Subspace-based estimation of multipath channel parameters for CDMA communication systems

Considers the estimation of multipath channel parameters to exploit time diversity in asynchronous code division multiple access communication systems. The multiuser channel estimation problem is decomposed into a series of single user problems through a subspace-based approach using the sample correlation matrix of the received signal. The authors estimate the channel impulse response associated with a given user by choosing filter coefficients that minimize the projection of the filtered spreading waveform into the estimated noise subspace. They then perform a least-squares fit of a parametric channel model to this unconstrained estimate to extract the desired channel coefficients. This method requires no training sequence, is near-far resistant, and is robust to changing channel conditions