The invention belongs to the technical field of Internet of Vehicles communication, and particularly provides a vehicle infrastructure cooperation-oriented BSM data anomaly correction method, which comprises the following steps: acquiring BSM data acquired by vehicle driving, the BSM data comprising a normal BSM and an abnormal BSM; constructing a ProtoNet model used for learning normal BSM features, selecting normal BSM data as a training data set, inputting the training data set into the ProtoNet model, and training parameters of the ProtoNet model; parameters of the ProtoNet model comprise encoder and memory network parameters; and according to normal BSM features learned by the model, an AnRectify model of self-supervised learning is adopted to correct the abnormal BSM. And the correction effect is optimized by using priori knowledge in the ProtoNet model, so that high-precision data correction is realized. According to the method, the accuracy of the BSM data can be effectively improved in a vehicle-road cooperation scene, misleading of abnormal data caused by environmental interference, sensor faults and the like to a driving decision is reduced, and the traffic safety and the system stability in the vehicle-road cooperation environment are improved.
本发明属于车联网通信技术领域,具体提供了一种面向车路协同的BSM数据异常校正方法,包括,获取车辆行驶所采集的BSM数据,BSM数据包括正常BSM与异常BSM;构建用于学习正常BSM特征的ProtoNet模型,选取正常BSM数据作为训练数据集输入ProtoNet模型,对ProtoNet模型的参数进行训练;ProtoNet模型的参数包括编码器和存储器网络参数;根据模型所学习的正常BSM特征,采用自监督学习的AnoRectify模型对异常BSM进行校正。利用ProtoNet模型中的先验知识对校正效果进行优化,实现高精度数据校正。该方法能够在车路协同场景下有效提高BSM数据的准确性,减少因环境干扰、传感器故障等引起的异常数据对驾驶决策的误导,提升车路协同环境中的交通安全性和系统稳定性。
BSM data abnormity correction method for vehicle-road cooperation
一种面向车路协同的BSM数据异常校正方法
2025-02-18
Patent
Electronic Resource
Chinese
Holographic vehicle track correction algorithm for data abnormity of road sensing equipment
European Patent Office | 2023
|Vehicle abnormity monitoring system and method for mine road
European Patent Office | 2024
|Vehicle-road cooperation system and vehicle-road cooperation method
European Patent Office | 2021
|Vehicle abnormity monitoring method based on driving data
European Patent Office | 2022
|