一种基于基准点映射的驾驶行为分析

An Analysis of Driving Behavior Based on Reference Point Mapping

在汽车驾驶行为分析中,一般通过车载传感器收集数据,并对数据进行处理,但是车载传感器收集的数据具有很多弊端,比如具有数据量大、无法分辨方向等。本文提出一种基于基准点映射的驾驶行为分析算法:根据原始行驶记录,使用一定规则和Geohash算法建立带有方向的基准点,并将行驶记录投影到对应的基准点中,采用k-Means聚类算法对基准点数据进行分类,减少有效数据点个数,为后续危险驾驶行为分析提供基础。

In the car driving behavior analysis,data collection and data processing are generally performed by means of vehicle sensors.However,there are many drawbacks for the collected data by vehicle sensors,such as a large amount of data and indistinguishable directions.This paper proposes a driving behavior analysis algorithm based on reference point mapping:According to the original driving record,a certain rule is used to establish the reference point with direction.The driving record is projected into the corresponding reference point classified by k-Means clustering to reduce the numbers of effective points,which is prepared for the subsequent cluster analysis and the dangerous driving behavior analysis.