Applied dynamic system modeling - Six degree-of-freedom simulation of forced unsteady maneuvers using recursive neural networks

Abstract

Techniques have been developed to computationally simulate six-degree-of-freedom forced unsteady maneuvers for a real vehicle using recursive neural network (RNN) technologies. The approach taken was to develop RNN 6-DOF maneuvering simulations using existing free-flight radio controlled model maneuvers. A modular design strategy was adopted which was comprised of component subsystems coupled within the recursive neural network architecture. In particular, semi-empirical component models for the propulsion and plane forces were developed. These time-varying models of the component forces were then coupled directly within the RNN 6-DOF algorithm. The RNN techniques are described in detail, and results using the combination of semi-empirical component models and RNN 6-DOF simulation techniques are described. It is concluded that RNN 6-DOF maneuvering simulations can provide accurate predictions of vehicle maneuvers, including maneuvers which are dominated by forced unsteady fluid dynamics. Across large numbers of maneuvers, the results indicate that these techniques provide accurate predictions for both maneuvers used to develop the RNN 6-DOF simulation and for validation maneuvers comprised of novel control sequences. (Author)