracking of the activity of atoms, that is, reducing the heartbeat interval when a data line is abnormal, making it easier to detect abnormal data lines, and determining when normal communication is restored between the atomic capabilities of the data lines with abnormal communication.

[0101] Optionally, before determining the data line based on multiple atomic capabilities, the method further includes:

[0102] Based on vehicle-to-everything (V2X) application data, several atomic capabilities are identified; these atomic capabilities include: data fusion atomic capability, on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0103] This invention, based on multi-source traffic data, introduces the concept of "vehicle-road cooperation," making it more promising for application.

[0104] In this embodiment of the invention, multiple atomic capabilities are determined based on multi-source traffic data (vehicle network application data) from the LTE-V2X communication technology standard.

[0105] In the context of the Internet of Vehicles (IoV), the uniqueness of each data source corresponds to specific atomic capabilities. These atomic capabilities within the MEC server are independent of each other, thus forming reusable and portable atomic capabilities: data communication atomic capabilities, OBU data parsing atomic capabilities, camera recognition algorithm atomic capabilities, radar data parsing atomic capabilities, and data fusion atomic capabilities. Each atomic capability is independent, reducing coupling and better mitigating the service downtime caused by the unavailability of single-line source data. This allows for the construction of a more flexible and efficient MEC management system under different business scenarios and resource configurations.

[0106] A detailed explanation of the function of each atom:

[0107] Data communication atomic capabilities: Based on 5G User Plane Function (UPF) devices, widely used data transmission protocols are defined, such as Message Queuing Telemetry Transport (MQTT). MQTT is suitable for low-bandwidth environments and unreliable network conditions, especially for mobile in-vehicle devices, i.e., OBU data access. Advanced Message Queuing Protocol (AMQP) is more reliable and adaptable to more complex communication models, differing from the simple publish-subscribe model. RabbitMQ and HornetQ are popular middleware implementations of this protocol. Specifically, MQTT and RabbitMQ middleware are deployed as containers on MEC nodes, mapping different communication ports to ensure that communication channels are independent and do not interfere with each other.

[0108] OBU data parsing atomic capability: Through the ASN.1 compiler, it parses and constructs structured data based on the LTE-V2X communication technology standard. Currently, the ASN.1 compiler supports C, C++, and Java languages. Specifically, it involves installing the ASN.1 plugin, creating an abstract model corresponding to the vehicle-side data parsing protocol, using the ASN (C / C++ / Java) compiler to generate model files in the corresponding languages, and performing encoding and decoding based on these classes to output structured data based on the LTE-V2X communication technology standard.

[0109] The camera recognition algorithm's atomic capabilities include: subscribing to traffic video in formats such as Real-Time Streaming Protocol (RTSP) video streams to achieve structured output of traffic elements. First, target detection technology is used to quickly and accurately locate vehicles. Then, based on vehicle location, its visual features are fully extracted to identify inherent vehicle attributes, generating structured labels for the vehicles (category, license plate, size, etc.). Finally, based on these structured labels, retrieval and re-identification technologies are combined to achieve the retrieval, location, and re-identification of specific vehicles within massive amounts of video data.

[0110] Radar data parsing capabilities: Utilizing algorithms built into the radar manufacturer, deep learning algorithms (such as YOLO) based on raw point clouds are used to detect traffic participants perceived by the radar, outputting corresponding structured perception data. YOLO treats object detection as a regression problem; after one inference step, the input image yields the location, category, and corresponding confidence probability of all objects in the image, completing the process from inputting the raw image to outputting object locations and categories.

