FAST ALGORITHMS FOR SEPARATING AND TRACKING SOURCES IN ADAPTIVE ARRAYS

Abstract

The use of adaptive array processing techniques to enhance desired directional signals and reject or detect unknown interferences is well known. However, many conventional algorithms still suffer from various limitations in enhancing directional signals and will not function properly in situations when the desired signal is strong and in motion or when there are multiple signal sources. In this thesis, a few new algorithms for separating and tracking multiple directional sources are proposed and investigated. After reviewing the general theory of adaptive signal processing and some relevant research topics, a new structure and algorithm is presented to separate uncorrelated directional sources by employing an adaptive matrix beamformer whose nulls are adjusted to track individual sources in an adaptive manner. An ex tension of the basic zero tracking technique for arbitrary two­ dimensional arrays is then carried out, resulting in an algorithm which tracks the environment in a time-multiplexed manner by solving two sets of linear equations periodically to calculate the preprocessing needed. Finally, a new adaptive algorithm that estimates both the optimal weight and its rate of change to achieve better tracking in non-stationary environments is investigated.