基于新能源汽车大数据的事故特征模式匹配追踪分析

针对传统新能源汽车事故分析存在受限因素多、相互关联复杂度高、资源消耗大、事故原因较难下定论的问题，以新能源汽车全生命周期数据为例，提出一套事故数据模式特征匹配追踪的方法。通过构造高斯高阶导数滤波器，对模板信号的高阶微分特征进行抽取并形成多维特征模板，再用滤波器组与目标信号作卷积，将两者作相关性运算，计算其相关系数，把相关系数理解成概率并作为匹配程度的度量，计算联合概率最终完成模板信号的匹配程度计算，实现目标信号的匹配追踪。验证表明该方法科学有效，能够区分正常车辆和事故车辆，可以为新能源汽车安全预警提供有效依据。

Feature Tracking and Pattern Matching for Accident Cars Based on Big Data of New-Energy Vehicles

In the traditional accident analysis of new energy vehicles, there are many restricted and complicated factors, high resource consumption and difficultiesin determining the causes of accidents. Based on the full life cycle data of new energy vehicles, a method of feature tracking and pattern matching for accident cars was proposed. Firstly by constructing a high-order Gaussian derivative filter, the derivative features of the template signal were extracted to form a multi-dimensional feature template.And then the target signal was convolved with a bank of the filters and the correlation coefficient was calculated.The correlation coefficient could be considered as the probability value and a measure of matching degree.Finally the joint probability was calculated to obtain the matching degree of the template signal and to achieve the tracking of the target signal. The verification results show that the proposed method can effectively distinguish the accident vehicles from the normal vehicles and provide a basis for safety pre-warning of new energy vehicles.