[0111] Data fusion atomic capability: This primarily employs high-order tensors to capture high-dimensional features of heterogeneous data, achieving fusion output of multi-source heterogeneous data. First, structured data from multiple sources is obtained using timestamps, defined as M types of heterogeneous data. A corresponding feature extraction module F, a feature fusion module I, and a feature decision matrix C are designed, and corresponding sub-networks are obtained for each. <F1,F2,……,F M In feature fusion module I, a high-order tensor R with heterogeneous data space dimensionality is constructed. R1 *R2* … *RM After training, the parameters of the tensor expanded along the i-th module can reflect the spatial dimensionality characteristics of the i-th type of heterogeneous data. In the feature decision module C, the definition of the feature decision module is determined according to the needs of the final business layer. By training on the features of the fused heterogeneous data, the potential relationships between heterogeneous data are explored more deeply, improving the feature expressiveness of the model on multi-source heterogeneous data, and finally outputting the fusion result.

[0112] It should be noted that the aforementioned atomic capabilities exist independently, enabling the establishment of heartbeat communication between them and the flexible combination of individual atomic capabilities.

[0113] Optionally, the method further includes:

[0114] For application scenarios involving multiple access edge computing nodes corresponding to the data lines, the corresponding atomic capabilities are invoked.

[0115] In this embodiment of the invention, multiple atomic capabilities on the MEC node are combined and deployed based on the heartbeat mechanism on the data line, which can realize flexible combination between atomic capabilities and adapt to more business scenarios. That is, for different application scenarios, the corresponding atomic capabilities can be called.

[0116] Optionally, the method further includes:

[0117] The nonlinear optimization model is constructed based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line.

[0118] Wherein, the sum of the weight values ​​of the data sources corresponding to the data lines is 1;

[0119] The weight value of the data source is directly proportional to the heartbeat interval of the data line.

[0120] In this embodiment of the invention, based on the multi-atom capability, the effectiveness and accuracy of the fusion result depend on the stable and efficient transmission of multi-source data. This embodiment of the invention constructs a nonlinear optimization model to perform regression calculation. The purpose of the regression is to maximize the evaluation index of the fused data and achieve the optimal solution for the final fused data quality.

[0121] Specifically, the constructed nonlinear optimization model is as follows:

[0122]

[0123]

[0124] Where i represents the data source of the data line, 1≤i≤N, w i p represents the weight values ​​of each data source in the fusion process. ik The value of evaluation metric k for data source i is expressed, where heartbeat represents the heartbeat interval, K is the number of evaluation metrics, and N is the number of data lines.

[0125] This embodiment uses a combination of regression and summation algorithms, and employs accuracy, recall, mean absolute variance, and coverage to evaluate the final results. Two constraints are set in the nonlinear optimization model: constraint (1) sums the data source weights to 1, which can achieve a weight center shift; constraint (2) represents a direct correlation between the weight value and the heartbeat interval. A larger weight value indicates normal communication of the data source, forming positive feedback and increasing the heartbeat interval; a smaller weight indicates abnormal communication of the data source, forming negative feedback and decreasing the heartbeat interval, thus enabling earlier detection of the data source's recovery results.

[0126] Optionally, before optimizing the heartbeat mechanism corresponding to the data line based on the communication state and nonlinear optimization model of the communication line, the method further includes:

[0127] Set an initial heartbeat interval on each data line;

[0128] The data line that receives feedback information within the initial heartbeat interval is identified as the data line with normal communication.

[0129] The data line that did not receive feedback information within the initial heartbeat interval is identified as the data line with the communication failure.

[0130] In this embodiment of the invention, in order to maintain normal communication between atomic capabilities, it is assumed that there are a total of N data lines, and each data line has a data source i (1≤i≤N). A heartbeat mechanism is established at the output end of each data line. First, the original heartbeat interval (e.g., 10ms) is input on each data line. That is, every 10ms, a connection request is sent to the atomic capability of the next receiving point. Receiving a "connection successful" feedback indicates normal communication. If no feedback is received within 10ms, it indicates a connection timeout and is judged as a communication anomaly.

[0131] Correspondingly, assuming there are n data lines maintaining normal communication (i.e., n lines generating feedback), the heartbeat interval is dynamically increased without affecting normal communication. That is, the initial heartbeat interval for normally communicating data lines is increased by a first preset interval. Conversely, assuming there are m data lines with abnormal communication (i.e., m lines not generating feedback), the weight of the abnormal data lines is appropriately reduced, shifting the larger weights to other surviving data sources (normally communicating data lines). In other words, the initial weight of the abnormal data lines is reduced by a first threshold, thus reducing the initial heartbeat interval of the abnormal data lines by a second preset interval.

