Performance analysis of signature waveform estimation in synchronous CDMA systems

We present an asymptotic performance analysis of a subspace-based method, proposed by Liu and Xu (see Twenty-Ninth Asilomar Conference on Signals, Systems, and Computers, 1995), for blind signature waveform estimation in synchronous code-division multiple-access systems. We summarize the algorithm, present a first order perturbation analysis of the noise eigenvectors, and derive expressions for the covariance of the estimated parameters. Using some examples of interest, we give Monte-Carlo simulations to validate the theoretical analysis and examine how the performance changes with varying parameters.