An independent component analysis (ICA) based CDMA receiver is proposed, which in addition to the independence of users also exploits the independence of pseudo-codes in user signal separation. The proposed scheme is termed as independent code division multiple access in contrast to orthogonal code division multiple access being used in conventional CDMA receivers. Independence of pseudo-codes has been invoked by introducing the so-called independent Gram-Schmidt process in a deflation based ICA algorithm. The independence via Gram-Schmidt process is claimed by taking the inner product of successive vectors in a nonlinearly transformed Hilbert space, incorporating higher order statistics, instead of that in the original Euclidean space. The proposed receiver has been shown to outperform the existing CDMA receivers, in terms of bit error rate, through simulations. Moreover, the introduction of nonlinear transformation is shown to have no adverse effect on the convergence of the original deflation based ICA algorithm in terms of number of iterations apart from the computational cost of nonlinear transformation at each step. Gold codes are used for the demonstration of the proposed scheme although more work is needed to be done in order to identify the code family more amenable to the independence property.

提出了一种基于独立分量分析（ICA）的CDMA接收器，除了用户的独立性外，还利用伪码在用户信号分离中的独立性。与在常规CDMA接收机中使用的正交码分多址相比，所提出的方案被称为独立码分多址。通过引入所谓的独立Gram-Schmidt来调用伪代码的独立性基于通货紧缩的ICA算法进行处理。通过在非线性变换的希尔伯特空间中采用连续向量的内积，并结合高阶统计量而不是原始欧几里得空间，来主张通过Gram-Schmidt过程的独立性。通过仿真，已证明拟议的接收机在误码率方面优于现有的CDMA接收机。而且，除了每一步的非线性变换的计算成本之外，在迭代次数方面，非线性变换的引入对原始的基于放气的ICA算法的收敛没有不利影响。尽管需要做更多的工作以识别更适合独立性的代码族，但仍需使用黄金代码来演示所提议的方案。