A New Fuzzy Control and Dynamic Modeling of Bicycle Robot

Abstract

A dynamic model of bicycle robot is established based on the Lagrange method, and its motion features, which include nonlinear and parameter time-varying, are analyzed. A new fuzzy control is proposed to make the bicycle robot system achieve favorable control effects including good dynamic and steady-state performance. This new fuzzy control can be divided into two components. In the first part, the fuzzy control method, which is on the basis of the individual riding experience, is used to promote the dumping bicycle to swing back to the equilibrium position. In the second part, the adaptive fuzzy PID method is employed to eliminate static error and guarantee the bicycle robot no vibration near the equilibrium position. Concerning reliability of switching process, the fuzzy method of smooth switching is used to guarantee the steady transition between these two different control strategies. Since it is difficult to select the initial values of the adaptive fuzzy PID method, a modified particle swarm optimization (MPSO) method is utilized to optimize the initial parameters off-line. To show high efficiency and the accuracy of this proposed algorithm, simulation results demonstrate that the stability of bicycle robot can be guaranteed by using this design scheme.