Performance of a low-complexity downlink channel estimator for long-code CDMA systems

Recently, a blind downlink channel estimator was proposed for long-code CDMA systems based on a covariance matching idea. By modeling long spreading codes as random codes and pre-computing some code-dependent quantities, the algorithm shows extremely low complexity. Due to inaccuracy of estimated data covariance from finite number of received data samples, its asymptotic performance depends on the sample size. In this paper, it is further analyzed in terms of the covariance and mean-square-error of the estimated channel vector.