A vehicle data processing method, device, equipment and medium

Abstract

A vehicle data processing method, device, equipment and medium are disclosed. The method comprises: obtaining a plurality of to-be-processed data streams of a time-sensitive network in a vehicle; performing data conversion on the plurality of to-be-processed data streams to obtain a plurality of feature vectors corresponding to the plurality of to-be-processed data streams; inputting the plurality of feature vectors into a priority evaluation model and a graph neural network respectively to obtain a priority evaluation score and a mapping score of each to-be-processed data stream; determining a priority list of the plurality of to-be-processed data streams according to the priority evaluation score and the mapping score, and processing each to-be-processed data stream according to the priority list. The technical solution can dynamically adjust the priority of each data stream in the vehicle time-sensitive network, thereby ensuring that each data stream obtains the best priority in different scenarios, so that the data stream can be transmitted and processed in time.

Description

Technical Field

[0001] This invention relates to the field of vehicle data communication technology, and in particular to a vehicle data processing method, apparatus, device, and medium. Background Technology

[0002] The automotive industry is rapidly moving towards software-defined vehicles (SDVs), where traditional hardware-centric vehicle systems are being replaced or enhanced by software-defined functions. This shift enables in-vehicle systems to be dynamically reconfigured, customized, and optimized. With the development of autonomous driving and vehicle-to-everything (V2X) technologies, the ability to process massive amounts of real-time data from various sensors and external networks is becoming crucial.

[0003] Modern vehicles rely heavily on Time-Sensitive Networking (TSN) to manage time-critical data, such as sensor data, driver assistance system data, and infotainment communication data. However, current TSN systems typically employ static or pre-configured prioritization mechanisms, lacking flexibility in dynamic environments; maintaining time-sensitive data transmission becomes extremely challenging when network load is high or failures occur. Summary of the Invention

[0004] This invention provides a vehicle data processing method, apparatus, device, and medium that can dynamically adjust the priority of each data stream in a vehicle time-sensitive network, thereby ensuring that each data stream obtains the optimal priority in different scenarios so that the data stream can be transmitted and processed in a timely manner.

[0005] According to one aspect of the present invention, a vehicle data processing method is provided, comprising: Acquire multiple data streams to be processed from the time-sensitive network in the vehicle; The multiple data streams to be processed are transformed to obtain multiple feature vectors corresponding to the multiple data streams to be processed. The multiple feature vectors are respectively input into the priority evaluation model and the graph neural network to obtain the priority evaluation score and mapping score of each data stream to be processed; A priority list of multiple data streams to be processed is determined based on the priority evaluation score and the mapping score, and each data stream to be processed is processed according to the priority list.

[0006] According to another aspect of the present invention, a vehicle data processing apparatus is provided, comprising: The data acquisition module is used to acquire multiple data streams to be processed from the time-sensitive network in the vehicle; The data conversion module is used to convert the multiple data streams to be processed to obtain multiple feature vectors corresponding to the multiple data streams to be processed. The score acquisition module is used to input the multiple feature vectors into the priority evaluation model and the graph neural network respectively to obtain the priority evaluation score and mapping score of each data stream to be processed; The priority list determination module is used to determine a priority list of multiple data streams to be processed based on the priority evaluation score and the mapping score, and to process each data stream to be processed according to the priority list.

[0007] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the vehicle data processing method according to any embodiment of the present invention.

[0008] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the vehicle data processing method according to any embodiment of the present invention.

[0009] The technical solution of this invention involves acquiring multiple data streams to be processed from a time-sensitive network in a vehicle; performing data transformation on the multiple data streams to obtain multiple feature vectors corresponding to the multiple data streams; inputting the multiple feature vectors into a priority evaluation model and a graph neural network respectively to obtain a priority evaluation score and a mapping score for each data stream to be processed; determining a priority list for the multiple data streams to be processed based on the priority evaluation score and the mapping score; and processing each data stream according to the priority list. This technical solution can dynamically adjust the priority of each data stream in the vehicle's time-sensitive network, thereby ensuring that each data stream obtains the optimal priority in different scenarios, so that the data streams can be transmitted and processed in a timely manner.

