Subspace algorithms for synchronization and tracking in CDMA

CDMA (code division multiple access) is a multiple access technique based on spread spectrum methods. To estimate the delay parameters appearing in the data model, several authors have proposed subspace-based solutions. Unfortunately, these methods, for example the eigenvector-based multiple signal classification (MUSIC) method are computationally quite expensive. In addition, when the number of users or propagation paths is unknown, it should be estimated accurately. To this end, one can use relatively the simple minimum description length (MDL) criterion, but it fails when the signal-to-noise ratio (SNR) is not sufficiently high. In this paper, we introduce subspace type delay estimation and tracking algorithms which are clearly more efficient than MUSIC with MDL. These algorithms use a new criterion for determining the model order. Simulations show that the new methods outperform standard MUSIC.