[0132] Optionally, after the nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method, the method further includes:

[0133] The heartbeat interval of normally communicating data lines with an increased first preset interval, and the weight values ​​of normally communicating data lines with a decreased first threshold, are input into the nonlinear optimization model for the next round of calculation.

[0134] In this embodiment of the invention, after optimizing the data line using a nonlinear optimization model, the optimization results of each round are retained. For data sources with communication anomalies, the weight results of the previous round are applied. The operation decreases the interval of a∈{1,2,3…,N}, increases the heartbeat interval of normally communicating data lines, and uses this interval as an input parameter to input the nonlinear optimization model for the next round of regression calculation.

[0135] Optionally, the weight values ​​of the communication anomalies that reduce the first threshold are input into the nonlinear optimization model for the next round of calculation, including:

[0136] When the utilization rate of the CPU of the central processing unit of the multi-access edge computing node corresponding to the data line is greater than the first preset utilization rate, the initial heartbeat interval of the data line with communication abnormality is input into the nonlinear optimization model for the next round of calculation.

[0137] It should be noted that while reducing the heartbeat interval, a threshold for server resource utilization is set. For example, if the CPU utilization rate of the server reaches above the first preset utilization rate (such as 90%), the heartbeat interval must be increased to prevent the server from crashing.

[0138] That is, for a certain line where the server resource utilization rate reaches the first preset utilization rate, the heartbeat of the previous round is also retained and used as an input parameter to participate in the regression calculation of the next round.

[0139] It should also be noted that in the case where the data sources on all data lines become inactive, a data emergency plan is adopted to temporarily maintain the data activity for the first preset duration.

[0140] This data emergency plan is mainly enabled in the data fusion atom and mainly relies on a prediction model that is continuously optimized by the multi-data fusion results stably output in the early stage. The relatively popular RNN_LSTM model can be used.

[0141] The following combines Figure 3 , and specifically describes the process of optimizing the heartbeat mechanism on the data line using a non-linear optimization model.

[0142] An initial heartbeat interval is set on each data line, and the values of the initial heartbeat intervals set on each data line are the same. The heartbeat interval is adjusted according to the service resource utilization rate where the atom ability is located. For example, when the service resource utilization rate (CPU utilization rate) where the atom ability is located is greater than the first preset utilization rate, the heartbeat interval is increased. It is judged whether the heartbeat feedback mechanism of the data source on data line n is normal. In the case of normal data line communication, a positive feedback is generated and the heartbeat interval is appropriately increased. In the case of abnormal communication (abnormal feedback), it is judged whether n>1 and n<N are satisfied, that is, whether not all data lines are abnormally communicating. If not all data lines are abnormally communicating, the weight value of the data line with abnormal communication is appropriately reduced, and the large weight is shifted to other surviving data sources (data lines with normal communication). That is, a negative feedback is generated and the heartbeat interval is appropriately reduced, and the next round of optimization is performed according to the increased heartbeat interval and the reduced heartbeat interval.

[0143] Optionally, for the application scenario of the multi-access edge computing node corresponding to the data line, the corresponding atom ability is called, including:

[0144] When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is less than the first duration, the accuracy value requirement is greater than the second threshold, and the recall rate value is greater than the third threshold, based on the heartbeat feedback mechanism, the first atom ability in the atom ability is called;

[0145] When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is greater than or equal to the first duration, the accuracy requirement is less than or equal to the second threshold, and the recall requirement is less than or equal to the third threshold, the first atomic capability and / or the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism.

[0146] When the multi-access edge computing node corresponding to the data line is used to connect with other multi-access edge computing nodes, the first atomic capability and the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism.

