Adaptive control system for flapping wing aircraft: A gain study on optimal control modification

Abstract

Prior work developed a model for a bird-like flapping wing aircraft using Modified strip theory for aerodynamics, a flight dynamics model based on Newtonian equations and an adaptive flight controller. This model was simulated in MATLAB/Simulink using existing data for the Slow Hawk Ornithopter. One of the control architectures developed was an Optimal Control Modification (OCM) of an Adaptive Neural Network Inverse Controller, which was able to successfully damp oscillations in system response in addition to adapting to modeling error and uncertainties. This work continues by performing a Gain Study on this OCM architecture with a PI controller. The proportional Gain (Kp), Integral Gain (Ki), Learning rate and the damping constant were optimized using the error metric introduced by Stepanyan6. In this error metric study, Kp was varied from 10 to 100 and a value of 35 was selected. Similarly, Ki was varied from 10 to 105 and a value of 10 was selected, and a good learning rate of 50000 was found.