计及安全性与舒适性的智能车辆换道轨迹规划研究

Lane Change Trajectory Planning of Intelligent Vehicle Considering Safety and Comfort

针对当前智能车辆在换道过程中的安全性以及乘员舒适性问题,本文提出一种基于风险场评估轨迹的二次筛选方法。首先,在Frenet坐标系下,将车辆运动解耦为横向和纵向两个维度,基于五次多项式生成所有横向d-t曲线簇和纵向s-t曲线簇;其次,由车辆的动力学特性和三圆碰撞模型设计轨迹初筛评价函数,选取合格轨迹作为候选轨迹;最后,参考人工势场理论的思想,引入行车过程中风险场的概念,根据换道效率,换道风险值和横、纵向冲击度建立总损失函数评价候选轨迹以进行二次筛选,选取最佳轨迹并完成可视化。为检验算法的可行性,通过搭建双车道的道路环境,设计障碍车不同速度和加速度的多场景进行弯道换道的仿真验证。研究结果表明,本文所提算法能够满足换道的安全性和舒适性需求。同时,在正常换道场景下,乘员在换道过程的97.5%时间处于舒适状态,在紧急避障场景下也能达到平衡安全性与舒适性的目标。

Aiming at improving the safety and passenger comfort of intelligent vehicles during lane change,a secondary screening method based on risk field evaluation trajectory was proposed.Firstly,in Frenet coordinate system,the vehicle motion is decoupled into lateral and longitudinal dimensions,and all lateral d-t curve clusters and longitudinal s-t curve clusters are generated based on quintic polynomials.Secondly,based on the dynamic characteristics of the vehicle and the three-circle collision model,the preliminary screening evaluation function of the trajectory is designed,and the qualified trajectory is selected as the candidate trajectory.Finally,referring to the idea of artificial potential field theory,the concept of risk field in the driving process is introduced,and the total loss function is established to evaluate candidate trajectories according to lane change efficiency,lane change risk value and lateral and longitudinal impact degree for secondary screening,and the optimal trajectories are selected and visualized by coordinate conversion.In order to test the feasibility of the algorithm,a two-lane road environment was built,and multiple scenes with different speeds and accelerations of obstacle vehicles were designed to simulate and verify the curve lane change.The results show that the proposed algorithm can meet the safety and comfort requirements of vehicle lane changing.At the same time,in the normal lane change scenario,the passengers are in a comfortable state in 97.5%of the time during the lane change process,and the goal of balancing safety and comfort can also be achieved in the emergency obstacle avoidance scenario.