Development of a small-scale electric vehicle for evaluation of torque vectoring and MPC path tracking

Abstract

In this thesis, a small-scale electrical vehicle with independently driven rear wheels is developed to implement and test torque vectoring and MPC using microcontrollers. The project investigated how torque distribution affects stability, maneuverability and trajectory tracking during representative driving maneuvers. A scale RC car chassis was purchased and equipped with custom mechanical components, electric motors, electronic speed controllers, battery, microcontrollers and an internal measuring unit to form a closed-loop control platform. The control system combined MPC using the TinyMPC library for path tracking with a PID-controller based torque vectoring strategy to influence yaw behavior. The system was evaluated through simulations and practical testing, where vehicle trajectories were extracted from video recordings using image processing.

The results showed that torque vectoring generated noticeable but moderate handling improvements. However, performance was limited by hardware constraints, simplified modeling, incomplete PID-controller tuning, limited torque authority and sensor limitations. Nevertheless, the project demonstrates that torque vectoring can improve controllability on a low-cost small-scale electric vehicle platform, as well as validating that MPC can be implemented on limited hardware, such as microcontrollers. It also shows that further development is required for more robust and quantitatively reliable results. Recommended improvements include systematic controller tuning, improved state estimation, onboard data logging, slip-angle control implementation, PWM-to-torque mapping, and more capable actuators.