Independent component analysis by general nonlinear Hebbian-like learning rules
Signal Processing - Special issue on neural networks
Blind multiuser detection: a subspace approach
IEEE Transactions on Information Theory
| Hi-index | 0.00 |
In this paper, a two phase algorithm is proposed for both blind synchronization and data sequence estimation of all users without any prior knowledge about spreading sequences in asynchronous unequal power multi-user direct sequence spread spectrum (DS-SS) systems. In the first phase, for blind synchronization, an eigenvalue variation (EV) based method is proposed, which uses all estimated eigenvalues related to signal, which are discriminated from noise eigenvalues by a threshold. In this paper, is shown EV to be a powerful tool for blind synchronization in eavesdropping scenarios in which unequal power signals are received from users. In the second phase, for blind data sequence estimation of all users, a variable step-size independent component analysis (ICA) algorithm based on negentropy maximization of active users is proposed using subspace as a preprocessing step. There is no need to know any spreading sequences for data estimation of users. Computer simulations confirm much better performance by the proposed algorithm at the cost of some more complexity compared with that of using only a pure subspace algorithm. Moreover, we compare the performance of the proposed blind synchronization with that of a successive blind synchronization, and we show that the proposed method is much faster.