Decentralized adaptive control for robot manipulators based on neural network

A decentralized adaptive control for trajectory tracking of robot manipulators is presented. The system is considered as a set of nonlinear subsystems with nonlinear uncertainties and interconnections. The tracking problem is tackled with decentralized controller. For each subsystem, an output feedback linearization is employed, and neural networks are used to compensate the system errors, the disturbances and the interconnections, which can simplify the design of the controller, improve the dynamic performance and make the system robust. Simulation results show that the system has a good tracking performance.