Multilevel Adaptive Control of Interconnected Nonlinear Systems: Application to Double Coupled Inverted Pendulums

This paper presents an adaptive backstepping-based multilevel approach for controlling nonlinear interconnected systems with unknown parameters. The proposed control system consists of an adaptive backstepping controller at the first level, in which the gains are optimally and automatically tuned using genetic algorithm; and a nonlinear controller at the second level to neutralize the interaction terms. The controller is applied to the uncertain double coupled inverted pendulums which are connected by a spring. The simulation results show that this new method is capable of controlling the system effectively, in both regulation and tracking tasks, and there is no influence of channels on each other.