提出了一种新的基于线性相关平行剖面模型(parallel profiles with linear dependencies,PARALIND)的MIMO-CDMA盲多用户检测算法。PARALIND模型的分解唯一性并结合通信信号的结构特征,算法在信道衰落和用户扩频码未知的情况下实现MIMO-CDMA信号的盲多用户检测。对恒模约束PARALIND模型的分解唯一性进行了讨论,并给出了唯一性定理。仿真结果表明,基于PARALIND模型的多用户检测算法具有良好的误符号率性能。算法在处理小数据块时同样有效。
A new parallel profiles with linear dependencies(PARALIND)-based blind multiuser detection algorithm in MIMO-CDMA system is proposed.Based on the uniqueness of PARALIND decomposition and the structure property of communication signals,this algorithm can detect MIMO-CDMA signals without any knowledge of the channel coefficients and users' spreading codes.The uniqueness property of the PARALIND model with constant modulus(CM) structure is discussed,and the related uniqueness theorem is presented.Simulation results show that the algorithm has good symbol error rate performance and works well in small data block condition.