Novel adaptive semi-blind multiuser detection algorithm based on LMS

From the foundation of the investigation on the least mean square (LMS) blind multiuser detection, an adaptive semi-blind multiuser detection algorithm based on the subspace constrained LMS is proposed for multicarrier CDMA (MC-CDMA) uplink. A semi-blind multiuser detector based on the minimum output energy (MOE) principle is designed, and a subspace constrained LMS algorithm is proposed to obtain the MOE weight vector adaptively. To reduce the computational complexity, an improved projection approximation sub-space tracking with deflation (PASTd) algorithm is adopted for adaptive signal subspace estimation. This proposed algorithm suppresses the interferers by exploiting the spreading sequences of all known users and mitigates the noise by constraining the weight vector in the signal subspace. As a result, the performance of the output signal-to-interference-plus-noise-ratio (SINR) and the bit error rate (BER) is improved. Simulation results demonstrate the effectiveness of the proposed algorithm.