[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart of a vehicle data processing method according to Embodiment 1 of the present invention; Figure 2 This is a flowchart of a vehicle data processing method according to Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the structure of a vehicle data processing device according to Embodiment 3 of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided according to Embodiment 4 of the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0014] It should be noted that the terms "first," "second," "initial," and "target," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0015] Example 1 Figure 1 This is a flowchart of a vehicle data processing method according to Embodiment 1 of the present invention. This embodiment is applicable to processing various data in a time-sensitive network within a vehicle. The method can be executed by a vehicle data processing device, which can be implemented in hardware and / or software and can be configured in an electronic device with data processing capabilities. Figure 1 As shown, the method includes: The technical solution of this embodiment can be executed by a Gen AI-driven dynamic TSN data stream prioritization system, specifically designed for the Software-Defined Vehicle (SDV) ecosystem. This system utilizes a Gen AI decision engine to dynamically adjust the TSN data stream priority in the SDV plane based on the real-time vehicle environment, thereby improving safety, operational efficiency, and modular adaptability in scenarios such as over-the-air (OTA) updates and component personalization. This innovative SDV plane can be integrated with existing SDN controllers to ensure seamless communication and network management. In this embodiment, the SDV plane acts as an intelligent orchestration layer, introducing adaptability and context awareness into the TSN data stream prioritization process.

[0016] S110. Acquire multiple data streams to be processed from the time-sensitive network in the vehicle.

[0017] Time-Sensitive Networking (TSN) is a protocol standard developed by IEEE 802.1, designed to provide a general time-sensitive mechanism for the data link layer of the Ethernet protocol. In vehicles, TSN ensures the real-time, deterministic, and reliable transmission of critical data. TSN provides a deterministic, low-latency Ethernet communication framework, which is crucial for applications such as autonomous driving, V2X (vehicle-to-everything) communication, and safety-critical systems. The data stream to be processed in a vehicle can also be referred to as a TSN data stream. The data stream to be processed can refer to the various data information streams transmitted in the time-sensitive network. In this embodiment, the time-sensitive network can manage various event-critical data, and the data stream to be processed can include, but is not limited to, various data streams from sources such as vehicle sensor data and external APIs. For example, external APIs may include the COVESA VSS API, the CAAM vehicle API, etc.

[0018] In this embodiment, multiple data streams to be processed from vehicle sensor data and external APIs can be acquired in the vehicle's time-sensitive network.

[0019] S120. Perform data transformation on multiple data streams to be processed to obtain multiple feature vectors corresponding to the multiple data streams to be processed.

[0020] Data transformation can be an operation that converts a data stream to be processed into structured feature vectors. Multiple feature vectors can refer to the individual feature vectors corresponding to each of the multiple data streams to be processed. In this embodiment, the multiple feature vectors correspond to the multiple data streams to be processed.

[0021] In this embodiment, context aggregation processing can be performed on each of the multiple data streams to be processed, and metadata encoding operations can be performed to convert them into structured feature vectors, thereby obtaining the feature vectors corresponding to each data stream in the multiple data streams to be processed.

[0022] Furthermore, the SDV plane in this embodiment may include a feature mapping layer, which can convert the input data of each data stream process into a structured feature vector for priority evaluation.

[0023] In this embodiment, optionally, data transformation is performed on multiple data streams to obtain multiple feature vectors corresponding to the multiple data streams to be processed, including: performing context data aggregation and data encoding operations on each of the multiple data streams to be processed to obtain multiple feature vectors corresponding to the multiple data streams to be processed.

[0024] Contextual data aggregation refers to extracting and integrating background information (context) related to the current data stream from multiple relevant data sources to enhance the semantic expressiveness and decision-making usefulness of the data. In this embodiment, in a Time-Sensitive Network (TSN), context refers to spatiotemporal, semantic, state, and environmental information related to the data stream, including but not limited to temporal context, spatial context aggregation, semantic context, state context, and environmental context. Temporal context can refer to the timestamp, periodicity, and temporal dependencies of the data stream. Spatial context can refer to the physical source and spatial coordinates of the data stream. For example, the physical source can be sensor locations, and the spatial coordinates can be the target location in the vehicle coordinate system. Semantic context can refer to the protocol type, message ID, and service type of the data stream. For example, the protocol type can be CAN, FlexRay, and SOME / IP, and the service type can be control commands and status feedback. State context can refer to the current state of the vehicle or subsystem, such as driving mode, fault status, and energy level. Environmental context can refer to external conditions, such as weather, road conditions, and lighting data. Data encoding operations can refer to encoding the aggregation context and metadata information of each data stream to obtain the corresponding structured feature vector.

[0025] In this embodiment, various data streams to be processed can be obtained from vehicle sensors and external APIs (such as COVESA VSS, CAAM), and the real-time context input of each data stream can be collected. For example, vehicle dynamics may include speed, mode, and braking state; environmental conditions may include weather, obstacles, and road hazards; and the functional classification of each data stream may include ADAS, infotainment, and diagnostic functions. Then, the aggregated context and metadata information of each data stream are encoded into a structured feature vector suitable for input into the Transformer model.

