摘要
针对驾驶风格辨识问题,提出了一种基于主客观融合的分类辨识方法。首先,构建驾驶风格主观量表,并映射形成主观风格因子;其次,构建数据采集系统获取自然驾驶数据,并对跟驰及换道行为场景进行挖掘;然后,提取行为关键特征与主观风格因子共同构成驾驶风格主客观融合辨识体系,基于主成分分析(PCA)及K均值方法对样本集进行降维及无监督聚类,并对聚类结果进行验证;最后,引入带有物理意义的具体特征参数的分布情况对激进型、中庸型及保守型驾驶风格进行辨识。结果表明,主观风格因子与客观行为特征参数的分布一致性较好,该融合方法可对驾驶风格进行有效辨识。
For the issue of driving style identification,this paper proposes a method based on subjective and objective fusion.Firstly,the subjective style scale is developed and transformed into subjective factors.Secondly,a data acquisition system is implemented to gather the natural driving data and the basic behavior scenarios of car following and lane changing are mined.Thirdly,the fusion of subjective and objective identification system of driving style is constructed by extracting the key behavioral characteristics and subjective factors.Based on Principal Component Analysis(PCA)and K-means methods,the sample set is processed and the clustering results are verified.Finally,specific characteristics with physical significance are introduced to identify radical,moderate and conservative driving styles.The results reveal a strong correlation between the distribution of subjective style factors and objective behavioral characteristics,and the fusion method can be used to identify driving style effectively.
作者
张慧
周景岩
付会通
邢智超
Zhang Hui;Zhou Jingyan;Fu Huitong;Xing Zhichao(CATARC(Tianjin)Automotive Engineering Research Institute Co.,Ltd.,Tianjin 300300;China Intelligent and Connected Vehicles(Beijing)Research Institute Co.,Ltd.,Beijing 100176)
出处
《汽车工程师》
2024年第11期13-19,共7页
Automotive Engineer
关键词
驾驶风格
主观风格因子
自然驾驶数据
行为关键特征
主客观融合
Driving style
Subjective factor
Natural driving data
Key behavioral characteristic
Fusion of subjective and objective
