Signal Reconstruction Based on Distributed Compressive Sensing Least Angle Regression Algorithm in L-CR system

In Low Earth Orbit (LEO) micro satellite Cognitive Radios (L-CR) system, multiple LEO satellite nodes are provided with spectrum sensing functionality. They perform sensing, transmission and processing of information from ground gateway station via distributed network. Ground sink reconstructs the original signal from LEO satellite forwarding signals. Considering the interference from the authorized primary user (PU) to the secondary user (SU) satellites in LEO system, SU sensing signal contains PU interference and noise. Hence, the fundamental problem is the implementation of efficient algorithm to realize noisy signal reconstruction at ground sink in L-CR system. Signal reconstruction based on distributed compressive sensing (DCS) in L-CR system is studied. According to the characteristics of L-CR system, the reconstruction mean square error (MSE) and the complexity performance are investigated for different convex recovery schemes in low signal to noise ratio (SNR) region respectively. Namely, basis pursuit de-noising (BPDN), homotopy method and least angle regression (Lars) algorithm. It is indicated that, BPDN has the minimum reconstruct MSE with the highest complexity, whereas Lars algorithm makes a tradeoff between reconstruction MSE and complexity. On the basis of that, signal reconstruction scheme based on DCS-Lars is proposed. Simulation results show that, the proposed DCS-Lars scheme can effectively recover sensing signals in low SNR region with better spectrum detection performance, and reconstruction complexity can be reduced simultaneously.