[0026] For example, in this embodiment, the feature mapping layer can take into account the data stream to be processed and the corresponding function configuration file for each data stream. The data stream to be processed may include real-time values ​​of VSS signal data (Vehicle Speed ​​Sensor) (e.g., / Vehicle / Speed, / ADAS / LaneKeepStatus), used to represent the dynamic operating status of the vehicle. The function configuration file (YAML / JSON) in this embodiment contains features such as ASIL, mode, criticality, and function for each VSS signal or data stream. In this embodiment, these combined attributes are encoded into numerical vectors as input to the priority evaluation engine.

[0027] This embodiment ensures that each data stream to be processed can be accurately represented, thereby enabling intelligent evaluation and adjustment of the priority of each data stream to be processed based on the fixed attributes of the data stream and the real-time vehicle context.

[0028] S130. Input multiple feature vectors into the priority evaluation model and the graph neural network respectively to obtain the priority evaluation score and mapping score of each data stream to be processed.

[0029] The priority evaluation model can be a pre-trained Transformer-based classification model. In this embodiment, the priority evaluation model can be a fine-tuned Transformer classification model (e.g., a BERT or GPT variant). In this embodiment, the priority evaluation model can assess the criticality of each data stream and output a priority evaluation score for each data stream. A graph neural network can construct a graph from the feature vectors corresponding to the input data streams and output a priority modifier or mapped priority score (i.e., a mapping score) for each data stream based on dependencies. Graph neural networks can be used to map each data stream to a priority level based on its attributes (e.g., latency requirements, bandwidth requirements, and functional interdependencies). In this embodiment, the graph neural networks are all pre-trained neural networks (GNNs).

[0030] In this embodiment, within the Time-Sensitive Network (TSN), mapping data streams to priorities using a Graph Neural Network (GNN) involves constructing a dependency graph between data streams and utilizing the GNN to learn the embedded representations of nodes (data streams), thereby automatically inferring their priorities to meet the TSN's requirements for low latency, high reliability, and determinism. The priority evaluation score can be obtained by classifying and evaluating the keyness level of the feature vectors corresponding to each data stream to be processed using a priority evaluation model. The mapping score can be a priority score obtained by mapping the dependencies of each data stream to be processed using a Graph Neural Network.

[0031] In this embodiment, each feature vector from multiple feature vectors can be input into a Transformer-based priority evaluation model. The priority evaluation model classifies the criticality level and assesses the relative urgency of the flow in the current context. Specifically, it can evaluate the criticality of each data flow based on data such as vehicle safety requirements, vehicle status, and vehicle environmental background. For example, safety requirements could be braking system data > infotainment data requirements; vehicle status could refer to highway or emergency conditions; and environmental background data could refer to weather or road hazard data. Then, the assessed relative urgency can be quantified into a priority evaluation score, and each feature vector from multiple feature vectors can be input into a graph neural network. The graph neural network constructs a corresponding graph based on the feature vectors, and determines the mapping score for each data flow based on the dependencies in the graph.

[0032] S140. Determine a priority list of multiple data streams to be processed based on the priority evaluation score and mapping score, and process each data stream to be processed according to the priority list.

[0033] The priority list can refer to a list of data obtained by sorting the priorities of multiple data streams to be processed. In this embodiment, the priority evaluation score and mapping score corresponding to each data stream to be processed can be fused to obtain the final priority score of each data stream to be processed. Then, the data streams to be processed are sorted according to the final priority scores to obtain the priority list corresponding to multiple data streams to be processed. The data streams are then transmitted and processed sequentially according to the sorting order in the priority list.

[0034] In this embodiment, the SDV plane acts as an intelligent orchestration layer, introducing adaptability and context awareness into the TSN prioritization process. The SDV plane may include a Gen AI decision engine and a stream priority generator. The Gen AI decision engine is the core component responsible for analyzing the contextual data of data streams and prioritizing all active data streams. The Gen AI decision engine can consist of a priority evaluation engine and a stream mapping engine; the priority evaluation engine is used to evaluate the priority score of each data stream, ensuring that critical operations are processed with the highest priority. The stream mapping engine is used to map each data stream to a priority based on its attribute information and the relationships between data streams, using graphical modeling; for example, attribute information can be latency requirements, bandwidth requirements, and functional interdependencies. The stream mapping engine adds system-level intelligence by simulating how data streams work together, ensuring consistent, efficient, and secure network behavior in complex SDV environments. It can determine how different data streams are related and assign optimal priorities based on these interdependencies. It ensures that functionally related streams (e.g., control signals from cameras + LiDAR + ADAS) are prioritized coherently rather than in isolation.

