Blind channel identification of dispersive CDMA channels

A new blind identification method based on signal subspace estimation is proposed. With the dispersive CDMA channel modeled as a multiple-input multiple-output signal model, the impulse response of the multiuser's channels can be identified without training sequences. The method characterizes parallel processing and lower algorithmic complexity, hence it is suitable for high-rate CDMA communication system. The computational requirement of the algorithm is o(N/sup 2/), N being the size of the covariance matrix. The algorithm is locally convergent and the undesired stationary points are unstable. Monte-Carlo simulations demonstrate the performance of the proposed methods in a channel identification context.