Multichannel group sparsity methods for compressive channel estimation in doubly selective multicarrier MIMO systems (extended version)

摘要

We consider channel estimation within pulse-shaping multicarriermultiple-input multiple-output (MIMO) systems transmitting over doublyselective MIMO channels. This setup includes MIMO orthogonal frequency-divisionmultiplexing (MIMO-OFDM) systems as a special case. We show that the componentchannels tend to exhibit an approximate joint group sparsity structure in thedelay-Doppler domain. We then develop a compressive channel estimator thatexploits this structure for improved performance. The proposed channelestimator uses the methodology of multichannel group sparse compressed sensing,which combines the methodologies of group sparse compressed sensing andmultichannel compressed sensing. We derive an upper bound on the channelestimation error and analyze the estimator's computational complexity. Theperformance of the estimator is further improved by introducing a basisexpansion yielding enhanced joint group sparsity, along with a basisoptimization algorithm that is able to utilize prior statistical information ifavailable. Simulations using a geometry-based channel simulator demonstrate theperformance gains due to leveraging the joint group sparsity and optimizing thebasis.