基于张量分解的DS-CDMA盲数据检测研究

Research on Blind Data Detection for DS-CDMA Based on Tensor Decomposition

为了节约建造天线的成本,本文借助张量分解技术,在DS-CDMA系统中建立PARAFAC模型,固定用户数量(6)、扩频码数(4)和符号快照的块长度数(50),改变信噪比和天线的数量,通过蒙特卡洛方法模拟生成多维阵列信号,执行COMFAC算法进行盲数据检测来求出误码率,进而在相同的信噪比条件下将误码率与天线数量进行数据拟合,最后结合拟合函数曲线和拟合方程得出最优天线数。研究结果表明拟合优度都达到90%以上,高信噪比的最优天线数少,稳定为5根;低信噪比的最优天线数多,并且不稳定。

In order to save the cost of antenna construction, the PARAFAC model is established in DS-CDMA system by means of tensor decomposition technology. Then the number of users (6), spread spec-trum code (4) and block length of symbolic snapshot (50) are fixed. Also the signal-to-noise ratio and the number of antennas are changed. Multi-dimensional array signal is simulated by Monte Carlo method. COMFAC algorithm is used to detect blind data to calculate bit error rate. Then, under the same SNR condition, the bit error rate and the number of antennas were fitted. Finally, the op-timal number of antennas was obtained by combining the fitting function curve and fitting equa-tion. The results show that the goodness of fit is more than 90%, and the number of optimal anten-nas with high signal-to-noise ratio is less, and the stable number is 5. The optimal antennas with low signal-to-noise ratio are numerous and unstable.