[0147] The first atomic capability includes at least one of the following: on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0148] It's worth noting that the most common MEC node deployment involves attaching a general-purpose x86 server to the base station to handle urgent and high-volume processing tasks that need to be handled within the base station. Taking an autonomous vehicle as an example, after the video signal captured by the vehicle's camera is received by the base station, it is processed directly on the MEC server, and adjustments are made to the driving situation according to rules. Meanwhile, the remote server performs tasks such as remote driving and rule updates.

[0149] The MEC node deployment scheme provided in this embodiment of the invention is mainly aimed at regional coverage. In the context of the above multi-source traffic data (vehicle network application data), according to the v2x protocol, the access layer MEC server, the local layer MEC server and the aggregation layer MEC server form a wireless communication network with full regional coverage.

[0150] For application scenarios involving MEC node deployment, the main scenarios for invoking the corresponding atomic capabilities include the following three types:

[0151] For application scenarios where latency requirements are less than the first threshold, accuracy requirements are greater than the second threshold, and recall requirements are greater than the third threshold—that is, services with high latency and reliability requirements—urgent and high-volume processing tasks need to be handled locally. This requires calling the Application Programming Interface (API) deployed in the local layer or access layer MEC (calling at least one of the following: On-Board Unit (OBU) data parsing atomic capabilities, camera recognition algorithm capabilities, radar data parsing capabilities, and data communication atomic capabilities) to complete perception and computation. For example, in an application scenario where a vehicle is passing through a signal-controlled intersection (lane) and the traffic light is about to turn red or is already red, but the vehicle fails to stop within the stop line and continues to move forward, the roadside sensing device detects a risk of the vehicle not following signal regulations or instructions and can issue a warning to the vehicle and simultaneously warn other vehicles to prevent a collision.

[0152] For application scenarios where latency requirements are greater than or equal to the first time interval, accuracy requirements are less than or equal to the second threshold, and recall requirements are less than or equal to the third threshold—that is, services with low latency and reliability requirements—routing management will invoke the corresponding APIs deployed in the aggregation layer MEC (i.e., calling at least one of the following atomic capabilities: On-Board Unit (OBU) data parsing, camera recognition algorithm, radar data parsing, data communication, and / or data fusion) to complete the service process. For example, in a scenario where a connected vehicle has subscribed to an intersection collision warning service, it will receive all information about traffic participants at the intersection broadcast by the MEC via the RSU; after receiving this broadcast information, the connected vehicle will make decisions and controls based on its own vehicle location and status information. For traffic flow information subscription broadcasts, the aggregation layer MEC will retrieve historical video data files, using the same time period each day (e.g., 17:30-18:00) + the same road segment as the filtering keywords, i.e., the constraint condition (WHERE time between ** and *** AND location like “%XX intersection%”). A multi-object detection + multi-object tracking algorithm will be used to count the traffic flow under this constraint condition. If the flow exceeds a certain threshold (count(*) > number), a congestion alert will be broadcast, suggesting subscribers plan their routes accordingly. At the same time, statistics will also be compiled based on different vehicle types (group by type). For example, if the number of trucks exceeds a certain threshold (count(truck) > number), a broadcast will indicate that the road segment is a high-truck zone, reminding subscribers to drive cautiously.

[0153] In the context of MEC nodes being used to connect with other multi-access edge computing nodes, i.e., large-scale multi-MEC node deployment schemes, to achieve real-time response of vehicle-to-everything (V2X) services, this invention implements "collaborative management of the edge cloud" (by calling at least one of the atomic capabilities of On-Board Unit (OBU) data parsing, camera recognition algorithm capabilities, radar data parsing capabilities, data communication atomic capabilities, and data fusion atomic capabilities) to achieve regional global scheduling optimization. For example, in application scenarios supporting traffic management big data platforms, not only is multi-source raw data such as vehicle-side road segments required, but also data calculated and analyzed by each MEC, as well as historical data from core cloud testing, are needed. This enables the provision of more real-time and remote dynamic traffic information to travelers, such as roadside traffic videos, congestion alerts, proactive avoidance information, tidal lane prompts, and road construction alerts, achieving real-time information interaction and targeted information push between the platform and individuals. It also promotes the transformation of urban traffic management strategies and algorithms from aggregated analysis modeling oriented towards groups to disaggregated analysis modeling oriented towards vehicles.

