摘要
开展了基于大数据识别驾驶员驾驶风格的方法,建立驾驶风格识别数据库,包含80名驾驶员并覆盖不同性别、年龄、驾龄、驾驶习惯等属性,从数据库中提取能够反映驾驶风格的工况,包括换道、转弯、跟车等7种工况总计万余条工况数据,最后利用K均值聚类方法和D-S证据理论决策融合方法进行聚类分析,训练并测试了驾驶风格识别模型。经过验证,所提出的驾驶风格识别方法查准率达到80%。
In this paper,driving style recognition method based on big data was researched,and a driving style recognition database was established,which included 80 drivers covering different attributes such as gender,age,driving age,driving habits,etc.Then,tens thousands of data in seven driving conditions,which could reflect driving style were extracted from the database,including lane change,turning,vehicle following and so on.Finally,the K-means clustering method and D-S evidence theory decision fusion method were used for cluster analysis.And the driving style recognition model was trained and tested.After verification,the precision rate of the recognition method proposed is up to 80%.
作者
吴振昕
何云廷
于立娇
付雷
陈盼
Wu Zhenxin;He Yunting;Yu Lijiao;Fu Lei;Chen Pan(Intelligent Connected Vehicle Development Institute of China FAW Group Co.,Ltd.,Changchun 130011)
出处
《汽车技术》
CSCD
北大核心
2018年第10期10-15,共6页
Automobile Technology
关键词
驾驶风格识别
工况辨识
机器学习
决策融合
Driving style recognition
Vehicle operating modes identification
Machine learning
Decision fusion