[0035] This embodiment determines the priority list of each data stream based not only on the importance of individual data streams but also on the interrelationships between them. For example, a single camera signal may not be important, but its priority increases when combined with braking control. This achieves smarter, more context-aware prioritization, optimizing network performance and security by avoiding isolated or conflicting decisions.

[0036] Understandably, in this embodiment, the priority evaluation engine analyzes vehicle safety requirements, vehicle status, and environmental context to generate a safety score; the stream mapping engine optimizes priority allocation by understanding the interdependencies of data streams, thereby automatically prioritizing multiple pending data streams (TSNs) of the time-sensitive network, minimizing manual reconfiguration work, and effectively reducing operating costs.

[0037] Furthermore, in this embodiment, after determining the optimal priority for each data stream to be processed based on the priority evaluation score and mapping score, the SDV plane generates a priority list based on the optimal priority of each data stream and sends it to the SDN controller. The SDN controller enforces these priorities by configuring the TSN switches, causing them to transmit and process data streams according to the assigned priorities. Specifically, in this embodiment, the priority score calculated by the priority evaluation engine and the mapping score obtained by the GNN in the flow mapping engine can both be passed to the flow priority generator. The flow priority generator merges the scores to obtain the final priority score. A structured priority list can be formed based on the final priority scores of each data stream, and then the priority list is passed to the TSN-SDN controller via the northbound API, thereby achieving seamless implementation of the Quality of Service (QoS) policy.

[0038] The technical solution of this invention involves acquiring multiple data streams to be processed from a time-sensitive network in a vehicle; performing data transformation on these multiple data streams to obtain multiple feature vectors corresponding to each data stream; inputting these feature vectors into a priority evaluation model and a graph neural network respectively to obtain a priority evaluation score and a mapping score for each data stream; determining a priority list for the multiple data streams based on the priority evaluation score and the mapping score; and processing each data stream according to the priority list. This technical solution can dynamically adjust the priority of each data stream in the vehicle's time-sensitive network, thereby ensuring that each data stream obtains the optimal priority in different scenarios, so that the data streams can be transmitted and processed in a timely manner.

[0039] Example 2 Figure 2 This is a flowchart of a vehicle data processing method according to Embodiment 2 of the present invention. This embodiment is an optimization based on the above embodiment. Specifically, the optimization involves inputting multiple feature vectors into a priority evaluation model and a graph neural network to obtain a priority evaluation score and a mapping score for each data stream to be processed. This includes: inputting multiple feature vectors into the priority evaluation model to output an initial priority evaluation score for each data stream to be processed; adjusting the initial priority evaluation score according to a set rule to obtain a priority evaluation score for each data stream to be processed; and inputting multiple feature vectors into the graph neural network to obtain a mapping score for each data stream to be processed. Figure 2 As shown, the method includes: S210. Acquire multiple data streams to be processed from the time-sensitive network in the vehicle.

[0040] S220. Perform data transformation on multiple data streams to be processed to obtain multiple feature vectors corresponding to the multiple data streams to be processed.

[0041] S230. Input multiple feature vectors into the priority evaluation model respectively, and output the initial priority evaluation score corresponding to each data stream to be processed.

[0042] The initial priority evaluation score can be obtained by evaluating the data streams to be processed using a priority evaluation model. In this embodiment, multiple feature vectors can be input into the priority evaluation model, which classifies the data streams to be processed into criticality levels and outputs a numerical priority score for each data stream, such as a score range of 0-100. A higher score indicates higher criticality; for example, the score of the data stream corresponding to the braking signal is higher than the score of the data stream corresponding to the infotainment media. This allows us to obtain the initial priority evaluation score for each data stream to be processed.

[0043] S240. Adjust the initial priority evaluation score according to the set rules to obtain the priority evaluation score corresponding to each data stream to be processed.

[0044] The rules can be pre-defined based on specific information such as vehicle safety requirements and vehicle status. The priority assessment score can refer to the priority score obtained after standardizing and filtering the initial priority assessment score through the set rules.

[0045] In this embodiment, the initial priority evaluation score can be standardized and filtered according to pre-defined rules. Specifically, the initial priority evaluation score can be adjusted by applying rules specific to a particular domain, thereby obtaining the priority evaluation score corresponding to each data stream to be processed, so as to ensure that non-critical streams do not interfere with other high-priority data streams.

[0046] S250. Input multiple feature vectors into the graph neural network to obtain the mapping score of each data stream to be processed.

