基于ICA和梅西算法的LSC-DS-CDMA信号伪码估计

Pseudo code estimation of long and short code DS-CDMA signals

长短码直扩码分多址(Long and Short Code Direct Sequence Code Division Multiple Access,LSC-DS-CDMA)信号包含多个用户的长码和短码,低信噪比时,各用户之间干扰大,基于三阶相关函数的长短码估计方法性能不佳。为此,提出一种基于FastICA和梅西算法的伪码估计算法。首先,根据长码和短码周期对信号进行2次分段,将单通道LSC-DS-CDMA信号构造为多通道信号模型,采用FastICA算法分离每个用户长短码组成的复合码片段并进行拼接;其次,将各用户复合码延迟相乘2次,依次消除幅度模糊和短码干扰;然后,采用滑动搜索和梅西算法,估计得到长伪码生成多项式,恢复出长码,消除拼接误差;最后,在复合码和长码中做相关运算估计得到短码。仿真结果表明,信噪比大于-7 dB时,算法能准确估计3个用户的长短码。

Since the long and short codes direct sequence spread spectrum code division multiple access(LSC-DS-CDMA)signal contains the long and short codes of multiple users,the interference between users is large under the background of low signal-to-noise ratio,and the performance of long and short codes estimation method based on third-order correlation function is poor.Therefore,a PN codes estimation method based on FastICA and Massey algorithm is proposed.Firstly,a multi-channel signal model is constructed from the single channel LSC-DS-CDMA signal by twice segmenting according to periods of the long codes and short codes,and then the composite codes fragments composed of long codes and short codes are separated by FastICA algorithm and spliced.Secondly,the amplitude ambiguity and short codes interference are eliminated by multiplying the delay of each user s composite codes twice.Then,the polynomials of long codes generation are estimated and the long codes are recovered to eliminate the splicing error by using sliding search and Massey algorithm.Finally,the short codes are estimated by correlation operation from the estimated composite codes and long codes.The simulation results show that the algorithm can accurately estimate the length code of three users when the SNR is above-7 dB.