面向多工况的智能车换道轨迹规划方法

Trajectory planning for lane changing of intelligent vehicles under multiple operating conditions

针对智能汽车在结构化道路上换道时存在的可行性和安全性问题,分不同工况进行轨迹规划的研究.在简单的无障碍换道时,提出一种新的改进余弦换道模型,使换道轨迹曲率连续、过渡平滑.当需要避障换道时,采用多项式来描述换道轨迹,提出基于评价指标的边界条件选取.同时,通过预测交通车的行驶轨迹,进行双5次多项式规划以避开交通车.在弯道换道时,也采用多项式的方法,考虑坐标的转化,得到以初始坐标系为基准的边界条件.仿真结果表明:提出的轨迹规划方法较已有算法在效率、舒适性、安全性等指标上具有良好的换道性能.

To ensure the feasibility and safety of lane changing for intelligent vehicles on structured roads, the trajectory planning was investigated under multiple operating conditions. For simple barrier-free lane changing, a newly modified cosine transformation model was proposed to make the path curvature continuous with smooth transition. When there were traffic vehicles in the driving path with needing to avoid obstacles, a polynomial was used to describe the trajectory for lane changing. The selection of boundary conditions was based on the evaluation indicators. The trajectory of traffic vehicle was predicted by the double five polynomial programming to avoid the traffic vehicle. For lane changing in a curve, polynomial methods were also used. Considering the transformation of coordinates, the boundary conditions were obtained based on the initial coordinate system. The simulation results show that compared with the existing algorithms,the proposed trajectory planning methods have excellent performance of lane changing in efficiency, comfort, safety and other indicators.