MAP Receiver with Spatial Filters for Suppressing Cochannel Interference in MIMO-OFDM Mobile Communications.

Fan Lisheng,Kazuhiko Fukawa,Hiroshi Suzuki

DOI: https://doi.org/10.1587/transcom.e92.b.1841

2009-01-01

IEICE Transactions on Communications

Abstract:This paper proposes joint maximum a posteriori (MAP) detection and spatial filtering for MIMO-OFDM mobile communications, which can maintain excellent receiver performance even over interference-limited channels. The proposed joint processor consists of a log likelihood generator and a MAP equalizer. The log likelihood generator suppresses cochannel interference by spatially filtering received signals and provides branch metrics of transmitted signal candidates. Using the branch metrics, the MAP equalizer generates log likelihood ratios of coded bits and performs channel decoding based on the MAP criterion. In the first step, the log likelihood generator performs spatio-temporal filtering (STF) of the received signals prior to the fast Fourier transform (FFT) and is referred to as preFFT-type STF. For further improvement, in iterative steps, the generator performs spatial filtering (SF) of the received signals posterior to FFT and is referred to as postFFT-type SF. Estimation of both tap coefficients of the spatial filters and channel impulse responses employs the recursive least squares (RLS) with smoothing. The reason for making a switch from preFFT-type STF into postFFT-type SF is that preFFT-type STF outperforms postFFT-type SF with a limited number of preamble symbols while postFFT-type SF outperforms preFFT- type STF when data symbols can be reliably detected and used for the parameter estimation. Computer simulations demonstrate that the proposed joint processing can achieve excellent BER performance under interference-limited channel conditions and that it can outperform the conventional joint processing of preFFT-type STF and the MAP equalizer.