Wide Band Spectrum Sensing in Cognitive Radios using Compressed Sensing based on Improved Matching Pursuit Algorithms

Cognitive Radio (CR), which is proposed to be a candidate technology for 5G wireless communications, has to continuously perform the task of monitoring spectrum holes over a wide range of licensed frequencies. Scanning through a wide band spectrum, particularly in the GHz range, poses challenges, especially for the Analog-To-Digital Converters that belongs to the front end of the CR receivers. Compressive Sensing is a method which addresses this problem, by collecting fewer samples compared to the ideal Nyquist rate. This paper proposes to use modified reconstruction algorithms based on orthogonal Matching Pursuit (OMP) for the recovery of the wide band Primary User (PU) signals from a very low number of collected samples. Here three algorithms are discussed viz. Stage wise Orthogonal Matching Pursuit (StOMP), Regularized Orthogonal Matching Pursuit (ROMP) and CoSaMP Compressive Sampling Matching Pursuit (CoSaMP). Simulation studies reveal that these three algorithms are better compared to the original OMP algorithm. Also, a study on the mean square performance is also done which reveals that CoSaMP outperforms other peer level algorithms.