Semiblind multichannel identification in asynchronous multiuser OFDM systems

This paper presents a novel subspace-based semiblind channel identification scheme for asynchronous multiuser orthogonal frequency division multiplexing (OFDM) systems with antenna diversity. The inherent property of asynchronous OFDM signals allows separation of the desired user´s channel from the interfering users´ channels via subspace decomposition. Moreover, the proposed semiblind channel estimator overcomes practical limitations of the conventional subspace method due to unknown channel lengths and noise-corrupted data by employing pilot symbols. Simulation results demonstrate that the semiblind channel estimator significantly outperforms a pilot-only estimator when most part of the noise subspace is exploited.