基于曲率自适应LTV-MPC的轨迹跟踪控制研究  

Trajectory Tracking Control Method Based on Curvature-adaptive LTV-MPC

作　　者：臧豫徽 杨爱喜 李兰友 ZANG Yuhui;YANG Aixi;LI Lanyou(School of Mechanical and Automotive Engineering,Anhui Polytechnic University,Wuhu 241000,China;Geely Automotive Institute,Hangzhou Vocational&Technical College,Hangzhou 310018,China;Zhejiang Asia Pacific Intelligent Network Automobile Innovation Center Co.,Ltd.,Hangzhou 311203,China)

机构地区：[1]安徽工程大学机械与汽车工程学院,安徽芜湖241000 [2]杭州职业技术学院吉利汽车学院,浙江杭州310018 [3]浙江亚太智能网联汽车创新中心有限公司,浙江杭州311203

出　　处：《安徽工程大学学报》2025年第1期22-30,共9页Journal of Anhui Polytechnic University

基　　金：浙江省科技计划项目(2022C04023)。

摘　　要：为了解决自动驾驶车辆在轨迹跟踪曲率变化时跟踪轨迹精度下降和行驶稳定性的问题,基于车辆的动力学三自由度模型建立状态空间方程,设计了一种道路曲率自适应线性时变模型预测控制(LTV-MPC)轨迹跟踪控制器。通过联合搭建Matlab/Simulink与Carsim仿真平台开展对比实验,仿真实验中首先分析对比了LTV-MPC控制器在不同的Q、R权重矩阵下跟踪精度的变化规律,随即拟合出随道路曲率变化的权重系数函数。最后经仿真实验对比分析,该曲率自适应LTV-MPC轨迹跟踪控制器相比LTV-MPC轨迹跟踪控制器,转向处平均横向跟踪误差减小30%,最大质心侧偏角减小2.9%,轨迹跟踪精度和车辆行驶稳定性提高。To solve the problems of trajectory tracking accuracy degradation and driving stability of autonomous driving vehicle when the trajectory tracking curvature changes,this paper designs a road curvature-adaptive linear time-varying model predictive controller(LTV-MPC)trajectory tracking controller based on the dynamic three-degree-of-freedom model of the vehicle.By building Matlab/Simulink and Carsim co-simulation platform for experiments,the paper first analyzes and compares the LTV-MPC in the different Q and R weight matrix coefficients under the tracking accuracy of the change rule,and then fits them with the road curvature changes in the weight coefficient function.Finally,after the simulation experiment comparison and analysis,the curvatureadaptive LTV-MPC trajectory tracking controller,compared with LTV-MPC trajectory tracking controller,the lateral tracking error is reduced by 30% in average,the maximum sideslip angle is reduced by 2.9%.Thus,the trajectory tracking accuracy and vehicle driving stability are improved.

关 键 词：自动驾驶 轨迹跟踪 模型预测控制 曲率自适应控制器 

分 类 号：U461[机械工程—车辆工程]