Fuzzy and recurrent neural network motion control among dynamic obstacles for robot manipulators

An integration of fuzzy controller and modified Elman neural networks (NN) approximation-based computed-torque controller is proposed for motion contr ol of autonomous manipulators in dynamic and partially known environments c ontaining moving obstacles. The fuzzy controller is based on artificial pot ential fields using analytic harmonic functions, a navigation technique com mon used in robot control. The NN controller can deal with unmodeled bounde d disturbances and/or unstructured unmodeled dynamics of the robot arm. The NN weights are tuned on-line, with no off-line learning phase required. Th e stability of the closed-loop system is guaranteed by the Lyapunov theory. The purpose of the controller, which is designed as a neuro-fuzzy controll er, is to generate the commands for the servo-systems of the robot so it ma y choose its way to its goal autonomously, while reacting in real-time to u nexpected events. The proposed scheme has been successfully tested. The con troller also demonstrates remarkable performance in adaptation to changes i n manipulator dynamics. Sensor-based motion control is an essential feature for dealing with model uncertainties and unexpected obstacles in real-time world systems.