[0154] According to the embodiments of the present invention, the atomic capabilities in MEC are classified according to the needs of transportation business, so as to realize combined deployment, which is more flexible and adaptable to more business scenarios.

[0155] Optionally, for application scenarios involving multiple access edge computing nodes corresponding to the data line, before invoking the corresponding atomic capabilities, the method further includes:

[0156] Based on the decision tree classification model, the application scenarios of the multi-access edge computing nodes corresponding to the data lines are determined.

[0157] In this embodiment of the invention, the application scenarios of MEC nodes are determined based on a decision tree classification model. When determining the application scenarios, the main parameters considered are latency, accuracy, and recall, using methods such as... Figure 4 The decision tree classification model shown classifies and determines the most suitable MEC deployment mode (access layer MEC, aggregation layer MEC). Furthermore, this model can be extended to more feature variables or feature parameters to adapt to more business needs.

[0158] According to the embodiments of the present invention, capabilities can be atomically split and combined based on business needs and resource configuration, and embedded and processed in different MEC node service schemes.

[0159] In this embodiment of the invention, the traffic business requirements are used as feature parameters input to the decision tree model to achieve automated selection and classification of the atomic capabilities of MEC, enabling combined deployment, which is more flexible and adaptable to more business scenarios.

[0160] In this embodiment of the invention, a heartbeat mechanism is introduced on the data lines between atomic capabilities to address the characteristics of multi-source traffic data. An adaptive heartbeat mechanism based on (positive and negative) feedback is proposed, which can realize reliable long-term connections of atomic capabilities. For lines with abnormal communication, a nonlinear optimization model is constructed with the goal of maximizing the evaluation index of the fused data, and an emergency handling scheme for shifting large weights to other surviving sources is realized.

[0161] like Figure 5 As shown, an embodiment of the present invention provides a data line optimization device, comprising:

[0162] The first determining module 501 is used to determine a data line based on multiple atomic capabilities; the data line is formed by establishing communication connections among multiple capability atoms.

[0163] The second determining module 502 is used to determine the communication status of the data line according to the heartbeat mechanism; the communication status includes normal communication and abnormal communication.

[0164] The optimization module 503 is used to optimize the heartbeat mechanism corresponding to the data line based on the communication status and nonlinear optimization model of the communication line.

[0165] The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method.

[0166] In this embodiment of the invention, a heartbeat mechanism is configured on the data lines between multiple atomic capabilities. Based on the heartbeat mechanism, the communication status of the data lines is determined. Furthermore, based on the communication status of the communication lines and a nonlinear optimization model, the heartbeat mechanism corresponding to the data lines is optimized. This enables timely tracking of the activity between atomic capabilities in MEC nodes, that is, timely detection of whether atomic capabilities can communicate normally, and timely detection of when atomic capabilities with abnormal communication will resume normal communication, thereby achieving collaborative management between atomic capabilities.

[0167] Optionally, the preset optimization method includes:

[0168] For data lines with normal communication, the initial heartbeat interval is increased by a first preset interval;

[0169] The initial weight value of the data line with communication failure is reduced by a first threshold, so that the initial heartbeat interval of the data line with communication failure is reduced by a second preset interval.

[0170] Optionally, the device further includes:

[0171] The third determining module is used to determine multiple atomic capabilities based on vehicle network application data; the multiple atomic capabilities include: data fusion atomic capability, on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0172] Optionally, the device further includes:

[0173] The calling module is used to invoke the corresponding atomic capabilities for application scenarios involving multiple access edge computing nodes corresponding to the data lines.

[0174] Optionally, the device further includes:

[0175] A module is established to construct the nonlinear optimization model based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line.

