CN-121979060-A - Whole vehicle data fusion system and method

CN121979060ACN 121979060 ACN121979060 ACN 121979060ACN-121979060-A

Abstract

The invention provides a whole vehicle data fusion system and method, wherein the system comprises a plurality of regional controllers, each regional controller is configured to acquire monitoring data of various vehicle-mounted devices in a region and fuse the monitoring data to generate standardized data services representing the state or environment of the region, a whole vehicle computing platform is configured to receive and conduct vehicle-level fusion and decision making based on all the standardized data services to generate a global driving behavior decision, decompose the global driving behavior decision into a cross-domain control instruction and send the cross-domain control instruction to the regional controllers, and a cloud service platform is configured to optimize algorithm parameters or models based on data fusion in the regional controllers and/or the whole vehicle computing platform based on state data and fault diagnosis data of vehicles and send optimized results to the whole vehicle computing platform and/or the regional controllers. The invention reduces the processing pressure and the load pressure of the whole vehicle computing platform and improves the response speed.

Inventors

ZHANG FENG
GAO LANG

Assignees

武汉江夏楚能汽车技术研发有限公司

Dates

Publication Date: 20260505
Application Date: 20260202

Claims (10)

1. The whole vehicle data fusion system is characterized by comprising: The system comprises a plurality of regional controllers, a plurality of vehicle-mounted devices and a plurality of data processing units, wherein the regional controllers are distributed in a plurality of regions of a vehicle and are configured to acquire monitoring data of a plurality of vehicle-mounted devices in the regions, and perform fusion processing on the monitoring data to generate standardized data services representing the states or the environments of the regions; The whole vehicle computing platform is configured to receive the standardized data services of each regional controller, perform vehicle-level fusion and decision-making based on all the standardized data services, generate a global driving behavior decision, decompose the global driving behavior decision into a cross-domain control instruction, and send the cross-domain control instruction to the regional controller; The cloud service platform is configured to receive state data and fault diagnosis data from a vehicle, optimize algorithm parameters or models according to which data in the regional controller and/or the whole vehicle computing platform are fused based on the state data and the fault diagnosis data, and send an optimization result to the whole vehicle computing platform and/or the regional controller.
2. The whole vehicle data fusion system of claim 1, wherein the zone controller is configured to initiate a fusion process in a periodically triggered or event triggered manner; The event triggering comprises at least one of detecting that the sensor senses a new target, the working condition of the vehicle meets a preset switching condition and receiving a fault warning signal exceeding a preset level.
3. The whole vehicle data fusion system of claim 1, wherein the zone controller is specifically configured to: acquiring monitoring data of various vehicle-mounted devices in the area; performing multi-dimensional verification on the monitoring data, wherein the multi-dimensional verification comprises identity verification, protocol verification, performance verification, address verification, data verification and channel verification; When the multi-dimensional verification is passed, a time stamp and a quality label based on a global clock of a vehicle are given to the monitoring data, available data in the monitoring data are determined based on the quality label, and the available data describing the same physical object or state are subjected to space-time association and fusion to generate the standardized data service; And when the multi-dimensional verification is not passed, calling a historical data service of the last period stored in the area controller, and taking the historical data service as the standardized data service.
4. The whole vehicle data fusion system of claim 1, wherein the zone controller is further configured to: when the monitoring data is abnormal, redundant switching is executed, or fusion processing is carried out based on historical monitoring data; and when the execution time of the fusion processing exceeds a first preset duration, switching to the simplified fusion processing with shorter execution time.
5. The whole vehicle data fusion system of claim 1, wherein the whole vehicle computing platform is specifically configured to: determining a plurality of data to be compared in the standardized data service of the area controller, wherein the data to be compared represents the same vehicle state or environment object; judging whether the plurality of data to be compared are consistent or not, and judging whether the number of the data to be compared is larger than a preset number or not; and when the plurality of data to be compared are consistent and the number of the data to be compared is larger than the preset number, vehicle-level fusion and decision are carried out on the basis of the plurality of data to be compared.
6. The whole vehicle data fusion system of claim 5, wherein the whole vehicle computing platform is further configured to: Generating a global vehicle dynamic state and a global environment map based on the plurality of comparison data; generating the global driving behavior decision based on the global vehicle dynamic state and the global environment map; And decomposing the global driving behavior decision into a group of cross-domain control instructions, and issuing the cross-domain control instructions to the regional controller.
7. The whole vehicle data fusion system of claim 1, wherein the whole vehicle computing platform is configured to: the method comprises the steps of executing vehicle-level fusion and decision tasks based on a preset priority scheduling strategy, wherein the priority scheduling strategy divides tasks of different functional domains into a plurality of priority levels and allocates duty ratios for high-priority tasks corresponding to high calculation forces; And monitoring the resource occupancy rate of the whole vehicle computing platform, and when the resource occupancy rate exceeds a preset threshold value, migrating the computing task of the whole vehicle computing platform to one or more regional controllers for execution.
8. The whole vehicle data fusion system of claim 1, wherein the whole vehicle computing platform is further configured to: And switching the whole vehicle level fusion function to different degradation modes for operation according to the detected fault range and severity of the regional controller and/or the whole vehicle computing platform, wherein the degradation modes comprise function simplification, fusion flow simplification or stopping whole vehicle level fusion.
9. The whole vehicle data fusion system of claim 1, wherein the cloud service platform is further configured to: Carrying out statistical analysis on the state data and the fault diagnosis data according to the instruction type and the vehicle operation condition to generate a statistical analysis result, and adjusting algorithm parameters or models according to the data fusion according to the statistical analysis result to obtain optimization parameters or optimization models; And packaging the optimization parameters or the optimization model into a software update version, and pushing the software update version to vehicles in batches in an OTA mode, wherein verification is carried out on a selected batch of test vehicles before pushing the software update version to the fleet in full quantity.
10. The whole vehicle data fusion method is characterized by comprising the following steps of: A plurality of regional controllers arranged in a plurality of regions of a vehicle are controlled to acquire monitoring data of a plurality of vehicle-mounted devices in the regions, and the monitoring data are subjected to fusion processing to generate standardized data service representing the state or environment of the regions; The whole vehicle computing platform is controlled to receive the standardized data services of each regional controller, vehicle-level fusion and decision are carried out based on all the standardized data services, global driving behavior decisions are generated, the global driving behavior decisions are decomposed into cross-domain control instructions, and the cross-domain control instructions are issued to the regional controllers; And the cloud service platform is controlled to receive state data and fault diagnosis data from the vehicle, optimize algorithm parameters or models according to which the data in the regional controller and/or the whole vehicle computing platform are fused based on the state data and the fault diagnosis data, and send an optimization result to the whole vehicle computing platform and/or the regional controller.

