Implementation of a functional link net-ANFIS controller for a robot manipulator

In this paper, a new hybrid network; FLN-ANFIS (functional link net-adaptive neuro fuzzy system), is proposed and used as a controller for robot path tracking purposes. The advantages of the ANFIS controller along with fast learning rate of FLN are combined to propose a new methodology for controlling robot arm movement. Furthermore, a classic PID controller as well as the intelligent controller such as NN, fuzzy and ANFIS are developed and their performances are compared both by simulation and implementation. A 5-bar linked robot is chosen as a test system for implementation. Simulation and implementation of different controllers under various conditions are carried out on this robot. Results show that the designed FLN-ANFIS network for robot control can provide very good performances through a wide range of operations by improving dynamic operations of the system.