Handling and Stability Analysis of an Autonomous Vehicle Using Model Predictive Control in a CarSim–Simulink Co-Simulation Environment

Abstract

Cars are a prevalent mode of transportation for both people and goods, with B-class hatchbacks being particularly popular in Indonesia. However, road traffic crashes remain a major concern, contributing millions of deaths annually, primarily due to human error. Autonomous vehicles offer a promising solution to mitigate these issues by reducing reliance on human control. In particular, Level 3 autonomous vehicles enhance road safety, enable independent mobility, reduce traffic congestion, and allow drivers to engage in non-driving tasks. This study proposes an autonomous vehicle model that employs a trajectory tracking approach using Model Predictive Control (MPC), a robust and widely adopted control strategy in autonomous systems. A three-degree-of-freedom (3-DOF) vehicle dynamic model was developed and analyzed through co-simulation using CarSim and Simulink to evaluate its performance during a double-lane change maneuver. The simulation results demonstrate that the vehicle accurately follows the reference trajectory and exhibits excellent dynamic performance. The roll angle remained consistently low, ranging between 0.024 and 0.026 radians—well below the rollover threshold of 0.14 radians—demonstrating strong roll stability. The slip angle varied between –0.013 and 0.0135 radians, nearly 12 times lower than the critical limit, indicating optimal traction and directional control. Lateral acceleration ranged from –3.59 m/s² to 3.41 m/s², and yaw rate remained within –7.78°/s to 7.25°/s, both well within safe operational bounds. These findings confirm that the proposed MPC-based control framework enables precise path tracking, robust stability, and reliable handling performance in dynamic driving scenarios.