Mm-wave MIMO channel modeling and user localization using sparse beamspace signatures

Millimeter-wave (mm-wave) communication systems operating between 30GHz and 300GHz are emerging as a promising technology for meeting the exploding bandwidth requirements of future wireless systems. In addition to large bandwidths, mm-wave systems afford high-dimensional multiple input multiple output (MIMO) operation with relatively compact arrays, and the corresponding narrow spatial beams make beamspace MIMO communication particular attractive. An important implication is that while the ambient spatial dimension is high, mm-wave MIMO channels exhibit a low-rank structure that is manifested in the sparsity of the beamspace MIMO channel matrix. In this paper, we develop a model for sparse mm-wave MIMO channels and propose an approach to mobile station (MS) localization that exploits changes in statistics of the sparse beamspace channel matrix as a function of the MS position. Unlike most existing methods, line-of-sight (LoS) propagation is not mandatory and the proposed approach benefits from the information provided by non-line-of-sight (NLoS) paths. Beamspace sparsity is exploited for developing a low-dimensional maximum-likelihood (ML) classifier that delivers near-optimal performance with dramatically reduced complexity compared to conventional designs. Numerical results illustrate the impact of the physical environment, grid-resolution, and MIMO dimensions on localization performance.