An Analysis of the Performances of Compressive Sensing Algorithms OMP and KLT for Cognitive Radio
|Research Area:||Volume 5,Issue 5, Sept. 2016||Year:||2016|
|Type of Publication:||Article||Keywords:||CS, DWT, KLT, Measurement Matrix, Measurement Vector, OMP, Signal Detection, Signal Recovery, Sparsity, Sparsity Order|
This paper provides a comparison between two widely used algorithms in the field of Compressive Sensing (CS), namely Orthogonal Matching Pursuit (OMP), and, Karhunen-Loève Transforms (KLT). As CS is one of the most essential techniques used by a Cognitive Radio (CR) for efficient usage of spectrum, it is required to be optimally simple, and, still, fast in working. The complexity here refers to the No. of computations a CR is required to make while using such algorithms and, this also, will in turn affect the effective requirement of hardware and power consumption. In this work, by means of simulations, we have tried to get an insight of working both this algorithms, OMP and KLT; and carried out the comparison between the two regarding their performances for the same experimental setup. We have discussed and evaluated their performances in terms of time, exact reconstruction of signal, percentage of error, and, complexity in terms of big-O, and, the probability of missed detection and probability of false alarm. From the simulation results we find that the OMP is quite promising CS tool as compared with the KLT in all these different aspects. As the CS is applicable to wideband spectrum sensing for CR and for varying sparsity environments, we are making comparison between the two that how the performance varies with different values of sparsity in frequency domain. We will carry out our further work on the bases of this work for modifying the OMP for CS.
Full text: IJEIR_2116_FINAL.pdf