Combined model- and rule-based controller synthesis with application to helicopter flight

This thesis deals with synthesis of combined (nonlinear) model-based and (fuzzy logic) rule-based controllers, along with their applications to helicopter flight control problem. The synthesis involves superimposing two control techniques in order to meet both stability and performance objectives. One is model-based control technique, which is based on inversion of an approximate model of the real system. The other is rule-based control technique that adaptively cancels the inversion errors caused by the approximate model inversion.;There are two major aspects of the research effort in this thesis. The first is the development of the adaptive rule-based (fuzzy logic) controllers. The linguistic rule weights and defuzzification output weights in the controllers are adapted for ultimate boundedness of the tracking errors. Numerical results from a helicopter flight control problem indicate improvement and demonstrate effectiveness of the control technique.;The second aspect of this research work is the extension of the synthesis to account for control limits. In this thesis, a control saturation related rule-bank in conjunction with the adaptive fuzzy logic controller is designed to trade-off system performance for closed-loop stability when the tendency towards control amplitude and/or rate saturation is detected. Simulation results from both a fixed-wing aircraft trajectory control problem and a helicopter flight control problem show the effectiveness of the synthesis method and the resulting controller in avoiding control saturations.