摘要
本文的主要目的是基于信息融合的方法设计出一套能准确辨识出驾驶行为的系统。本系统使用六轴加速度计采集加速度信息,通过多尺度多重分形(MMA)算法(首次将该算法用作特征提取的方法)从加速度信号中提取出可反映不同驾驶行为的波动特征。并采集电动汽车的OBD接口获取的包括速度、功率、电流等车载OBD信息并提取特征。分别通过随机森林(RF)算法对驾驶员的驾驶行为进行辨识。提出一种新的信息融合的方法,采用该方法对加速度信息和OBD信息进行融合,发现信息融合的方法可以更有效的辨识出电动汽车的驾驶行为。
The goal of this paper is to design a driving behavior identification system in electric vehicle through the information fusion.In the system,the information of acceleration is obtained by installing a six-axis accelerometer in the electric vehicle,followed by feature extraction by multi-scale multi-fractal(MMA)algorithm(it is the first time to use the method to extract features)to characterize the fluctuations in detail,which can distinguish the driver’s driving behavior.The information including speed,power,and current,are also acquired through the OBD interface to assist driving behavior identification.Random forest(RF)algorithm is first used to identify the driving behavior based on acceleration information or OBD information only.In this paper,a method that used to fuse the information is proposed.Then the method is applied to evaluating the driving behavior.It is found that this method can effectively identify the driving behavior in electric vehicle.
作者
陶红兴
莫凌飞
严如强
TAO Hongxing;MO Lingfei;YAN Ruqiang(School of Instrument Science and Engineering,Southeast University,Nanjing 210096,China)
出处
《传感技术学报》
CAS
CSCD
北大核心
2018年第3期355-362,共8页
Chinese Journal of Sensors and Actuators
基金
国家十二五科技支撑计划项目(2015BAG09B01)
关键词
驾驶行为辨识
OBD
MMA算法
信息融合
随机森林
driving behavior identification
acceleration
OBD,MMA
information fusion
random forest