[0176] Wherein, the sum of the weight values ​​of the data sources corresponding to the data lines is 1;

[0177] The weight value of the data source is directly proportional to the heartbeat interval of the data line.

[0178] Optionally, the device further includes:

[0179] The settings module is used to set the initial heartbeat interval on each data line;

[0180] The fourth determining module is used to determine that the data line that receives feedback information within the initial heartbeat interval is the data line with normal communication.

[0181] The fifth determining module is used to determine that the data line that has not received feedback information within the initial heartbeat interval is the data line with communication abnormality.

[0182] Optionally, the device further includes:

[0183] The processing module is used to input the heartbeat interval of the normally communicating data lines with an increased first preset interval, and the weight values ​​of the abnormally communicating data lines with a decreased first threshold, into the nonlinear optimization model for the next round of calculation.

[0184] Optionally, the processing module includes:

[0185] The processing unit is configured to input the initial heartbeat interval of the communication-abnormal data line into the nonlinear optimization model for the next round of calculation when the utilization rate of the CPU of the central processing unit of the multi-access edge computing node corresponding to the data line is greater than a first preset utilization rate.

[0186] Optionally, the calling module includes:

[0187] The first invocation unit is used to invoke the first atomic capability among the atomic capabilities based on the heartbeat feedback mechanism when the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is less than the first duration, the accuracy requirement is greater than the second threshold, and the recall value is greater than the third threshold.

[0188] The second calling unit is used to call the first atomic capability and / or data fusion atomic capability in the atomic capabilities based on the heartbeat feedback mechanism when the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is greater than or equal to the first duration, the accuracy requirement is less than or equal to the second threshold, and the recall value is less than or equal to the third threshold.

[0189] The third calling unit is used to call the first atomic capability and the data fusion atomic capability in the atomic capabilities based on the heartbeat feedback mechanism when the multi-access edge computing node corresponding to the data line is used to connect with other multi-access edge computing nodes.

[0190] The first atomic capability includes at least one of the following: on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0191] Optionally, the device further includes:

[0192] The sixth determination module is used to determine the application scenario of the multi-access edge computing node corresponding to the data line based on the decision tree classification model.

[0193] It should be noted that the data line optimization device provided in the embodiments of the present invention is a device capable of executing the above-described data line optimization method. Therefore, all embodiments of the above-described data line optimization method are applicable to this device and can achieve the same or similar technical effects.

[0194] like Figure 6 As shown, this embodiment of the invention also provides a network node, including a processor 600, a transceiver 610, a memory 620, and a program stored in the memory 620 and executable on the processor 600; wherein the transceiver 610 is connected to the processor 600 and the memory 620 via a bus interface, and the processor 600 performs the following process to read the program from the memory:

[0195] A data line is determined based on the capabilities of multiple atoms; the data line is formed by establishing communication connections between multiple capability atoms.

[0196] The communication status of the data line is determined based on the heartbeat mechanism; the communication status includes normal communication and abnormal communication.

[0197] Based on the communication status and nonlinear optimization model of the communication line, the heartbeat mechanism corresponding to the data line is optimized;

[0198] The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method.

[0199] Optionally, the preset optimization method includes:

[0200] For data lines with normal communication, the initial heartbeat interval is increased by a first preset interval;

[0201] The initial weight value of the data line with communication failure is reduced by a first threshold, so that the initial heartbeat interval of the data line with communication failure is reduced by a second preset interval.

[0202] Optionally, before determining the data line based on multiple atomic capabilities, the processor 600 is further configured to:

[0203] Based on vehicle-to-everything (V2X) application data, several atomic capabilities are identified; these atomic capabilities include: data fusion atomic capability, on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0204] Optionally, the processor 600 is further configured to:

[0205] For application scenarios involving multiple access edge computing nodes corresponding to the data lines, the corresponding atomic capabilities are invoked.

[0206] Optionally, the processor 600 is further configured to:

[0207] The nonlinear optimization model is constructed based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line.

