Abstract: This paper discusses radar High Resolution Range Profile(HRRP) recognition. The kernel atching pursuit algorithm, a computational efficient kernel machine training algorithm, is used to design the classifier. The power spectrum of HRRP, a translation-invariant feature, is used as the feature vector for recognition. Experimental results based on measured data show that the proposed technique achieves good recognition performance and low computational complexity.

Key words: high resolution range profile, kernel matching pursuit, translation-invariant feature, template matching method