基于轨迹-速度耦合策略的复杂道路汽车行驶速度决策

发布时间：2018-05-03 12:24

本文选题：人车路系统 + 智能驾驶　；参考：《西南交通大学学报》2015年04期

【摘要】：为了提供复杂道路上汽车自动驾驶的目标速度,提出了基于前视轨迹曲率的速度决策算法:首先决策出前视断面的轨迹点,算出每个点位的轨迹曲率并将其作为输入数据;然后从时间最省、驾驶最舒适、定速巡航以及混合模式中选择其一作为决策目标;在速度界限以及纵向/侧向舒适性约束下进行滚动时域优化,决策出前视断面上的期望速度值,随着车辆的行驶前视断面依次向前滚动,最终得到沿行驶距离变化的速度曲线.用复杂山区道路的实测数据验证了模型的精度,以2条公路和1条赛道作为仿真算例,结果表明:(1)通过组合不同的轨迹决策目标和速度决策目标,能够得到多种驾驶模式的速度曲线,进而模拟出多种实际驾驶行为;(2)由于前视轨迹是在通道边界内生成,弯道半径、转角、回旋线、路宽、偏转方向等用于确定通道边界的道路变量,都会改变轨迹特性进而影响行驶速度,因此,本文算法能够适应复杂的道路几何条件.
[Abstract]:In order to provide the target speed of auto driving on complex roads, a speed decision algorithm based on forward trajectory curvature is proposed. Firstly, the trajectory points of forward looking sections are determined, and the trajectory curvature of each point is calculated and used as input data. Then choose one of the most time-saving, the most comfortable driving, fixed speed cruising and mixed mode as the decision goal, and optimize the rolling time domain under the speed limit and longitudinal / lateral comfort constraints. The expected velocity value on the forward looking section is determined and the velocity curve varying along the driving distance is obtained by rolling forward with the forward looking section of the vehicle in turn. The accuracy of the model is verified by using the measured data of complex mountain roads. The simulation examples of two highways and one track show that: 1) by combining different trajectory decision targets and speed decision targets, The speed curve of various driving modes can be obtained, and then many actual driving behaviors can be simulated. (2) because the forward view track is generated within the channel boundary, the radius of the bend, the turning angle, the roundabout, the width of the road, The deflection direction and other road variables used to determine the channel boundary will change the trajectory characteristics and then affect the speed of the road. Therefore, the algorithm in this paper can adapt to the complex road geometry conditions.
【作者单位】：重庆交通大学山地城市交通系统与安全重庆市重点实验室;中铁二院工程集团有限责任公司;西南交通大学交通运输与物流学院;
【基金】：国家自然科学基金资助项目(51278514) 高等学校博士学科点专项科研基金资助项目(20135522110003)
【分类号】：U491.25

【参考文献】