A Novel Method for Robust Control Using Taguchi Method and Genetic Algorithm in QFT Controller

This paper proposes a novel method to design and optimize a robust controller for a SCARA robot using quantitative feedback theory (QFT). In every physical system, there are number of factors that cause uncertainty in the performance. A robot arm is an example of such systems. Although QFT design technique has been successfully used for plants having structured parameter uncertainty, there are some difficulties that a designer encounters. In this paper we investigated the effects of parameter uncertainties of a SCARA robot on frequency response of open loop system. Taguchi’s experimental design technique is used for determination of the uncertain parameters, which have the greatest influence on the outcome through a very limited number of experiments. With consideration of important parameters, the next step in QFT design procedure is loop-shaping. In the presented method the controller is designed directly by choosing and optimization of coefficients of transfer function by using genetic algorithm. In optimization procedure, stability and bounds of the system were considered as the constraints of the problem. Non-linear simulations on the tracking problem are performed and the results highlight the success of the designed controllers. The results indicate that applying the proposed technique successfully overcomes the obstacles to robust control of non-linear SCARA robots.