基于模型预测的智能汽车轨迹跟踪控制  

Track Tracking Control of Smart Vehicle Based on Model Prediction

作　　者：周君武 杨航 覃立仁 梁国全 ZHOU Jun-wu;YANG Hang;QIN Li-ren;LIANG Guo-quan(Dongfeng Liuzhou Automotive Co.,Ltd.;Institute of Machinery and Automotive Engineering,Guangxi University of Science and Technology,Liuzhou 545000,China)

机构地区：[1]东风柳州汽车有限公司 [2]广西科技大学机械与汽车工程学院,广西柳州545000

出　　处：《汽车电器》2022年第7期1-3,共3页Auto Electric Parts

摘　　要：针对智能驾驶汽车轨迹跟踪问题,本文验证在五次多项式工况下,模型预测控制的轨迹跟踪效果。本文建立车辆运动学模型,为了便于建立基于模型预测的轨迹跟踪控制器,将所建立的非线性车辆运动学模型线性化,再通过向前欧拉法将系统离散化,得到基于线性时变的预测模型。为了使汽车可以快速且平稳地跟踪目标轨迹,建立包含系统状态量和控制增量的目标函数。最后在Matlab/Simulink中对设计的轨迹跟踪器在五次多项式工况下进行测试,与前轮反馈控制(Stanley)对比,验证此工况下所建立的基于模型预测控制的轨迹跟踪器与Stanley控制相比,可以更准确地跟踪期望轨迹。To solve the problem of intelligent driving vehicle track tracking,this paper verifies the track tracking effect of model predictive control under quintic polynomial working conditions.This paper establishes the vehicle kinematics model,in order to establish a track tracking controller based on model prediction,linearizes the established non-linear vehicle kinematics model,and then discretizes the system by forward Euler method to get a prediction model based on linear time-varying.In order for the car to track the target track quickly and smoothly,an objective function containing the system state and control increment is established.Finally,the designed trajectory tracker is tested in Matlab/Simulink under quintic polynomial working conditions,and compared with the front wheel feedback control(Stanley),it is verified that the model predictive control based trajectory tracker under this working condition can track the desired trajectory more accurately than the Stanley control.

关 键 词：智能汽车 运动学模型 模型预测 轨迹跟踪 

分 类 号：U463.6[机械工程—车辆工程]