基于改进人工势场法的车辆换道轨迹规划

Vehicle Lane Change Trajectory Planning Based on Improved Artificial Potential Field Method

摘要: 针对当下智能车辆在换道过程中的安全性及舒适性问题，提出了一种基于人工势场法的轨迹筛选方法。首先，基于五次多项式生成所有换道轨迹簇，并以车辆的动力学特性对轨迹簇进行初次筛选，留下符合运动学特性的轨迹；其次，在传统人工势场模型的基础上增加形状系数对该模型进行改进，基于改进的人工势场模型分别对换道环境中的障碍车辆及道路边界等建立安全势场；最后以换道过程中的风险值及横向冲击度建立损失函数对余下轨迹进行二次筛选，得到最优换道轨迹。为验证其可行性，以双车道结构化道路作为换道环境，以主动换道和强制换道作为换道场景。结果表明提出的轨迹规划算法可以满足换道安全性和舒适性的要求。

Abstract: A trajectory screening method based on the artificial potential field approach is proposed to address safety and comfort issues in intelligent vehicle lane changes. Initially, all lane change trajectory clusters are generated using quintic polynomials, and then screened based on vehicle dynamic characteristics to retain trajectories that conform to kinematic characteristics. Additionally, a shape factor is incorporated into the traditional artificial potential field model for further improvement. This improved model establishes safety potential fields for obstacle vehicles and road boundaries in the lane changing environment. Finally, a loss function incorporating risk value and lateral impact during lane change is used for secondary screening of remaining trajectories, resulting in optimal lane change trajectories. To validate its feasibility, a two-lane structured road is utilized as the lane changing environment with active and forced lane changing scenarios. The results demonstrate that the proposed trajectory planning algorithm meets requirements for both safety and comfort during lane changes.