A new adaptive ordering method for successive decision feedback multiuser detection

The paper analyzes the importance of adaptive user detection ordering in an adaptive successive multiuser decision feedback detection scheme with imperfect feedback. The analysis shows that a better performance in terms of bit error rate (BER) is obtained if the less erroneous users are detected first. We propose a new user detection ordering method based on mean squared error (MSE) measurements in the training period. A significant BER gain is obtained with respect to the case when the users are detected based on the decrease of the received powers or in an arbitrary order. An adaptive least mean square (LMS) algorithm is used to determine the users detection order as well as to estimate adaptive filter coefficients.