摘要
对智能汽车避障过程进行研究,提出衍生状态格的概念,并在此基础上确定避障轨迹规划算法。该算法结合避障车辆和障碍物的运动状态和位置信息,将复杂道路环境中期望避障轨迹的求取问题转换为避障车辆和状态格之间的轨迹规划问题。由相关算法生成的避障轨迹能针对性地考虑障碍物的状态并适用于多障碍物环境,提出状态格成本的概念,再采用Dijkstra搜索算法在多条可行避障轨迹中进行寻优。仿真结果表明,汽车在障碍规避过程中横摆角速度和侧向加速度值均符合稳定性要求,验证了该算法的可行性。最后通过CarSim/LabView硬件在环试验对该方法进行了进一步的验证,所得结论与仿真结果基本一致。
The obstacle avoidance process of intelligent vehicle is studied,a concept of derived state lattice is proposed,and on this basis,a trajectory planning algorithm for obstacle avoidance is determined.The algorithm combines the movement state and location information of target vehicle and obstacles,and converts the problem of seeking the desired obstacle avoidance trajectory in complex road environment into a trajectory planning one between target vehicle and state lattice.The obstacle avoidance trajectory generated by the corresponding algorithm is suitable for multi-obstacle environment with pertinent consideration of the state of obstacles,another concept of state lattice cost is also put forward,and the Dijkstra search algorithm is used to find the optimum from several feasible obstacle avoidance trajectories.The results of simulation show that both the yaw rate and lateral acceleration of vehicle meet the requirements of stability in its obstacle avoidance maneuver,verifying the feasibility of the algorithm.Finally,a CarSim/LabView hardware-in-the-loop test is also conducted to further verify the method proposed,with a conclusion basically agrees with simulation results.
作者
胡延平
田博
陈无畏
张锐陈
Hu Yanping;Tian Bo;Chen Wuwei;Zhang Ruichen(School of Mechanical Engineering,Hefei University of Technology,Hefei230009;School of Automotive and Traffic Engineering,Hefei University of Technology,Hefei230009)
出处
《汽车工程》
EI
CSCD
北大核心
2019年第8期967-974,共8页
Automotive Engineering
基金
国家自然科学基金(U1564201和51675151)
安徽省科技重大专项(17030901060)
中央高校基本科研业务费专项资金(JZ2016HGBZ1009)资助
关键词
状态格
避障
轨迹规划
状态格成本
搜索算法
state lattice
obstacle avoidance
trajectory planning
cost of state lattice
search algorithm