Filter Bank Approach to the Estimation of Flexible Modes in Dynamic Systems.

摘要

The problem of estimating frequencies of sinusoids buried in noise has been of great interest in both military and civilian applications. In particular, in Control Systems with flexible appendages the sinusoidal vibrations can cause instabilities and degrade the performance of the overall system. In this thesis the problem of identifying frequencies of disturbances in flexible systems using advanced Digital Signal Processing techniques such as filter banks and Quadrature Mirror Filters is addressed. In a number of situations there is a need to design a controller for a system with flexible modes. In particular, in space applications solar panels and robotic arms introduce flexible modes in the system which degrades the performance. In these kinds of applications, the frequencies of the flexible modes can not be modeled accurately a priori and they can change according to the operating conditions. The proposed approach is tested by computer simulations.,,,,,,Satellite systems are being deployed on a regular basis these days with various missions, namely civilian or military. Among them, at a great percentage, are communications and observation. However, it has been observed that for these cases the existing vibrations from the flexible surfaces of the satellite and the atmospheric interferences add colored noise and consequently undermine the system's performance. In this thesis an approach to estimate frequencies in noise based on Filter Banks and Quadrature Mirror Filters is developed. The general idea is to decompose the signal into components at various frequency bands and compute the relative energy in the single bands.