Frequency-Diversity RCS Based Target Recognition with ICA Projection

Abstract

In this paper, a frequency-diversity radar cross section (RCS) based target recognition scheme with independent component analysis (ICA) projection is proposed. The goal is to identify the similarity between the unknown and known targets through collected frequency-diversity RCS. Note that the unknown target means the test target and that known targets mean previously seen targets in a database. To enhance the recognition ability, frequency-diversity RCS data are projected into the ICA space, and the recognition is performed using features of ICA space. The ability to tolerate noise effects for proposed recognition scheme is also investigated. The frequency-diversity technique can greatly reduce the efforts of collecting RCS because only a small number of measuring locations are required to achieve accurate recognition. With the use of ICA projection, the recognition scheme will have good abilities in both discriminating targets and tolerating noise effects.