Direct Adaptive Control for Underactuated Mechatronic Systems using Fuzzy Systems and Neural Networks: A Pendubot Case

This paper describes the implementation of a direct adaptive control of a nonlinear underactuated mechatronics system known as the Pendubot robot using fuzzy systems and neural networks. A PD fuzzy controller is employed to control the two links motion from the free hanging position to the vertical position (the swing-up controller). Then, an intelligent adaptive fuzzy radial Gaussian neural networks system is used to control the Pendubot at the vertical position (the balancing controller) by five rules only in case of parameters uncertainty. This algorithm is proven to be globally stable, with errors converging to a neighbourhood of zero. Finally, the simulation results confirm the theoretic analysis