Abstract
We treat the problems of propagation delay and channel estimation as well as data detection of orthogonally modulated signals in an asynchronous DS-CDMA system over fading channels using the maximum likelihood (ML) approach. The overwhelming computational complexity of the ML algorithm makes it unfeasible for implementation. The emphasis of this paper is to reduce its complexity by some approximation methods. The derived approximative ML schemes are compared with conventional algorithms as well as some others, e.g., the parallel interference cancellation (PIC) for data detection and the subspace-based algorithm for acquisition