Subspace based estimation of parameters and linear space-time multiuser detection for WCDMA systems

The previously developed subspace techniques for estimation of parameters and recovery of WCDMA signals embedded in frequency selective channels have been probed as a very promising techniques with great theoretical and practical importance. Particularly, this paper extends a blind channel estimation algorithm based on intersection of subspaces. Such a scheme is used together with a JADE ESPRIT algorithm, which calculates jointly the relative delays and AOAs of multipath signals based on the shift invariance properties of WCDMA signals. Afterwards, a less complex linear space-time decorrelator with a sliding window of 3 bits over a frequency selective Rayleigh fading channel is studied. We investigate the BER performance of the system by Monte Carlo simulations