A new adaptive fuzzy control policy against conventional methods. Statistical analysis of real time control performance

During a real time control process, a set of controllers that use different control policies have been used to control the temperature of a simulated room during 200 minutes under the same conditions. All the realized simulations have been repeated five times with two different sets of set points. A quantitative comparison between the performance of the conventional controllers and the new Adaptive and Self-Organizing Fuzzy Controller proposed in this work will be performed. This latter can adapt and self-organize their internal parameters in real time using just a few qualitative information about the plant (the monotonicity sign and the delay time) and without allocate a rule base. Using all results of the performed simulations as a set of sampling for the statistical tool ANOVA (Analysis Of Variance), we have been able to emphasize the extrapolability of the obtained results (the committed error and the power consumption) and illustrate the great contribution that brings the new Adaptive and Self-Organizing Fuzzy Control policy proposed in this paper.