Bilinear Neural Network Tracking Subspace for Blind Multiuser Detection

In order to lower the level of multi-access interference (MAI) and improve the reception performance in CDMA communication systems, a new blind multiuser detection algorithm is proposed. Different from some other blind detection approaches, the proposed algorithm tracks the eigenvalues and eigenvectors of signal subspace by bilinear neural network based on a new principal component analysis (PCA) method. Simulation results show that the proposed algorithm improves the tracking capability and bit error rate (BER) performance of blind detector compared with the MMSE and PASTd algorithm. In consideration of its comparatively low computational complexity, the proposed detection has some value in practice.