Modeling, Parameter Identification, and Steering Angle Control of Electric Power Steering for Trajectory Following in Autonomous Vehicles

This paper presents modeling, parameter identification, and control of the Electric Power Steering (EPS) in the autonomous driving mode (driverless mode). The model consists of three parts, the first part estimates the mechanical system of the EPS by a second order one degree of freedom model, the lumped parameters of the mechanical model, moment of inertia, damping and coulomb friction constant are identified using Recursive Least Squares (RLSE). The second part of the model is the estimation of the dynamics of the electric motor by a second order transfer function using (RLSE). The third part of the model is the estimation of the aligning moment torque resulting from road disturbance. The aligning moment is estimated as a function of lateral forces, pneumatic trail, and kinematic trail. Lateral forces are estimated as a function of tire slip angle. Extended Kalman filter (EKF) is implemented to estimate the vehicle side slip angle and yaw rate to be used in the tire slip angle estimation. The model is used in control of the Steering Wheel Angle (SWA) of the EPS for trajectory following. The model is also used to provide a complete steering actuation model for simulation purposes. The control method used to control the EPS angle is Sliding Mode Control (SMC). Side slip angle and aligning moment were validated experimentally in the vehicle, estimated side slip angle showed comparable results with Oxford INS/GPS. Simulation results proved excellent tracking of the SMC SWA controller and compensation for the aligning moment. Simulations were conducted using the EPS model constructed from real time data.