On the performance comparison of compressed sensing based detectors for sparse signals Compressive detectors for sparse signals

Compressed sensing (CS) explores sparsity nature of a given signal in a given domain and allows the entire signal to be determined from only a few measurements than that required by the Nyquist sampling rate. This advantage of CS framework is utilized in the design of sparse signal detectors. Specifically compressive and sub-space compressive detectors are the various CS based sparse signal detectors. In this paper we compare the detection performances of traditional, compressive and sub-space compressive detectors using Monte Carlo simulations. For both known and unknown parameter vector cases in the sparasity models, these simulations have been performed. The sub-Nyquist number of measurements required by the compressive and subspace compressive detectors to achieve same performance as that of a traditional detector are found out using known parameter vector sparsity model. For unknown parameter vector in the sparsity model case, the performance of the detectors is compared for various values of signal to noise ratio (SNR). The results show that sub-space compressive detector with fewer measurements than a compressive detector, provides better detection performance than the latter.