Vehicle Test Platform with Experimentally Correlated Tire Model for Torque Vectoring Control

This paper presents a testing platform for torque vectoring controls utilizing a 10 degree of freedom (DOF) vehicle simulation and a ⅕ scale radio control test vehicle. These vehicle simulations or “Digital Twins” have been widely adopted throughout the automotive industry for their lower operating costs and ease of implementation. Virtual models are not perfect representations of reality, however, and physical testing is still necessary to validate systems for use in the real world. This is especially true when testing safety-critical features such as Advanced Driver Assist Systems (ADAS). As a result, a simulation environment working in conjunction with a test vehicle represents an optimal hybrid approach. In this work, a 1/5 scale radio controlled vehicle with torque vectoring capability is designed and assembled. A high fidelity vehicle model is then constructed in the Matlab/Simulink environment to model test vehicle behavior. The test vehicle features independent suspension for off road capability. As a result, the digital twin is capable of modeling all angular degrees of freedom of the vehicle body. The vehicle model must also estimate wheel forces during high-sideslip maneuvers. The Pacejka Magic Formula is used for its accurate representation of tire behavior in highly transient driving scenarios. MF coefficients are estimated from GPS, inertial, and odometry measurements collected throughout defined test maneuvers. Vehicle model behavior is benchmarked against the test vehicle across all degrees of freedom modeled. An S-curve maneuver is performed in simulation and experimentation to ensure accuracy and consistency across transient and steady state behavior. In future work, focus will turn to creating an ADAS control system which re-stabilizes a vehicle after a collision using torque vectoring.