On the Cramér-Rao Bound of Time-Varying Narrowband Leaked MIMO OFDM Channels

Abstract

In this paper, the problem of time-varying narrowband leaked Multiple-Input Multiple-Output Orthogonal Frequency Division Multiplexing (MIMO OFDM) channel estimation is considered. Practical constraints on communication systems, e.g. finite bandwidth and block length cause leakage rendering a sparse channel into a non-sparse one. The current work extends prior work on single carrier and OFDM; herein, it is shown that leaked MIMO OFDM channel is also separable in the delay and Doppler domains. A convex optimization approach, based on the atomic norm heuristic and bilinear representation of the channel, is adapted for channel estimation. The Cramer-Rae Bound (CRB) on the leaked MIMO OFDM channel estimation is derived. Simulation results show superiority of the atomic norm-based bilinear channel estimator over the classical l 1 element-wise sparse method and an adaptive MIMO OFDM channel estimator; with an average 6.3 dB and 2 dB improvement, respectively. Comparison of the CRB with simulation, shows that the proposed algorithm is essentially optimal.