A new approach to adaptive blind MMSE multiuser detection

A new adaptive blind multiuser detector is proposed. The algorithm is founded on minimizing the mean squared error (MSE) but does not require training sequences, and converges to the minimum MSE (MMSE) detector. Like the blind minimum-output-energy (MOE) detector, the proposed detector requires only the signature waveform and timing of the user of interest. We show that the blind MMSE and MOE detectors have similar transient and steady-state performance, and the same level of robustness against imprecise estimation of the desired user's received signature waveform. However, the proposed adaptive detector has 2N+1 fewer multiplications and N fewer summations than the MOE detector, where N is the length of the tap-weight vector.