基于混合整数规划的智能车横纵向一体化滚动优化决策  

Longitudinal and lateral integrated moving horizon decision making of intelligent vehicle based on mixed integer programming

作　　者：戴启坤 刘俊[1,2] 郭洪艳 陈虹[1,3] 孙义[1,2] DAI Qi-kun;LIU Jun;GUO Hong-yan;CHEN Hong;SUN Yi(State Key Laboratory of Automotive Simulation and Control,Jilin University,Changchun Jilin 130022,China;College of Communication Engineering,Jilin University,Changchun Jilin 130022,China;College of Electronic and Information Engineering,Tongji University,Shanghai 201804,China)

机构地区：[1]吉林大学汽车仿真与控制国家重点实验室,吉林长春130022 [2]吉林大学通信工程学院,吉林长春130022 [3]同济大学电子与信息工程学院,上海201804

出　　处：《控制理论与应用》2024年第1期21-29,共9页Control Theory & Applications

基　　金：国家自然科学基金项目(U19A2069,61790563);吉林省科学技术厅项目(20200401088GX,20200501011GX)资助.

摘　　要：本文针对智能车辆的行为决策问题,设计了基于混合整数规划的智能车横纵向一体化滚动优化决策方法.该方法首先将纵向车速表示为非整数,将期望车道表示为整数控制量,建立了混合整数智能车决策简化模型;然后,设计了横纵向一体化滚动优化决策方法,决策出纵向车速和换道动作,根据系统输出与非线性约束的时域关系证明了优化问题的递归可行性并通过遗传算法求解非线性混合整数规划优化问题.基于车辆动力学仿真软件veDYNA和Simulink进行了联合仿真,并在红旗E-HS3智能车上开展了实车试验,结果表明,本文提出的基于混合整数规划的智能车横纵向一体化决策方法能够实现超车、避障、跟车、停车和弯道工况下的行为决策.Aiming at decision making of intelligent vehicle,longitudinal and lateral integrated moving horizon decision making on intelligent vehicle based on mixed integer programming is designed.Firstly,a simplified mixed integer intelligent vehicle decision making model is established,where the longitudinal vehicle speed is expressed as a non-integer,and the desired lane is expressed as an integer input.Then the longitudinal and lateral integrated moving horizon decision making method is designed,which determines the longitudinal vehicle speed and lane change action.The recursive feasibility of the optimization problem is proved based on the time-domain relationship between the system output and the nonlinear constraints,and the nonlinear mixed integer programming optimization problem is solved by genetic algorithm.The simulation verification is carried out in vehicle dynamics simulation software veDYNA and Simulink,and the experiment was carried out on the Hongqi E-HS3 intelligent vehicle.The results show that the longitudinal and lateral integrated moving horizon decision making on intelligent vehicle based on mixed integer programming proposed in this paper can realize decision making in overtaking,obstacle avoidance,following,stopping and curve scenarios.

关 键 词：决策 混合整数规划 模型预测控制 横纵向一体化 决策模型 自动驾驶 

分 类 号：U463.6[机械工程—车辆工程]