A blind symbol recovery for dual antenna OFDM systems

Blind symbol recovery is attractive in OFDM systems because it saves bandwidth that would otherwise be lost to training sequences or pilot tones. Here, a space-frequency implementation of iterative least square constant modulus algorithm (ILS-CMA) on dual antenna OFDM systems is explored and shown to have modest computational complexity compared to the conventional OFDM symbol recovery scheme using channel estimation and frequency domain equalization. It is shown through computer simulations that the symbol error rate (SER) performance of ILS-CMA is close to or even better than the perfect dual antenna combiner for the cases both with or without a cyclic prefix (CP). Also, a fast convergence behavior (less than 10 iterations) is achieved at 20 dB SNR or above with some modest convergence error value.