Development and control of a bicycle robot based on steering and pendulum balancing

This research focuses on the development of an electric bicycle robot which balances itself while moving at a constant forward speed. The bicycle robot is balanced by applying a pendulum mass moving together with heading steering to share the balancing load. The nonlinear dynamics model of a bicycle along with a pendulum is derived from the Euler-Lagrange equation of motion and nonholonomic constraints with respect to translation and rotation relative to the ground plane. This nonlinear model is then linearized around the upright position and combined with the DC motor model to obtain the completed linearized dynamics model. The robot's prototype is designed and built. The linear quadratic regulator (LQR) is implemented on the system to control its balance. The simulation results show that, by varying the control weighting matrix to increase the use of the pendulum's input, the system obtains better performance in both balancing and straight-line path tracking. In addition, the state weighting matrix can be varied to further improve the path tracking performance while still keeping the stability. Real experimental results are also in line with the simulation results.