Quick Spectrum Access via Block Sampling

This paper considers spectrum access in wideband channels and its objective is to design an agile and reliable mechanism for identifying spectrum opportunities. Driven by the need for reducing the time required for identifying spectrum opportunities, the idea of data-adaptive and sequential block sampling is proposed, through which instead of examining each channel individually, a cognitive user takes samples that are linear combinations of simultaneous measurements from multiple channels. If such coarse samples indicate that the block of channels contains at least a vacant (unused) channel, then the channels are examined individually in order to accumulate more information about their spectral occupancy states, and otherwise, the entire block is discarded and the process resumes sequentially by examining the next block of channels.