基于神经网络自适应MPC智能车辆轨迹跟踪仿真

打开文本图片集

中图分类号：U461 文献标识码：A DOI：10.3969/j.issn.1674-8484.2025.04.014

Abstract：The weight matrixof traditional model predictive control （MPC） controlers usuallyrelies on manual experience for parametertuning，making it dificult toadapt tocomplex dynamic environments.Therefore， a method for adaptiveadjustment of MPC weight matrices based on backpropagation （BP） neural networks was proposed.Firstly，the inteligent vehicle dynamics model with MPC control was established to analyze theinfluenceof diferent weight coefficientson thevehicletrajectory tracking performance，secondly thedata were constructed to train the BP neural network model，and the BP neural network adaptive MPCcontroler was constructed using the Matlab/Simulink module to jointly simulate with Carsim，and finally，a doubleshift simulation condition was designed from diferent speedsand road adhesion coeficients to validate the robustnessof thecontrollerunderdiferent working conditions.Theresults show thatthe BPneural networkbasedadaptive MPC controllerachieves favorable control performance across diferent speeds when the road surfaceadhesion coefficient is O.85.Ataspeedof 65km/h ，thevehicleunder the fixed-weightMPCcontrol approaches destabilization，whereas the root-mean-squares （RMS）of the lateral displacement deviation and lateral angledeviation for the adaptive controllerare reduced by 44.17% and 66.66% ，respectively.The proposed controlleralso exhibits strong performance on road surfaces with varying adhesion coeficients—most notably onslippery roadswith anadhesion coefficient of O.35.When traveling at 30km/h under such conditions， the RMS values of the two deviations are decreased by 27.49% and 49.54% compared to the fixed-weight MPC controller.This neural network-based approach foradaptive adjustment of MPC controllerweightscan provide valuable insights for enhancing trajectory tracking performance in medium-and high-speed cooperative control of intelligentconnected vehicles，as wellas inautonomous navigation systems forspecial operationvehicles.

Keywords：inteligent connected vehicle; neural network;adaptive;trajectory tacking;modelpredictive control （MPC）

随着城市化进程加速，传统的车辆已难以满足现代社会的需求，智能车辆因其高度的感知能力及智能化控制越来越受到人们的重视。（剩余12374字）