Abstract:

This paper proposes a weighted feature reduction method based on Fisher’ discriminant ratio(FDR) for a time-shift invariant feature, power spectrum, in radar automatic target recognition using the high-resolution range profile(HRRP). The proposed weighted feature reduction method uses the FDR vector of the target power spectrum to iteratively search for an optimal weight vector, and reduce feature dimensionality according to the optimal weight vector. Compared with using the raw power spectrum feature and some existing feature reduction methods based on Fisher’s linear discriminant, the proposed weighted feature reduction method can not only reduce the feature dimensionality, but also improve the recognition performance with low computation complexity. In the recognition experiments based on measured data, the proposed method is robust to different test data and achieves good recognition results.

Key words: radar automatic target recognition(RATR), high-resolution range profile(HRRP), power spectrum, feature reduction, Fisher's discriminant ratio(FDR)