[0047] In this embodiment, each feature vector from multiple feature vectors can be input into a graph neural network. The graph neural network constructs a graph for each feature vector to obtain a data flow graph corresponding to each data flow to be processed. Then, based on the interdependence between each data flow node in the constructed data flow graph, the mapping score corresponding to each data flow node is determined, thereby obtaining the mapping score of each data flow to be processed.

[0048] In this embodiment, since the data flow in the vehicle is usually functional or temporally dependent, it can be naturally represented using a graph structure: nodes are represented as data flows, and edges are represented as dependencies or shared resources (e.g., bandwidth, time windows, sequence correlation) between data flows. Graph neural networks (GNNs) are well-suited to learn from such structured relationships and capture the global context of the network.

[0049] In this embodiment, optionally, multiple feature vectors are input into a graph neural network to obtain a mapping score for each data stream to be processed, including: inputting the feature vector of each data stream to be processed into the graph neural network, constructing a data stream graph based on the feature vector of each data stream to be processed by the graph neural network, and determining the mapping score of each data stream to be processed based on the data stream graph.

[0050] In this context, a data flow graph can be a graph structure constructed by taking the data flow to be processed as nodes and their corresponding feature vectors as node attributes, based on the interdependencies between the data flows.

[0051] In this embodiment, the feature vector of each data stream to be processed can be input into a graph neural network. The graph neural network can represent each data stream as a node in the data stream graph, and assign a corresponding feature vector to each node (data stream). For example, the feature vector may include static attributes (ASIL level, source ECU and function class), dynamic VSS input (vehicle speed and environment), etc. Then, the interdependencies between the nodes (e.g., camera feeds and sensor data of ADAS) are captured as edges of the graph. The edges of the data stream graph can represent the interdependencies or shared resources between data streams, such as using the same communication channel, time correlation or functional association (e.g., radar and camera). Thus, the data stream graph can be constructed based on the determined nodes and the edges between nodes. Then, according to the GNN, a mapping priority score (or priority modifier) ​​can be output for each data stream to be processed based on the dependencies in the data stream graph.

[0052] In this embodiment, a Graph Neural Network (GNN) is used to process the data stream, encoding node-level features to understand the influence between nodes (e.g., the presence of other data streams may increase the importance of a particular data stream), thereby capturing global graph-level patterns. This embodiment can train the GNN using historical or simulated driving scenarios. The ground truth labels can represent priority adjustments due to interdependencies; for example, when both radar and cameras are streaming, one may receive higher priority depending on the situation.

[0053] In this embodiment, when processing the data stream in the time-sensitive network in the vehicle, a real-time graph is constructed using real-time stream information through the SDV plane. Then, the GNN outputs a priority modifier (or mapping priority) for each stream based on the dependency relationship, and these values ​​are passed to the stream priority generator.

[0054] Understandably, in this embodiment, the output of the graph neural network (GNN) is not the final priority, but a refined / modulated version that takes into account interdependencies. It needs to be fused with the priority score output by the Transformer-based priority evaluation model to produce the final priority transmission order.

[0055] In this embodiment, with this setup, a graph neural network can map each data stream to a mapping score based on the attribute information (such as latency requirements, bandwidth requirements, and functional interdependencies) contained in its feature vector. The graph neural network can dynamically learn the optimal mapping for each data stream to be processed, so as to prioritize each data stream.

[0056] S260. Determine a priority list of multiple data streams to be processed based on the priority evaluation score and mapping score, and process each data stream to be processed according to the priority list.

[0057] In this embodiment, optionally, determining a priority list of multiple data streams to be processed based on priority evaluation scores and mapping scores includes: determining a target priority score for each data stream to be processed based on priority evaluation scores and mapping scores; and sorting the multiple data streams to be processed according to the target priority scores to obtain a corresponding priority list.

[0058] The target priority score can be considered as the final priority score of each data stream to be processed. In this embodiment, the target priority score can be obtained by fusing the priority evaluation score and the mapping score. In this embodiment, the priority evaluation score and the mapping score can be fused according to their respective weight factors to obtain the target priority score of each data stream to be processed. Then, each data stream to be processed is prioritized according to its target priority score to obtain a prioritized list.

[0059] For example, in this embodiment, the priority streams can be output as a structured list as follows: [{"stream_id": "camera_front", "priority": 1}, {"stream_id": "lidar", "priority": 2}, {"stream_id": "radar", "priority": 3}, ...].

