Blind equalization in asynchronous DS-CDMA systems based on linear prediction approach

In this paper, direct equalization based on linear prediction is presented for asynchronous DS-CDMA systems operating in a multi-path environment. The classical subspace-based methods are computationally costly and do not tolerate mismatched channel lengths, which are the major limitations to their practical applicability. The important property of computational simplicity and performance robustness makes the Linear Prediction Algorithm (LPA) one of the potentially attractive solutions to the blind channel estimation/equalization problem. One possible drawback of this approach is that the channel estimation may suffer from system noise and computation errors. Therefore, we first exploit the recently proposed linear prediction approach to estimate the column vector subspace of the channel. Then we design the direct blind zero-forcing and MMSE equalizers. This avoids the channel estimation error and the resulting methods are therefore more accurate, more robust and are also near-far resistant. Simulations show that our direct equalization methods are effective.