汽车工程 ›› 2024, Vol. 46 ›› Issue (10): 1780-1789.doi: 10.19562/j.chinasae.qcgc.2024.10.006

• • 上一篇    下一篇

基于自主漂移的自动驾驶车辆极限工况轨迹规划与控制

卢少波1(),代灵峰1,王晨辉1,刘丙军2,褚志刚1,谢文科1   

  1. 1.重庆大学机械与运载工程学院,重庆 400030
    2.长安汽车股份有限公司,重庆 400023
  • 收稿日期:2024-05-20 修回日期:2024-07-16 出版日期:2024-10-25 发布日期:2024-10-21
  • 通讯作者: 卢少波 E-mail:lsb@cqu.edu.cn
  • 基金资助:
    国家自然科学基金(51675066);重庆市科技创新与应用发展专项(CSTB2023TIAD-STX0039)

Trajectory Planning and Control of Autonomous Vehicle Under Extreme Conditions Based on Autonomous Drift

Shaobo Lu1(),Lingfeng Dai1,Chenhui Wang1,Bingjun Liu2,Zhigang Chu1,Wenke Xie1   

  1. 1.College of Mechanical and Vehicle Engineering,Chongqing University,Chongqing  400030
    2.Chongqing Changan Automobile Co. ,Ltd. ,Chongqing  400023
  • Received:2024-05-20 Revised:2024-07-16 Online:2024-10-25 Published:2024-10-21
  • Contact: Shaobo Lu E-mail:lsb@cqu.edu.cn

摘要:

为兼顾自动驾驶车辆在极限工况下的稳定性与轨迹跟踪性能,提出了一种基于自主漂移的自动驾驶车辆轨迹规划与控制方法。基于神经网络设计了神经网络轮胎动力学模型,提升了传统魔术轮胎公式的精度。为进一步拓展自动驾驶车辆极限工况下的稳定边界,基于漂移时轮胎饱和及最大侧滑特性结合质心侧偏角-横摆角速度相平面约束设计了漂移稳定边界,采用非线性模型预测控制(NMPC)在更大稳定范围内规划了安全漂移轨迹,并对规划轨迹进行了漂移跟踪控制。Simulink/CarSim联合仿真结果表明,该方法可充分利用漂移运动优势,在极限工况下确保车辆不发生失控,同时准确跟踪期望轨迹。

关键词: 极限工况, 轨迹跟踪, 稳定性控制, 神经网络轮胎模型, 非线性模型预测控制

Abstract:

To consider both stability and trajectory tracking performance of autonomous vehicles operating in extreme conditions, a trajectory planning and control method based on autonomous drift is proposed. A neural network tire dynamics model is designed based on neural network to improve the accuracy of the traditional magic tire formulation. In order to further expand the stability boundaries under the extreme working conditions of autonomous vehicles, the drift stability boundaries are designed based on the tire saturation and maximum sideslip characteristics combined with the center-of-mass lateral deflection angle-transverse swing angular velocity phase plane constraints during drift, and the nonlinear model predictive control (NMPC) is used to plan a safe drift trajectory within a wider stability range, and the drift tracking control is carried out for the planned trajectory. The results of the joint simulation of Simulink/CarSim show that the method can fully utilize the advantages of drift motion to ensure that the vehicle does not go out of control under extreme working conditions, while accurately tracking the desired trajectory.

Key words: extreme conditions, trajectory tracking, stability control, neural network tire model, nonlinear model predictive control