Explicit MPC for column-type EPS systems

Abstract

For decades, electric power steering (EPS) systems have replaced hydraulic-assist steering systems. The main purpose of EPS systems is to generate an assist torque using an electric motor for the improvement of the driving comfort. The proposed controller is designed against a model predictive control (MPC) and linear quadratic regulator (LQR) for a comparison in terms of generating a desired value of the assist torque. The motion equations of a column-type EPS system are derived in this study. Meanwhile, MPC has been regarded as an effective control method for constrained problems because this method can anticipate future events of the system. However, the repeated online optimisation imposes huge computational complexity. The proposed scheme explicitly obtains the current control law by adopting a multi-parametric quadratic programming technique. The proposed controller is evaluated in MATLAB/Simulink and a hardware-in-the-loop simulation using the Infineon automotive TriCore AURIX TC277 microcontroller. The simulation and experimental results confirmed that the proposed controller successfully tracks the reference value while reducing the computational burden. It is also shown by conducting experiments with various vehicle speeds that the tracking ability of the explicit model predictive controller (MPC controller) is superior to that of the LQR controller.