Subspace-Based Blind Adaptive Multiuser Detection Using Kalman Filter

A new blind adaptive multiuser detection scheme based on a hybrid of Kalman filter and subspace estimation is proposed. It is shown that the detector can be expressed as an anchored signal in the signal subspace and the coefficients can be estimated by the Kalman filter using only the signature waveform and the timing of the desired user. The resulting subspace-based algorithm brings the benefit of lower computational complexity than the full-rank approach, and the theoretical analysis indicates that the proposed algorithm is also superior in convergence performance. The adaptive implementation in dynamic environment such as variable number of users is obtained by seamlessly integrating a subspace tracking algorithm. The new proposed subspace-based method is effective in AWGN channels as well as in slowly time-varying Rayleigh fading channels. Moreover, the proposed detector is much more robust against the signalling waveform mismatch and inaccurate knowledge of the amplitude of the desired signal than the full-rank one, as demonstrated by computer simulations.