基于U-D分解滤波的盲空时多用户检测算法

Blind space-time multiuser detector based on U-D factor filter

多用户检测作为CDMA系统的抗多址干扰环节,在16bit或32bit专用微处理器或DSP上实现,采用卡尔曼滤波算法的最佳多用户检测算法必然面临数值计算稳定性的问题.文中提出利用U D分解渐消记忆滤波算法估计CDMA系统多用户接收器的最优判决向量,构造出一种收敛速度快、跟踪性能好、数值稳定性好的高性能盲空时多用户检测算法.仿真实验表明,该方法具有很强的抗多址干扰能力和较高的数值鲁棒性.

As a method for eliminating multiple access interference (MAI) in CDMA, blind space-time multiuser detection based on the Kalman filter will face the problem of numeric stability when used in 16 bits or 32 bits microprocessor or DSP. So a high performance blind space-time multiuser detection algorithm is proposed by introducing the U-D factor Kalman filter to estimate the optimal decision vector of the CDMA multiple-users receiver. This algorithm can converge rapidly, track in high accuracy and has satisfied numeric stability. Simulation results show that this algorithm can eliminate multiple access interference (MAI) efficiently and is robust to numerical calculation.