[0060] This embodiment, through such a setting, can provide real-time priority for each data stream to be processed based on vehicle operation and environmental conditions, ensuring that critical systems such as braking or ADAS always enjoy the highest priority, even in high-demand scenarios, thereby improving the dynamic adaptability of data processing and vehicle safety.

[0061] In this embodiment, optionally, determining the target priority score for each data stream to be processed based on the priority evaluation score and the mapping score includes: determining the first weighting factor and the second weighting factor corresponding to the priority evaluation score and the mapping score; and fusing the priority evaluation score, the mapping score, the first weighting factor, and the second weighting factor to obtain the target priority score.

[0062] The first weighting factor can be the weighting factor corresponding to the priority evaluation score. The second weighting factor can be the weighting factor corresponding to the mapping score. In this embodiment, the first and second weighting factors can be dynamically set according to actual needs. The fusion processing can be a weighted summation operation.

[0063] Specifically, in this embodiment, each data stream evaluated by the priority evaluation engine has a corresponding initial priority score, i.e., a priority evaluation score. For example, the initial priority score can be a floating-point value between 0 and 1 or a category level such as "high," "medium," or "low." Additionally, a dependency-based adjusted score or priority modifier, i.e., a mapping score, is obtained from the graph neural network (GNN) output by the flow mapping engine for each data stream. In this embodiment, the priority evaluation score and the mapping score are weighted and summed using a fusion function, combining their respective first and second weighting factors, to obtain the target priority score. Specifically, the target priority score can be: Target Priority Score = w1 * Priority Evaluation Score + w2 * Mapping Score; where the first weighting factor w1 and the second weighting factor w2 can be adjustable weights based on performance testing.

[0064] In this embodiment, by using this setting, the priority scores obtained from the two evaluation methods can be fused to obtain the target priority score, thereby determining the real-time priority of each data stream to be processed, so as to perform subsequent transmission or processing operations on different data streams according to the priority of each data stream.

[0065] In this embodiment, optionally, after determining the priority list of multiple data streams to be processed, the method further includes: dynamically adjusting the data transmission routing strategy of the data streams to be processed based on the priority list of multiple data streams to be processed.

[0066] In this embodiment, the data transmission routing strategy refers to the optimal transmission path for transmitting the data streams to be processed. After determining the priority list of multiple data streams to be processed, the priority list can be sent to the SDN controller that supports TSN, allowing for dynamic adjustment of the data transmission routing strategy to ensure compliance with real-time Quality of Service (QoS) policies. This embodiment can also pass the priority list to the TSN configuration engine, thereby configuring QoS and scheduling rules on the network based on the priority list.

[0067] Specifically, in this embodiment, the data content in the SDN controller supporting TSN can be adjusted according to the priority list. Specifically, the following adjustments can be made: 1. Adjust queue scheduling: Assign flows to specific queues based on priority (e.g., high-priority traffic to a fast queue). TSN scheduling mechanisms such as Time-Aware Shapers (IEEE 802.1Qbv) ensure deterministic delivery. 2. Change flow paths: Reroute critical data via low-latency or less congested paths. Avoid faulty nodes or links based on priority. 3. Modify VLAN priority / PCP bits: Set the Priority Code Point (PCP) value in the Ethernet header according to flow priority. 4. Update flow rules in switches: Dynamically push new flow rules to TSN switches using OpenFlow or similar protocols. For example, set a matching action rule: "If / vehicle / speed flow → forward to queue 1 with strict priority". 5. Adjust transmission gate timing: Dynamically modify the transmission plan for each queue (especially when using Qbv) to ensure the time period for high-priority flows. 6. Bandwidth Reservation (Qcc): Flow Reservation Protocol (SRP) and Centralized Network Configuration (CNC) logic reallocate bandwidth to high-priority flows.

[0068] In this embodiment, the priority list of multiple data streams to be processed can be sent to the SDN controller. The SDN controller then enforces these priorities by configuring the TSN switch, so that it transmits and processes the data streams according to the assigned priorities, ensuring that the data streams to be processed can be processed in a timely manner.

[0069] The system in this example may also include a control plane and a data plane. The control plane can be used to manage the interaction between the SDV plane and the vehicle's TSN infrastructure. For example, an SDN controller supporting TSN can be used to implement priority lists generated by the SDV plane; dynamically adjust routing policies to ensure real-time QoS compliance; and implement CUC (Centralized User Configuration) and CNC (Centralized Network Configuration) for data flow and network management and synchronization. The data plane may include physical and virtual TSN infrastructure. For example, TSN SDN switches can be used for routing, processing endpoints (senders and receivers) of priority data flows, and ensuring reliable delivery of time-sensitive data to appropriate listeners with minimal latency.

