Blind identification and equalization of CDMA time-varying dispersive channels in multipath-Doppler domain

We have developed a discrete-time canonical model of CDMA time-varying (TV) channels based on canonical samples of the channel's spreading function in the multipath-Doppler domain. With this model, we propose a new two-stage blind identification and equalization algorithm based on the second-order statistics (SOS) of the observations. The ideal single-input multi-output (SIMO) model is equivalent to an additive Gaussian noisy (AGN) SIMO model so as to cancel the error propagation effects of this two-stage method. Compared with the subspace method, this new algorithm needs much less computation of the input correlation of the equivalent multi-input single-output (MISO) model. Simulations demonstrate the effectiveness of the algorithm.