Sparse representation based tracking of frequency hopping primary user for cognitive radio

Cognitive radio has been proposed as a solution to the apparent problem of spectrum scarcity. In this work, a novel technique for tracking a slow frequency hopping spread spectrum signal using sparse representation classification (SRC) over a learned-exemplar dictionary is presented. On comparing with an FFT based spectrum sensing method, the proposed technique is found perform better for a range of low SNR values. The advantage of SRC with the proposed learned-exemplar dictionary over that with simple-exemplar dictionary is also highlighted.