Abstract—The structure frequency bending modes are the limiting factor to achieve a stability margins and performance of the flight vehicle Autopilots. This is due to a lightly damping structure of these bending frequency modes. The structure engineer could evaluate these frequencies at different design operating points according to the locations of rate gyro sensors on the vehicle beam. Previously, the classical adaptive notch filter can be designed in order to overcome the instability which may occur due to uncertainty of these modes. In this paper, the modified adaptive notch filter is presented in which the classical integral of frequency tracking system in classical adaptive notch filter is replaced by a Radial Basis Function Neural Network for integrator. This modification attempts to enlarge the tracking range of the classical adaptive filter so that the frequency modes with large uncertain values can be gain stabilized. Finally, to demonstrate the objective of this study work, the Simulation results for a typical flexible dynamic model are included at the end of this paper.

Index Terms—Flexible model dynamics, high frequency modes, adaptive filter, structure adaptive filter, neural network.

Ali Elmelhi is with University of Tripoli, Faculty of Engineering, Tripoli-Libya (e-mail: ali_elmulhi@ yahoo.com).

 

Cite:Ali Elmelhi, "Modified Adaptive Notch Filter Based on Neural Network for Flexible Dynamic Control," International Journal of Computer and Electrical Engineering vol. 6, no.2, pp. 185-190, 2014.