Neuro-fuzzy trajectory tracking control for the Nexus 4-wheeled omnidirectional mobile robot

This work presents the design of a single-layer neuro-fuzzy controller to solve the trajectory-tracking problem of an omnidirectional mobile robot with 4 Mecanum wheels. The controller consists of a PD with fuzzy-neural compensation, which does not require knowledge of the plant and is a robust controller that can deal with disturbances. Unlike a conventional neural controller, in the proposed neuro-fuzzy controller a vector of fuzzy products is implemented as the activation function of the neural network. Due to the computational cost that this implies, the controller is designed by subsystems. The neural network weights are updated online using filtered error and adaptive laws. Experimental results are presented on the Nexus 4-wheeled omnidirectional mobile robot, which verify the effectiveness of the proposed controller.