[0070] This embodiment features a Gen AI-driven dynamic TSN data stream prioritization system designed for seamless integration into the SDV ecosystem to address the following trends: Over-the-air updates: Supporting the addition of new components or features without reconfiguring TSN priorities. The SDV plane dynamically updates the priority list to adapt to new demands. Personalized configuration: Providing customized network configurations for individual drivers by configuring feature profiles to determine patterns and adjust safety requirements in the feature mapping layer, such as prioritizing infotainment systems during relaxed driving or safety-critical systems during highway driving; adapting to changes in vehicle architecture, such as replacing hardware modules or upgrading software. In this embodiment, the system can also dynamically reallocate bandwidth based on current needs to optimize network resource usage, improve operational efficiency, and align with the automotive industry's software-centric development, supporting OTA updates, modular architecture, and other trends.

[0071] This embodiment integrates advanced AI models (such as Transformers and graph neural networks) into the SDV plane to ensure optimal priority in different scenarios, enabling dynamic reconfiguration of network priorities during OTA updates. During OTA updates, new components or software modules are installed, introducing new data flows or altering existing ones. The SDV plane detects these changes via the vehicle API, uses its AI model to re-evaluate the priorities of all data flows in real time, generates an updated priority list containing the added or modified data flows, and sends it to the SDN controller. The SDN controller reconfigures TSN network scheduling and QoS rules accordingly by reconfiguring TSN switches, eliminating the need for manual reconfiguration. This embodiment also allows for context-driven prioritization based on real-time environment and operational data, handling dependencies between various data flows in a resource-efficient manner.

[0072] This embodiment illustrates some SDV use cases implemented using this technical solution. For example, by configuring an emergency braking scenario in the SDV, in high-speed emergency situations, the Gen AI decision engine identifies braking and sensor data streams as critical data and assigns them the highest priority. For OTA update integration, when a new ADAS module is added via OTA, the system dynamically recalibrates the TSN priority to include the module without interrupting existing functionality. Infotainment-priority configuration means that during leisurely driving, the infotainment stream may receive higher priority, while in dangerous situations, safety-critical data streams dominate.

[0073] The technical solution of this invention involves acquiring multiple data streams to be processed from a time-sensitive network in a vehicle; performing data transformation on the multiple data streams to obtain multiple feature vectors corresponding to the multiple data streams; inputting the multiple feature vectors into a priority evaluation model to output an initial priority evaluation score for each data stream; adjusting the initial priority evaluation score according to a set rule to obtain a priority evaluation score for each data stream; inputting the multiple feature vectors into a graph neural network to obtain a mapping score for each data stream; determining a priority list for the multiple data streams based on the priority evaluation score and the mapping score; and processing each data stream according to the priority list. This technical solution can dynamically adjust the priority of each data stream in the vehicle's time-sensitive network, thereby ensuring that each data stream obtains the optimal priority in different scenarios, so that the data streams can be transmitted and processed in a timely manner.

[0074] Example 3 Figure 3 This is a schematic diagram of the structure of a vehicle data processing device according to Embodiment 3 of the present invention. Figure 3 As shown, the device includes: The data acquisition module 310 is used to acquire multiple data streams to be processed from the time-sensitive network in the vehicle; The data conversion module 320 is used to convert multiple data streams to be processed, and obtain multiple feature vectors corresponding to the multiple data streams to be processed. The score acquisition module 330 is used to input multiple feature vectors into the priority evaluation model and the graph neural network respectively to obtain the priority evaluation score and mapping score of each data stream to be processed; The priority list determination module 340 is used to determine a priority list of multiple data streams to be processed based on the priority evaluation score and the mapping score, and to process each data stream to be processed according to the priority list.

[0075] Optional, the data conversion module 320 is specifically used for: For each of the multiple data streams to be processed, context data aggregation and data encoding operations are performed to obtain multiple feature vectors corresponding to the multiple data streams to be processed.

[0076] Optionally, the score acquisition module 330 includes: The initial evaluation unit is used to input multiple feature vectors into the priority evaluation model and output the initial priority evaluation score for each data stream to be processed. The first score evaluation unit is used to adjust the initial priority evaluation score according to the set rules to obtain the priority evaluation score corresponding to each data stream to be processed; The second score evaluation unit is used to input multiple feature vectors into the graph neural network to obtain the mapping score for each data stream to be processed.

[0077] Optional, a second score assessment unit, specifically used for: The feature vector of each data stream to be processed is input into the graph neural network, and the graph neural network constructs a data stream graph based on the feature vector of each data stream to be processed. The mapping score for each data stream to be processed is determined based on the data stream graph.

