A multiuser detection method based on support vector machine

Abstract

In this paper, a multi-user detector based on a Support Vector Machine (SVM) is proposed, which divides the receiving vector into two classes, +1 and -1, to attain detection. Differing from the MMSE detector, the SVM method can find an optimal hyperplane to separate the +1 and -1 from the training data. Simulation results show that under the Rayleigh channel, this detector can achieve a relatively low BER in comparison with the minimum mean square error (MMSE) detector.