Active disturbance rejection control applied to automated steering for lane keeping in autonomous vehicles

The automated steering for lane keeping is an important technique used in the autonomous vehicles. To achieve satisfactory lane keeping performance, the steering control scheme is required to be robust against the vehicle uncertainties and the external disturbances, and be easy to be implemented. To this end, the active disturbance rejection control (ADRC) is applied to the steering controller design in this paper. The ADRC scheme estimates the vehicle uncertainties and the external disturbances in real-time and then compensates them actively. The stability analysis based on the concept of Lyapunov exponents shows that the ADRC control system is exponentially stable around the equilibrium point. Simulations under varying vehicle parameters and disturbances show that the ADRC scheme is able to keep the vehicle within the lane with maximum lateral offset of 0.1 m. Finally, the ADRC scheme is implemented on a scale vehicle. Some necessary electronic devices are installed into the scale vehicle to realize the automated steering. The comparisons between the ADRC scheme and the PID controller show that the scale vehicle controlled by the ADRC scheme performs better with maximum 0.03 m and 0.16 m lateral offsets during the straight and curved lane keeping, respectively. It is also shown that the ADRC scheme is able to perform the lane change. Since the ADRC scheme and the software developed in this paper are independent on a specific platform, they can be implemented on a full-size vehicle.