识别驾驶人意图的车道偏离防避有界控制  

Driver's Intention Identification Based Lane Departure Avoidance Bounded Control

作　　者：王其东[1,2,3] 汪选要 黄鹤[3] WANG Qi-dong;WANG Xuan-yao;HUANG He(School of Mechanical Engineering, Anhui University of Science and Technology, Huainan 232001, Anhui, China;Department of Mechanical Engineering, Hefei University, Hefei 230601, Anhui, China;School of Automotive andTraffic Engineering, Hefei University of Technology, Hefei 230009, Anhui, China)

机构地区：[1]安徽理工大学机械工程学院,安徽淮南232001 [2]合肥学院机械工程系,安徽合肥230601 [3]合肥工业大学汽车与交通工程学院,安徽合肥230009

出　　处：《中国公路学报》2018年第3期105-115,共11页China Journal of Highway and Transport

基　　金：国家自然科学基金项目(51175135;51405004)

摘　　要：为了防止车辆偏离车道导致交通事故的发生和避免车道偏离防避系统(Lane Departure Avoidance Systems,LDAS)对驾驶人行为不必要的干预,提出基于中心区操纵特性阈值法和基于D-S(Dempster-Shafer)证据理论的车辆偏离车道驾驶人意图识别准则,并运用CarSim/Simulink联合仿真对比2种识别准则的有效性。建立转向盘角速度为输入的车路模型,设计LDAS滑模转向控制器,基于预瞄点的侧向偏移量和横摆角速度设计LDAS的期望横摆角速度观测器,并与基于道路曲率和预瞄点侧向偏移量的期望横摆角速度的LDAS进行性能对比。运用相平面法确定保证LDAS车辆稳定性的前轮转向角最大值,并基于CarSim/LabVIEW RT硬件在环试验平台验证基于BP神经网络训练获得D-S证据理论的初始概率赋值的驾驶人意图决策算法的有效性。结果表明:所提出的识别准则能够及时识别车辆偏离车道时的驾驶人意图,为LDAS控制器介入赢得了宝贵的时间,所设计的期望横摆角速度观测器具有很好的稳定性,所提出的方法能够有效避免车辆偏离车道。In order to prevent traffic accidents resulted from the lane departure,and to avoid unnecessary intervention on driver＇s behavior for lane departure avoidance systems（LDAS）,identifying criteria of the driver＇s intention for the lane departure were proposed based on method of threshold value of on-center handling characteristics and D-S evidence theory.And the effectiveness of the two identification criteria were compared by dint of CarSim/Simulink combined simulation.A vehicle-road dynamic model taking steering wheel angular rate as an input was established,whilst a sliding mode steering controller was designed for LDAS.A desired yaw rate observer was designed based on the lateral deviation at look-ahead point and yaw rates,and the performance of LDAS of desired yaw rates based on the road curvature and lateraldeviation at look-ahead point was compared.The maximum value of the front wheel steering angle for guaranteeing the vehicle lateral stability of LDAS was determined based on the phase plane method.The effectiveness of driver＇s intention decision algorithm based on the BP neural network training was verified based on CarSim/LabVIEW RT hardware in loop test to get initial probability assignment of the D-S evidence theory.The proposed identifying criteria can timely identify driver＇s intention when lane departure occurs,and it gains valuable time for the controller intervention of LDAS.The designed yaw rate observer presents good stability.Furthermore,the proposed method can effectively avoid the lane departure.

关 键 词：交通工程 车道偏离 D-S证据理论 识别驾驶人意图 滑模转向控制 中心区操纵特性 

分 类 号：U461[机械工程—车辆工程]