欢迎访问《汽车安全与节能学报》，
汽车安全与节能学报

JASE ›› 2019, Vol. 10 ›› Issue (3): 326-333.DOI: 10.3969/j.issn.1674-8484.2019.03.008

• 汽车安全 • 上一篇    下一篇

基于模型预测控制的无人驾驶汽车横纵向运动控制

谢  辉，刘爽爽   

  1. （天津大学 机械工程学院，天津 300072，中国）
  • 收稿日期:2019-03-26 出版日期:2019-09-30 发布日期:2019-10-01
  • 作者简介:第一作者：谢辉 (1970—)，男( 汉)，天津，教授。E-mail: xiehui@tju.edu.cn。 第二作者：刘爽爽(1993—)，女( 汉)，山东，硕士研究生。E-mail: 2016201143@tju.edu.cn。
  • 基金资助:

    天津市科技计划项目(17ZXRGGX00140)。

Lateral and longitudinal motion control of unmanned vehicles using model predictive control

XIE Hui, LIU Shuangshuang   

  1. (School of Mechanical Engineering, Tianjin University, Tianjin 300072, China)
  • Received:2019-03-26 Online:2019-09-30 Published:2019-10-01

PDF

1423

摘要：

        针对具有高度非线性、强耦合的无人驾驶汽车运动控制问题，提出了一种基于模型预测控制 (MPC)的横纵向综合控制方法。参考速度曲线由参考路径弯曲度确定，再通过分层纵向控制器，实 现速度跟踪。通过MPC，用纵向上位控制器来计算期望加速度；基于逆纵向动力学模型，用下位控 制器来协调驱动和制动。基于运动学模型，以当前系统状态和纵向速度预测序列为输入，用横向控制 器来预测出系统状态变化，以便求解车辆前轮转角。仿真结果表明：与参考速度和参考路径相比，纵 向速度均方根误差为0.086 km/h，横向位置均方根误差为94 mm。因而，该方法可实现车辆运动的 有效控制。

关键词: 无人驾驶汽车 , 运动控制 , 协调控制 , 模型预测控制（MPC）

Abstract:

 A coordinated lateral and longitudinal control method was developed by using a model predictive control (MPC) to solve the motion control with high nonlinearity and strong coupling for unmanned vehicles. According to the curvature of the reference path, the reference speed profile was determined, and then tracked by a layered longitudinal controller. An upper longitudinal controller was used to calculate the desired acceleration through the MPC, while a lower controller was used to coordinate the driving and braking by an inverse-longitudinal-dynamic-model. A lateral controller was used to predict the changes of the system state based on the kinematic model to solve the vehicle front- wheel steer- angle when taking the current system state and a longitudinal-velocity-prediction-sequence as the inputs. The simulation results show that the root mean square error is 0.086 km/h for longitudinal vehicle speed compared to the reference speed, and is 94 mm for lateral position compared to the reference path. Therefore, this method achieves an effective control of vehicle motion.

Key words: unmanned vehicles , motion control ,  coordinated control , model predictive control (MPC)