Model-based Development Of Emedded Racing Vehicle Control Systems: A Formula Student Torque Vectoring Case Study

Abstract

This thesis presents a model-based V-model development cycle for embedded vehicle control unit development in a formula student team. The work focuses on designing and implementing a torque vectoring controller for a fully electric formula student race car, addressing the need for structured software development amid limited time and resources. An adapted V-model development cycle is proposed, integrating model-in-the-loop (MiL), processor-in-the-loop (PiL), and hardware-in-the-loop (HiL) methodologies to systematically define, implement, and verify control software. The developed torque vectoring system utilizes a modular three-stage architecture consisting of a yaw rate reference generator, a high-level yaw moment controller with gain-scheduled PI and feedforward control, and a low-level motor torque distributor. This system was developed in Simulink and deployed on a real-time-enabled Raspberry Pi 5 platform using a PREEMPT_RT Linux kernel. Simulation results using the VI-CarRealTime platform and the X-in-the-loop methods demonstrate that the developed controller improves vehicle yaw rate tracking in selected driving scenarios and delivers measurable lap time improvements when running on track compared to a baseline configuration without torque vectoring. The proposed development approach offers a scalable, simulation-driven pathway for embedded controls in resource-constrained environments and provides actionable insights for formula student teams transitioning to electric vehicle platforms.