基于改进Stanley算法的无人车路径跟踪融合算法研究

Research on Fusion Algorithm of Unmanned Vehicle Path Tracking Based on Improved Stanley Algorithm

针对斯坦利(Stanley)跟踪算法无法更好地同时满足无人驾驶路径跟踪的精确度和平滑性要求的问题,根据车辆的航向角、横向偏差、车速等特性,基于合适的预瞄距离,采用纯跟踪(Pure Pursuit)算法对Stanley算法中车轮转角的计算方式进行改进,提出一种新的融合算法,实时计算车辆在当前车速下合适的车轮转角。仿真结果表明,相比于Stanley算法,所提出的融合算法在不失跟踪精确度的情况下,不同车速下跟踪平滑性均有较大提升。实车试验结果表明,在20 km/h车速下,所提出融合算法的跟踪路径比原Stanley算法的跟踪路径有更好的精确度和平滑性。

For the problem that Stanley tracking algorithm cannot better meet the accuracy and smoothness requirements of unmanned driving path tracking simultaneously,this paper used pure pursuit algorithm based on the appropriate preview distance,to improve the calculation method of wheel angle in Stanley algorithm according to the characteristics of vehicle heading angle,lateral deviation and vehicle speed,and proposed a new fusion algorithm,which could real-time calculate the appropriate wheel angle of the vehicle at the current speed.The simulation results show that compared with Stanley algorithm,the proposed fusion algorithm can greatly improve the tracking smoothness under different vehicle speeds without compromising the tracking accuracy.The real vehicle test results show that the tracking path of the proposed fusion algorithm has better accuracy and smoothness than that of the original Stanley algorithm at the speed of 20 km/h.