当前位置: X-MOL 学术Mech. Syst. Signal Process. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
A robust lateral tracking control strategy for autonomous driving vehicles
Mechanical Systems and Signal Processing ( IF 7.9 ) Pub Date : 2021-03-01 , DOI: 10.1016/j.ymssp.2020.107238
Wei Zhang

Abstract A robust steering torque control strategy for lateral tracking functionality of autonomous driving vehicles that perform active steering, active accelerating and active braking is researched in this paper. The main target of this research is to track references of lateral position and heading angle, which are provided by upstream motion planning module, via torque that robotic arm applies to steering hand wheel. Firstly, a system with tracking errors is generated and analyzed. To keep control system’s robustness against time varying parameters, such as speed, center of mass, etc., gain-scheduling approach is utilized to obtain proper feedback gain. To achieve performance robustness, methods of linear matrix inequalities are used in design to maintain a performance function. Control performance is validated in Hardware in the loop environment built by ETAS labcar and Matlab/Simulink for a lane changing scenario and two typical parking scenarios. Results show that the proposed control strategy can regulate vehicle to follow target trajectory precisely even when speed varies. In lane changing scenario, steady state error is close to zero and maximum lateral position error is about ±30 cm. In parking scenarios, tracking error of lateral position is about ±3 cm, and of heading angle is about ±0.1 rad at end points. In addition, in comparison with linear quadratic regulator (LQR) and model predictive control (MPC), the proposed control outperformances in the three test scenarios.

中文翻译:

一种鲁棒的自动驾驶车辆横向跟踪控制策略

摘要 本文研究了具有主动转向、主动加速和主动制动功能的自动驾驶车辆横向跟踪功能的鲁棒转向转矩控制策略。本研究的主要目标是通过机械臂施加在方向盘上的扭矩来跟踪上游运动规划模块提供的横向位置和航向角的参考。首先,生成并分析具有跟踪误差的系统。为了保持控制系统对时变参数(如速度、质心等)的鲁棒性,利用增益调度方法来获得适当的反馈增益。为了实现性能鲁棒性,在设计中使用线性矩阵不等式的方法来维持性能函数。控制性能在 ETAS labcar 和 Matlab/Simulink 为车道变换场景和两个典型停车场景构建的环路环境中的硬件中得到验证。结果表明,即使在速度变化时,所提出的控制策略也可以调节车辆精确地遵循目标轨迹。在换道场景中,稳态误差接近于零,最大横向位置误差约为±30 cm。在停车场景中,横向位置的跟踪误差约为±3 cm,终点的航向角约为±0.1 rad。此外,与线性二次调节器 (LQR) 和模型预测控制 (MPC) 相比,所提出的控制在三个测试场景中表现优异。结果表明,即使在速度变化时,所提出的控制策略也可以调节车辆精确地遵循目标轨迹。在换道场景中,稳态误差接近于零,最大横向位置误差约为±30 cm。在停车场景中,横向位置的跟踪误差约为±3 cm,终点的航向角约为±0.1 rad。此外,与线性二次调节器 (LQR) 和模型预测控制 (MPC) 相比,所提出的控制在三个测试场景中表现优异。结果表明,即使在速度变化时,所提出的控制策略也可以调节车辆精确地遵循目标轨迹。在换道场景中,稳态误差接近于零,最大横向位置误差约为±30 cm。在停车场景中,横向位置的跟踪误差约为±3 cm,终点的航向角约为±0.1 rad。此外,与线性二次调节器 (LQR) 和模型预测控制 (MPC) 相比,所提出的控制在三个测试场景中表现优异。1 rad 在端点。此外,与线性二次调节器 (LQR) 和模型预测控制 (MPC) 相比,所提出的控制在三个测试场景中表现优异。1 rad 在端点。此外,与线性二次调节器 (LQR) 和模型预测控制 (MPC) 相比,所提出的控制在三个测试场景中表现优异。
更新日期:2021-03-01
down
wechat
bug