采用变时域模型预测的车辆路径跟踪控制方法

为了解决自动驾驶车辆变速行驶时模型预测路径跟踪控制器的可靠性问题,提出一种变预测时域自适应路径跟踪控制方法.首先,推导简化后适用于仿真验证的车辆三自由度动力学模型,引入松弛因子以避免求解过程中出现非可行解,并将跟踪控制转化为二次规划求解问题;然后,确定模型预测控制器的重要设计参数,分析车速和预测时域的变化关系,拟合预测时域与车速的函数曲线,设计一种变时域自适应路径跟踪控制器;最后,搭建Carsim/Matlab/Simulink联合仿真平台进行验证.结果表明:变时域自适应路径跟踪控制器能够随着车速变化实时更新预测时域,可保证车辆具有较好的跟踪精度和稳定性.

Vehicle Path Tracking Control Method Using Varying Horizon of Model Predictive Control

In order to solve the reliability problem of the model predictive path tracking controller for autonomous vehicles with variable speed, a varying prediction horizon adaptive path tracking control method was proposed. First, a simplified vehicle three freedom degrees dynamic model was deduced that can be applied to simulation and verification,the relaxation factors were introduced to avoid infeasible solutions when solving problems, and the path tracking control was transformed into quadratic problems. Then, the model predictive controller’s important design parameters were confirmed, the varying relationship between vehicle speed and prediction horizon was analyzed, the function curve of prediction horizon and vehicle speed was fitted to design a varying prediction horizon adaptive path tracking controller. Finally, a Carsim/Matlab/Simulink co-simulation platform was established for verification. The results show that the varying horizon adaptive path tracking controller can update the prediction horizon in real-time as vehicle speed changes, and ensure good tracking accuracy and stability of vehicles.