Blind channel identification for OFDM systems with receive antenna diversity

We address the problem of channel estimation in orthogonal frequency division multiplexing (OFDM) systems. Exploiting receive antenna diversity, a second-order statistics-based (SOS) blind technique is proposed. Our method differs from the existing SOS-based techniques in that channel estimation is carried out using the frequency domain (i.e., post-FFT) signals whereas existing methods use the time domain signals (i.e., pre-FFT). In the proposed method, channel identifiability is guaranteed regardless of the channel zero locations, so long as any roots common to all the diversity channels are either at the origin or on the unit circle. Only short data records are required to achieve good performance. For PSK transmission, this method enables channel estimation even from a single OFDM symbol at high SNR, which is impossible with the existing SOS-based methods. Further when only a small number of subcarriers can be used for channel estimation because of computational complexity, we determine the optimal set of subcarriers in terms of estimation accuracy.