Compressive, collaborative spectrum sensing for wideband Cognitive Radios

One of the primary tasks of a Cognitive Radio (CR) is to monitor a wide spectrum and detect vacant channels for secondary transmission opportunities. However, the requirement of prohibitively high sampling rates to monitor a wideband, makes this a challenging task. In this paper, we present a novel wideband spectrum sensing model that reduces the sampling requirement to a sub-Nyquist rate, proportional to the number of occupied channels in the wideband spectrum. The sampling scheme is efficiently implementable using low-rate analog-to-digital converters (ADCs). The sensing algorithm uses techniques borrowed from theoretical computer science and compressive sampling, to detect the occupied channels with a high probability of success. The algorithm is implementable for spectrum sensing in a single CR, as well as in a decentralized CR-network with minimal communication between one-hop neighbors. We provide theoretical expressions for probability of detection and run-time requirements of the scheme. Our simulations show that the proposed scheme exhibits a performance similar to a Nyquist-rate energy detector, even with low SNR conditions and high under-sampling factors.