Robot Manipulation by a New Hybrid Feedback Linearization Adaptive Neuro-Fuzzy Control Method

Although frequently used, intelligent controllers such as neural networks and fuzzy logic have problems dealing with robot manipulators. The large number of DOF and well-being dependency of system parameters on each other and also on input parameters leads to two problems: large number of fuzzy rules; complexity of the system behavior. These problems can be fixed by combining two intelligent methods of adaptive neuro fuzzy and feedback linearization methods. At first, the nonlinear and dependent dynamical equations of the robot are converted to a linear, independent and predetermined system of equations, using feedback linearization method. Then for each state variable, an adaptive neuro fuzzy controller is designed to stabilize the variable. In other words, a MIMO system is converted to some SISO systems and then controlled.This paper investigates the issue of trajectory tracking control of a SCARA manipulator. A new control scheme is proposed based on adaptive neuro-fuzzy network technology and feedback linearization approach for tracking a planned trajectory. The simulation results are provided to verify the effectiveness of the proposed new methodology.