Description

Whole vehicle data fusion system and method Technical Field The invention relates to the technical field of whole vehicle data processing, in particular to a whole vehicle data fusion system and method. Background With the continuous improvement of the intelligent and networking degree of automobiles, the electronic and electric architecture of the automobiles is undergoing a deep transformation from a traditional distributed architecture to a regional oriented architecture. In this process, the in-vehicle network forms a hybrid architecture where signal-oriented communications (e.g., CAN, LIN bus) coexist with service-oriented communications (e.g., SOME/IP, DDS, etc., IP-based protocols). The signal data has periodic and low granularity characteristics, and the service data is event-driven and high granularity information, which have obvious differences in data format, semantic definition and timing characteristics. This heterogeneity makes it difficult for data from different network domains, different protocols, to be uniformly understood and processed, resulting in difficulties in cross-domain data fusion and high system integration complexity. The existing scheme mostly adopts a central centralized data processing mode, namely, all original or primarily processed sensor and controller data are converged to a central computing unit for fusion and decision. The method has the obvious defects that on one hand, the centralized transmission of mass data brings huge pressure to the bandwidth of a vehicle-mounted network, a central computing unit is easy to become a system performance bottleneck, the processing efficiency is restricted, and on the other hand, for the application with extremely high requirements on real-time performance such as intelligent driving, chassis control and the like, the delay introduced by the data through long-path transmission and centralized queuing processing is difficult to meet the requirement of millisecond-level response, and the reliability and the safety of the system function are influenced. Therefore, it is needed to provide a whole vehicle data fusion system and method, which can improve the processing efficiency while realizing cross-domain data fusion, so that the vehicle operation has reliability and safety. Disclosure of Invention In view of the foregoing, it is necessary to provide a system and a method for integrating data of a whole vehicle, which are used for solving the technical problems that the data processing efficiency is low, the requirement of millisecond-level response cannot be met, and the reliability and safety of the system function are affected due to the fact that the data processing efficiency is low caused by a central centralized data processing mode in the prior art. In order to solve the above technical problems, in a first aspect, the present invention provides a whole vehicle data fusion system, including: The system comprises a plurality of regional controllers, a plurality of vehicle-mounted devices and a plurality of data processing units, wherein the regional controllers are distributed in a plurality of regions of a vehicle and are configured to acquire monitoring data of a plurality of vehicle-mounted devices in the regions, and perform fusion processing on the monitoring data to generate standardized data services representing the states or the environments of the regions; The whole vehicle computing platform is configured to receive the standardized data services of each regional controller, perform vehicle-level fusion and decision-making based on all the standardized data services, generate a global driving behavior decision, decompose the global driving behavior decision into a cross-domain control instruction, and send the cross-domain control instruction to the regional controller; The cloud service platform is configured to receive state data and fault diagnosis data from a vehicle, optimize algorithm parameters or models according to which data in the regional controller and/or the whole vehicle computing platform are fused based on the state data and the fault diagnosis data, and send an optimization result to the whole vehicle computing platform and/or the regional controller. In one possible implementation, the zone controller is configured to initiate the fusion process in a periodically triggered or event-triggered manner; The event triggering comprises at least one of detecting that the sensor senses a new target, the working condition of the vehicle meets a preset switching condition and receiving a fault warning signal exceeding a preset level. In one possible implementation, the area controller is specifically configured to: acquiring monitoring data of various vehicle-mounted devices in the area; performing multi-dimensional verification on the monitoring data, wherein the multi-dimensional verification comprises identity verification, protocol verification, performance verification, address verification, data veri