[0078] Optionally, the priority list determination module 340 includes: The target score determination unit is used to determine the target priority score for each data stream to be processed based on the priority evaluation score and the mapping score. The sorting unit is used to sort multiple data streams to be processed according to the target priority score to obtain the corresponding priority list.

[0079] Optional, target score determination unit, specifically used for: Determine the first and second weighting factors corresponding to the priority evaluation scores and mapping scores; The priority evaluation score, mapping score, first weighting factor, and second weighting factor are fused together to obtain the target priority score.

[0080] Optionally, the device may also include: The adjustment module is used to dynamically adjust the data transmission routing strategy of the data streams to be processed based on the priority list of the multiple data streams to be processed after determining the priority list of the multiple data streams to be processed.

[0081] The vehicle data processing device provided in this embodiment of the invention can execute a vehicle data processing method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0082] Example 4 Figure 4 This is a schematic diagram of an electronic device according to Embodiment 4 of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0083] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0084] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0085] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as vehicle data processing methods.

[0086] In some embodiments, the vehicle data processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the vehicle data processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the vehicle data processing method by any other suitable means (e.g., by means of firmware).

[0087] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0088] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0089] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0090] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0091] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0092] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0093] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0094] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A vehicle data processing method, characterized in that, include: Acquire multiple data streams to be processed from the time-sensitive network in the vehicle; The multiple data streams to be processed are transformed to obtain multiple feature vectors corresponding to the multiple data streams to be processed. The multiple feature vectors are respectively input into the priority evaluation model and the graph neural network to obtain the priority evaluation score and mapping score of each data stream to be processed; A priority list of multiple data streams to be processed is determined based on the priority evaluation score and the mapping score, and each data stream to be processed is processed according to the priority list.

2. The method of claim 1, wherein, The multiple data streams to be processed are transformed to obtain multiple feature vectors corresponding to the multiple data streams to be processed, including: For each of the multiple data streams to be processed, context data aggregation and data encoding operations are performed to obtain multiple feature vectors corresponding to the multiple data streams to be processed.

3. The method of claim 1, wherein, The multiple feature vectors are respectively input into the priority evaluation model and the graph neural network to obtain the priority evaluation score and mapping score of each data stream to be processed, including: The multiple feature vectors are respectively input into the priority evaluation model, and the initial priority evaluation score corresponding to each data stream to be processed is output. The initial priority evaluation score is adjusted according to the set rules to obtain the priority evaluation score corresponding to each data stream to be processed; The multiple feature vectors are input into the graph neural network to obtain the mapping score for each data stream to be processed.

4. The method of claim 3, wherein, The multiple feature vectors are input into a graph neural network to obtain the mapping score for each data stream to be processed, including: The feature vector of each data stream to be processed is input into a graph neural network, and the graph neural network constructs a data stream graph based on the feature vector of each data stream to be processed. The mapping score for each data stream to be processed is determined based on the data stream graph.

5. The method of claim 1, wherein, A priority list of multiple data streams to be processed is determined based on the priority evaluation score and the mapping score, including: The target priority score for each data stream to be processed is determined based on the priority evaluation score and the mapping score. Multiple data streams to be processed are sorted according to the target priority score to obtain a corresponding priority list.

6. The method according to claim 5, characterized in that, The target priority score for each data stream to be processed is determined based on the priority evaluation score and the mapping score, including: Determine the first weighting factor and the second weighting factor corresponding to the priority evaluation score and the mapping score; The priority evaluation score, the mapping score, the first weighting factor, and the second weighting factor are fused together to obtain the target priority score.

7. The method of claim 1, wherein, After determining the priority list of multiple data streams to be processed, the following is also included: The data transmission routing strategy for the multiple data streams to be processed is dynamically adjusted based on their priority list.

8. A vehicle data processing apparatus characterized by comprising: include: The data acquisition module is used to acquire multiple data streams to be processed from the time-sensitive network in the vehicle; The data conversion module is used to convert the multiple data streams to be processed to obtain multiple feature vectors corresponding to the multiple data streams to be processed. The score acquisition module is used to input the multiple feature vectors into the priority evaluation model and the graph neural network respectively to obtain the priority evaluation score and mapping score of each data stream to be processed; The priority list determination module is used to determine a priority list of multiple data streams to be processed based on the priority evaluation score and the mapping score, and to process each data stream to be processed according to the priority list.

9. An electronic device, comprising: The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the vehicle data processing method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the vehicle data processing method according to any one of claims 1-7.

Images