[0208] Wherein, the sum of the weight values ​​of the data sources corresponding to the data lines is 1;

[0209] The weight value of the data source is directly proportional to the heartbeat interval of the data line.

[0210] Optionally, before optimizing the heartbeat mechanism corresponding to the data line based on the communication status and nonlinear optimization model of the communication line, the processor 600 is further configured to:

[0211] Set an initial heartbeat interval on each data line;

[0212] The data line that receives feedback information within the initial heartbeat interval is identified as the data line with normal communication.

[0213] The data line that did not receive feedback information within the initial heartbeat interval is identified as the data line with the communication failure.

[0214] Optionally, after the processor 600 optimizes the heartbeat mechanism corresponding to the data line using the nonlinear optimization model according to a preset optimization method, the processor 600 is further configured to:

[0215] The heartbeat interval of normally communicating data lines with an increased first preset interval, and the weight values ​​of normally communicating data lines with a decreased first threshold, are input into the nonlinear optimization model for the next round of calculation.

[0216] Optionally, the processor 600 is specifically used for:

[0217] When the utilization rate of the CPU corresponding to the atomic capability of the data line with the communication failure is greater than a first preset utilization rate, the initial heartbeat interval of the data line with the communication failure is input into the nonlinear optimization model for the next round of calculation.

[0218] Optionally, the processor 600 is specifically used for:

[0219] When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is less than the first duration, the accuracy requirement is greater than the second threshold, and the recall value is greater than the third threshold, the first atomic capability in the atomic capabilities is invoked based on the heartbeat feedback mechanism.

[0220] When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is greater than or equal to the first duration, the accuracy requirement is less than or equal to the second threshold, and the recall requirement is less than or equal to the third threshold, the first atomic capability and / or the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism.

[0221] When the multi-access edge computing node corresponding to the data line is used to connect with other multi-access edge computing nodes, the first atomic capability and the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism.

[0222] The first atomic capability includes at least one of the following: on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

[0223] Optionally, before invoking the corresponding atomic capabilities for the application scenario of multiple access edge computing nodes corresponding to the data line, the processor 600 is specifically used for:

[0224] Based on the decision tree classification model, the application scenarios of the multi-access edge computing nodes corresponding to the data lines are determined.

[0225] Among them, Figure 6 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits of one or more processors represented by processor 600 and memory represented by memory 620 together. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 610 can be multiple components, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium, including wireless channels, wired channels, optical fibers, etc. For different user equipment, the user interface 630 can also be an interface capable of connecting external or internal devices, including but not limited to keypads, displays, speakers, microphones, joysticks, etc.

[0226] The processor 600 is responsible for managing the bus architecture and general processing, while the memory 620 can store the data used by the processor 600 during operation.

[0227] Optionally, the processor 600 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD), and the processor may also adopt a multi-core architecture.

[0228] This invention also provides a readable storage medium storing a program that, when executed by a processor, implements the data line optimization method described in any of the above embodiments.

[0229] The above describes the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also within the scope of protection of the present invention.

Claims

1. A method for optimizing data lines, characterized in that, include: The data path is determined based on the capabilities of multiple atoms; The data line is formed by establishing communication connections among multiple capability atoms; The communication status of the data line is determined based on the heartbeat mechanism; The communication status includes normal communication and abnormal communication; Based on the communication status of the data line and the nonlinear optimization model, the heartbeat mechanism corresponding to the data line is optimized; The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method. The method further includes: The nonlinear optimization model is constructed based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line; wherein, the sum of the weight values ​​of the data sources corresponding to the data line is 1; the weight value of the data source and the heartbeat interval of the data line are directly proportional.

2. The data line optimization method according to claim 1, characterized in that, The preset optimization methods include: For data lines with normal communication, the initial heartbeat interval is increased by a first preset interval; The initial weight value of the data line with communication failure is reduced by a first threshold, so that the initial heartbeat interval of the data line with communication failure is reduced by a second preset interval.

