DYNAMIC CONTROL OF ROBOTIC MANIPULATORS USING ARTIFICIAL NEURAL NETWORK

Abstract

The aim of the thesis is to develop an efficient dynamic control such as point to point and continuous path control strategy for Robot manipulator using Artificial neural network (ANN) and Adaptive Neuro fuzzy inference system (ANFIS),owing to the advantage of their learning ability & unique characteristics, which enables to control Robot manipulators In this work, a robot controller based on neural network is presented. This controller has been applied to a single link robot arm and three link SCARA (Selective Compliance Assembly Robot Arm) which has a highly nonlinear structure. The model based approaches for robot control (such as the computed torque technique) require high computational time and can result in a poor control performance, if the specific model-structure selected does not properly reflect all the dynamics. In addition, conventional-PD controller could not cope up with unmodeled dynamics. Moreover, Fuzzy logic can be used to map complex nonlinear relations by a set of IF-THEN rules. The membership functions are designed by intuitive human reasoning. This poses three problems. One, for different control application a new set of membership functions have to developed and second, latent stability problem and third, once these membership functions are developed and implemented there is no means of changing them. This means fuzzy logic lacks a learning function. Neural networks on the other hand self-organize the mapping relationship by learning. A dynamic model has been assumed here where a controller is associated with each joint and separate RBF (Radial Base Function) neural networks are used as assistants to the PD controllers in order to minimize tracking errors. All above features naturally allow one to consider investigating the feasibility of neural networks. The proposed ANFIS methodology combines artificial neural networks with fuzzy logic. The fuzzy sets are used to formalize the level of human perception of the physical system. The neural networks, on the other hand, perform all the necessary computations and with regard to their learning capabilities, they enable an adaptation of the existing controller through its learning to the changes in the system behavior. In this work, the ANFIS (Adaptive Neuro-Fuzzy inference system) for the dynamic control (point-to-point as well as continuous path control) of the three-link SCARA manipulator is designed. This new method for control combines th advantages of neural networks (learning and adaptability) with the advantages of fuzzy logic (use of expert knowledge) to achieve the goal of robust control of robot dynamic systems. Simulation results show very good tracking performance. In addition, a visual display of three-link SCARA manipulator is made by using C++. Further, in this work a practical approach to implement the Neuro. -Fuzzy technique has been discussed for future