Automated calibration of a fuzzy logic controller using a cell state space algorithm

A method for automatically fine-tuning the output function parameters of a fuzzy logic controller based on the cell mapping concept is presented. The method takes a computational approach to the analysis of phase-space-based information about the global behavior of the system. The desired optimal control of a system is determined with a cell-state-space-based optimal control algorithm. Output function parameter modification is accomplished through application of an error gradient estimation algorithm. The optimal control algorithm generates an optimal control table. The optimal control table is used by the gradient descent algorithm to identify the values of the output function parameters that best match the desired optimal control. Using this method, a given fuzzy rule base can be calibrated automatically. A demonstration of the method is shown for the time-optimal set point control of a DC motor.<>