OFDM Channel Estimation using Total Variation Minimization in Compressed Sensing

Compressed sensing (CS) is a newly emerging technology, which states that a signal can be sampled at a rate much smaller than what is commonly prescribed by Shannon-Nyquist. CS combines sampling and compression in to a single non-adaptive linear measurement process. In this paper, we are investigating the application of compressed sensing to the problem of estimating a doubly selective channel in Orthogonal Frequency Division Multiplexing (OFDM) system. We presented a compressed sensing based pilot-aided channel estimation method, in which Total variation minimization by augmented Lagrangian and alternating direction algorithms (TVAL3) is used as a compressed sensing algorithm. Most of the existing pilot-aided channel estimators, e.g. Least square estimator, depends on the use of large number of pilots to increase the accuracy of estimation, which results in lack of spectral efficiency. Results of the TVAL3 method is compared with the existing LS method. Simulation results shows that TVAL3 based channel estimator is proves to be a very good alternative to LS channel estimator, yielding good estimation quality despite using only fewer number of pilots. The TVAL3 based compressive channel estimator establishes the fact that Total variation (TV) minimization has application in channel estimation.