Blind adaptive sensor-array processing techniques for extracting highly corrupted radio communication signals.

Abstract

This thesis examines the techniques for blindly extracting highly corrupted radio communication signals with the use of sensor arrays. Initially, existing work is reviewed. The blind signal extraction techniques reviewed do not require known a priori training signals, signals direction of arrival, sensor array geometry or calibration. However they are found generally to be dependent upon some prior signal assumptions made as their basis for separating signals. A new blind adaptive filtering technique called the Self-Adapting sensor array MUlti-Signal blind Extractor, or S-AMUSE, is proposed by the author to provide a general truly blind adaptive filtering framework based on a multi-stage approach. The technique is described in detail and results are presented via Monte Carlo simulations. The performance results obtained are highly promising.