An application of evolutionary fuzzy modeling to spacecraft fuzzy controller synthesis

Evolutionary fuzzy modeling (EFM), its application in building complex nonlinear fuzzy control systems, and robustness and optimality related issues have been one of our research focus. EFM permits the integration of subjective expert rules with data driven newly discovered rules and ensures robust performance using a relatively small number of rules. Performance and robustness comparisons of fuzzy controller synthesis using manually generated rules, and EFM evolved fuzzy rules and membership function parameters have indicated the usefulness of EFM. In this study, a single axis (pitch angle) fuzzy controller is first synthesized using EFM for spacecraft attitude control. A 3-axis fuzzy controller is then built from the synthesized single axis controller. Performance comparisons using ad hoc and GA evolved control parameters are then presented. We first introduce the need for robust controller synthesis techniques such as EFM for spacecraft control. The EFM schedule and its related issues are then discussed. This is followed by details of applying EFM to space station 3-axis attitude control. The paper concludes with a discussion on the observed results and further directions