3. The data line optimization method according to claim 1, characterized in that, Before determining the data path based on multiple atomic capabilities, the method further includes: Based on vehicle-to-everything (V2X) application data, several atomic capabilities are identified; these atomic capabilities include: data fusion atomic capability, on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

4. The data line optimization method according to claim 1, characterized in that, The method further includes: For application scenarios involving multiple access edge computing nodes corresponding to the data lines, the corresponding atomic capabilities are invoked.

5. The data line optimization method according to claim 1, characterized in that, Before optimizing the heartbeat mechanism corresponding to the data line based on the communication status and nonlinear optimization model of the data line, the method further includes: Set an initial heartbeat interval on each data line; The data line that receives feedback information within the initial heartbeat interval is identified as the data line with normal communication. The data line that did not receive feedback information within the initial heartbeat interval is identified as the data line with the communication failure.

6. The data line optimization method according to claim 1, characterized in that, After the nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method, the method further includes: The heartbeat interval of normally communicating data lines with an increased first preset interval, and the weight values ​​of normally communicating data lines with a decreased first threshold, are input into the nonlinear optimization model for the next round of calculation.

7. The data line optimization method according to claim 6, characterized in that, The weight values ​​of the communication anomalies that reduce the first threshold are input into the nonlinear optimization model for the next round of calculation, including: When the utilization rate of the CPU corresponding to the atomic capability of the data line with the communication failure is greater than a first preset utilization rate, the initial heartbeat interval of the data line with the communication failure is input into the nonlinear optimization model for the next round of calculation.

8. The data line optimization method according to claim 4, characterized in that, For application scenarios involving multiple access edge computing nodes corresponding to the data lines, the corresponding atomic capabilities are invoked, including: When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is less than the first duration, the accuracy requirement is greater than the second threshold, and the recall value is greater than the third threshold, the first atomic capability among the atomic capabilities is invoked based on the heartbeat feedback mechanism. When the multi-access edge computing node corresponding to the data line is applied to an application scenario where the latency requirement is greater than or equal to the first duration, the accuracy requirement is less than or equal to the second threshold, and the recall requirement is less than or equal to the third threshold, the first atomic capability and / or the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism. When the multi-access edge computing node corresponding to the data line is used to connect with other multi-access edge computing nodes, the first atomic capability and the data fusion atomic capability in the atomic capabilities are invoked based on the heartbeat feedback mechanism. The first atomic capability includes at least one of the following: on-board unit (OBU) data parsing atomic capability, camera recognition algorithm capability, radar data parsing capability, and data communication atomic capability.

9. The data line optimization method according to claim 4, characterized in that, For application scenarios involving multiple access edge computing nodes corresponding to the data lines, before invoking the corresponding atomic capabilities, the method further includes: Based on the decision tree classification model, the application scenarios of the multi-access edge computing nodes corresponding to the data lines are determined.

10. A data line optimization device, characterized in that, include: The first determining module is used to determine the data line based on multiple atomic capabilities; The data line is formed by establishing communication connections among multiple capability atoms; The second determining module is used to determine the communication status of the data line based on the heartbeat mechanism; The communication status includes normal communication and abnormal communication; An optimization module is used to optimize the heartbeat mechanism corresponding to the data line based on the communication status and nonlinear optimization model of the data line. The nonlinear optimization model optimizes the heartbeat mechanism corresponding to the data line according to a preset optimization method. The device further includes: A module is established to construct the nonlinear optimization model based on the data source corresponding to the data line, the weight value of the data source, the preset evaluation index of the data source, and the heartbeat interval of the data line; wherein, the sum of the weight values ​​of the data sources corresponding to the data line is 1; the weight value of the data source and the heartbeat interval of the data line are directly proportional.

11. A network node, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the method for optimizing a data line as described in any one of claims 1 to 9.

12. A readable storage medium, characterized in that, The readable storage medium stores a program that, when executed by a processor, implements the data line optimization method as described in any one of claims 1 to 9.

Images