基于多约束随机模型预测控制的无人车运动规划与控制

Path Tracking Method of Intelligent Vehicle Based on Multi-Constrained Stochastic Model Predictive Control

摘要:

提出了一种基于随机模型预测控制的无人车运动规划方法。在道路坐标系下，采用质点运动模型和高斯分布对周围动态车辆的预测轨迹进行位置不确定性表征，并使用随机模型预测控制（SMPC）中的机会约束进行描述，以此建立车辆空间位置的约束。通过变步长求解方式获得基于运动学模型的初始控制序列。基于此初始解，考虑车辆动力学信息，引入基于横摆角速度和质心侧偏角关系的稳定性约束，求解最优控制量。在多种工况下，通过仿真试验验证了所提出方法的有效性和稳定性。

Abstract:

A path tracking method of intelligent vehicle based on stochastic model predictive control is proposed. The predicted trajectories of surrounding dynamic vehicles are characterized by positional uncertainty using a prime motion model and Gaussian distribution in the road coordinate system， and described using chance constraints in stochastic model predictive control （SMPC） as a way to establish constraints on the spatial location of the vehicles. The initial control sequence based on the kinematic model is obtained by means of a variable step size solution. Based on this initial solution， a stability constraint based on the relationship between the angular velocity of the transverse pendulum and the lateral eccentricity of the center of mass is introduced by considering the vehicle dynamics information to solve the optimal control volume. The effectiveness and stability of the proposed method are verified by simulation tests under various operating conditions.