Piecewise affine control for autonomous bicycle using gyroscopic effect

Abstract

This thesis proposes the new idea of autonomous bicycle control using gyroscopic effect by piecewise affine control method. We considers the naturally unstable autonomous bicycle system at constant forward and rotational speeds. The bicycle is attached with a gyroscopic flywheel acting as an actuator for roll angle stabilization. The system parameters are measured from the adult size bicycle body and some parameters are obtained indirectly from CAD program. The nonlinear model of the bicycle system is approximated by a set of linear model or piecewise affine models which minimizes the model error even outside the operating regions. To stabilize this nonlinear system, we use the piecewise affine model to design the feedback controller. The controller synthesis problem is cast as a Linear Matrix Inequalities problem. The feasible feedback control gain is derived based on a globally quadratic Lyapunov function to guarantee the system stability for all regions. The simulation confirms the effectiveness of this approach.