Abstract: In order to improve the trajectory tracking accuracy of unmanned vehicles, an LQR trajectory tracking control method for unmanned vehicles based on improved quantum genetic algorithm (IQGA) was proposed. Firstly, in order to solve the problems of significant steady-state error and insufficient predictability of the LQR controller, a feedforward module and a prediction module were designed. Secondly, addressing the challenge of selecting weight parameters for LQR controllers, the quantum genetic algorithm was employed to optimize and select these parameters. Then, addressing the issues in quantum genetic algorithms where the rotation angle of the quantum rotation gate was fixed, which made the algorithm easy to fall into the local optimal value and the lack of population diversity, an adaptive dynamic rotation angle adjustment strategy and a discrete coefficient determination mechanism were designed. Finally, the improved LQR control algorithm (IQGA-LQR) was tested for double lane change trajectory tracking on the Simulink-Carsim co-simulation platform. The simulation results show that compared with the traditional LQR, GA-LQR and QGA-LQR algorithm, the peak lateral errors of the IQGA-LQR algorithm are reduced by 63.1%, 50.15% and 39.44% respectively; and root mean square errors of the IQGA-LQR algorithm are reduced by 50.68%, 41.03% and 34.86%, respectively.

Key words: vehicle engineering; QGA; IQGA; LQR; trajectory tracking control; unmanned driving

摘要： 为了提高无人驾驶车辆的轨迹跟踪精度，提出一种基于改进量子遗传算法的无人驾驶车辆LQR轨迹跟踪控制方法。首先，针对LQR控制器存在稳态误差较大以及预见性不足的问题，设计了前馈模块以及预测模块。其次，针对LQR控制器权重参数难以选取的问题，使用量子遗传算法对参数进行优化选取。然后，针对量子遗传算法中量子旋转门的旋转角度固定导致算法容易陷入局部最优值以及种群多样性不够丰富的问题，设计了自适应动态旋转角度调整策略以及离散系数判定机制。最后，改进后的LQR控制算法(IQGA-LQR算法)在Simulink-Carsim联合仿真平台上进行了双移线轨迹跟踪测试，仿真结果表明：相较于传统LQR、GA-LQR、QGA-LQR算法，IQGA-LQR算法的横向误差峰值分别降低了63.10%、 50.15%、 39.44%，均方根误差分别降低了50.68%、41.03%、34.86%。

关键词: 车辆工程；QGA；IQGA；LQR；轨迹跟踪控制；无人驾驶