Comparison of MPC controller parameter for autonomous lane change

Kanan, Muhammad Nur Amin (2022) Comparison of MPC controller parameter for autonomous lane change. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Comparison of MPC controller parameter for autonomous lane change.pdf - Submitted Version Download (2MB)

Abstract

The system that is related to the autonomous vehicle steering system is a Multi-Input MultiOutput (MIMO) system. Due to the interplay between the inputs and outputs, it is not easy to construct a controller for this system using a conventional controller. Next, designing the best controller requires a larger parameter tuning on a controller to ensure the system fully functions. As a result, implementing an MPC controller will solve this problem, This project aim is to investigate and compares the effect of varying MPC parameters (prediction horizon, control horizon and sample time) on the vehicle model's lateral position and yaw angle by tuning the MPC controller toolbox in MATLAB with the default response of yaw angle and lateral position compared with tuned response. An MPC controller linear model for an Autonomous Vehicle Steering System is generally used to direct an autonomous vehicle's steering system to realize the main objective. As a result of this project's analysis, we discovered that changing the parameters on the prediction horizon that produce a high rate of change for the first 2.5 seconds, which is 1.05 persent for Lateral Position and for Yaw Angle 0.02 persent for sample time at the final value of the graph is the best way to achieve the most optimised tuning, and the result can find one of the best parameters to refine the most optimised controller and which is the closest value to the value of the output reference.

Actions (login required)

View Item