Decoupling of Longitudinal and Lateral Control in Autonomous Vehicles Using Model-Based Predictive Control

EasyChair Preprint 11126

Abstract

This work addresses the development of a decoupled lateral and longitudinal controller for autonomous vehicles (AVs). The present work uses a model-based predictive controller (\textit{Model Predictive Controller}, MPC) to perform the lateral control and a proportional-integral (PI) controller to perform the longitudinal control of the vehicle. Lateral control aims to keep the vehicle on the desired trajectory. The longitudinal controller intends to keep the vehicle at a fixed speed, previously established, generating the desired acceleration from the reference longitudinal velocity. The trajectories and the reference velocity are generated by a trajectory-planner that uses a receding-horizon control strategy based on mixed integer-quadratic programming (MIQP). The evaluation of the proposed method was performed using the MATLAB/SIMULINK tool. Simulation experiments consider a maneuver of a vehicle traveling along a one-way road in the presence of obstacles.

Keyphrases: autonomous vehicle, mixed integer quadratic programming, predictive control, proportional-integral control

@booklet{EasyChair:11126,
  author    = {Ícaro B. Viana and João P. S. Rodrigues and Alan M. da Rocha and Marcelo M. S. de Souza},
  title     = {Decoupling of Longitudinal and Lateral Control in Autonomous Vehicles Using Model-Based Predictive Control},
  howpublished = {EasyChair Preprint 11126},
  year      = {EasyChair, 2023}}