Adaptive control of autonomous bicycle kinematics

Abstract

In this paper, an adaptive Self-Tuning Regulator is proposed to control an autonomous bicycle kinematics. The model used for describing the dynamics of the bicycle is Whipple's bicycle model. The roll (lean) and the steer angle of the bicycle the two outputs of the model and the torques across the roll and steer angle as the two control variables. The identification technique used in here is recursive least square identification and the proposed control strategy is state feedback control with the controller gain calculated using linear quadratic regulator. The autonomous bicycle was tested for varying velocities and it was observed that the adaptive controller gives a good control action of the bicycle model not just within the stable velocity region but outside it as well.