Maximum-likelihood delay estimation for DS-CDMA communication systems

The delay estimation algorithms of a synchronous DS-CDMA system are considered under multipath fading channels. A novel maximum-likelihood delay estimation algorithm is addressed for resolving closely spaced delays in indoor wireless communications. The observation space can be partitioned into a signal and a noise subspace by eigenvector decomposition of the time-averaged correlation matrix generated from the output of a sliding correlator without prior knowledge of the signal vectors. Exploiting known statistical properties of the estimated subspaces to maximize the likelihood of the projections forms the maximum-likelihood delay estimation. The simulation results show that the proposed delay estimation algorithm is superior to the conventional sliding correlator and MUSIC-type algorithm.