Optimal Fuzzy Supervised PID Controller using Ant Colony Optimization Algorithm

PID controllers are the well known and most widely used controllers in the industries. The reason behind this is because of its simple structure and reliability. Nonetheless when the plant to be controlled is highly non linear or is subjected to disturbances or we have less knowledge about it, under these conditions poor performance is obtained when we are using fixed parameter PID as controller. Thus an expert supervisor is required to online tune its controller parameters. But because of the known shortcomings of a human supervisor such as chances of poor tuning or damage to the plant caused due to carelessness of supervisor, automatic tuning is preferred. Using a fuzzy block for the role of supervisor is a good option because of its known qualities of mimicking human operator and robustness. However an expert is still necessary to determine the parameters of membership functions, fuzzy rules and scaling factors of the fuzzy block. Also a good degree of knowledge about system is also required for tuning parameters of fuzzy block. So we can use of an optimization technique which can tune the parameters of the fuzzy block which in turn gives us a superior performance. Several techniques are available, here we have made use of ACO (Ant Colony Optimization) algorithm which follows the food gathering ability of one of the most successful species on earth i.e. ants. Simulation on MATLAB and SIMULINK show better performance obtained by using above mentioned tuning scheme as compared to well known Zieglar-Nicolas scheme for a second order system.