Traction Control of Heavy Vehicles using Road Information

Abstract

Loss of traction, especially while climbing a hill on low-friction surfaces with heavy vehicles, significantly compromises safety and performance. This thesis presents an approach that integrates future road information, such as friction coefficients and slope, to control vehicle velocity and optimize wheel slip in real time. Using road information data, the controller setup anticipates the wheel speed needed for a reference velocity and gives the necessary torque to be at that wheel speed while also adhering to the constraints. A decoupled architecture with three modules is used to leverage different control systems for different subproblems. Firstly, a differential lock controller decides which differential setting should be used. Secondly, a Model Predictive Controller (MPC) tracks a reference velocity while adhering to constraints. Finally, a Wheel Speed Controller (WSC) is used to give out the necessary torque to control the wheel speed. The decoupling is mainly done to keep the computation times close to real-time implementations for heavy-duty truck platforms. Extensive simulations were done on both synthetic and real-life data to validate the controller pipeline for varying road profiles. Overall, the results indicate that preview based traction control can noticeably improve operational safety and performance for heavy vehicles in off-road and low-friction scenarios.