Neural-network-based adaptive H∞ position tracking control for a humanoid robot based on neural network

Abstract

System modeling and controller problems are studied for a humanoid with a holonomic omnidirectional platform consisted of three lateral orthogonal wheel assemblies. Based on decentralized-body modeling idea, the relative accuracy of mechanism modeling method is combined with the advantage of connectionism of the neural network for the modeling of this humanoid robot. Using the resulting model, a new neural-network-based adaptive H∞ position tracking controller is proposed. The robust nonlinear H∞ control approach and model direct adaptive neural network technique are integrated together naturally. The closed-loop system robust stability is proved. The simulation results indicate that the proposed methodology is valid and effective.