Abstract—In this paper, an improved multi-channel blind de-convolution algorithm for code division multiple access signals is proposed. Through minimizing the average squared cross-output-channel-correlation, the proposed scheme achieves a better source separation performance for the linear convolved mixture system. Unlike most of the existing linear convolved mixture separation algorithms in the literature, the improved algorithm can obtain the original users’ signal even in low signal-to-noise ratio in a non-cooperation scenario. Simulation results show that the proposed optimal scheme not only achieves superior performance on the bit error rate to those of the existing ones, but also provides a convergent solution under different channel conditions.

Index Terms—CDMA signal, non-cooperation, blind de-convolution, gradient algorithms

Cite: Hao Cheng, Na Yu, and Jun Liu, "Improved Multi-Channel Blind De-Convolution Algorithm for Linear Convolved Mixture DS/CDMA Signals," Journal of Communications, vol. 9, no. 10, pp. 785-791, 2014. Doi: 10.12720/jcm.9.10.785-791