Compressive oversampling using circulant matrices for lossy wireless channels

Compressive sensing has been explored over the past decade in many fields as a data compression, encryption, channel estimation, outlier detection technique etc. One such application of compressive sensing is improving robustness of communication in the presence of data losses in wireless channel. The basic idea is taking more number of measurements (oversampling) than the optimal number based on the lossy channel. These extra measurements help in original signal recovery at the receiver although some measurements are lost. Random matrices (like Gaussian distributed, Bernoulli etc.) are the most popular measurement/sensing matrices used in compressive sensing. But generating measurements using random matrices requires more operations resulting in slower data acquisition. Hence in this paper we have shown the performance analysis of compressive oversampling using circulant measurement matrices, which have faster implementation than random matrices, for the application of sparse signal recovery in the presence of data loss in wireless channels.