Abstract
Embedded signaling is an efficient method to blindly estimate different parameters in wireless communication systems. In this paper, we propose a new nodes identification method based on embedded signaling for wireless relay systems. Data packets, modulated with high order modulations, are transmitted by a source node and forwarded by a relay node to a destination node. In the proposed identification scheme, source and relay nodes incorporate precoding sequences, based on codewords generated from block codes, into the data sequence. A novel soft treatment based on the computation of log-likelihood ratios is performed at the destination node to identify source and relay nodes involved in packets transmissions. This soft treatment avoids noise amplification encountered in previous works. Two types of identification schemes are proposed: the hierarchical scheme and the non-hierarchical scheme. The hierarchical scheme allows the destination to recover hierarchically the identities of transmitting source and relay nodes of received packets in two shots. The non-hierarchical scheme allows the destination to efficiently recover in one shot the identities of transmitting nodes. We study the complexities of both identification schemes and compare their performance in terms of identification error rate. Since the knowledge of the channel is necessary in the identification process, we propose a blind estimation technique allowing the destination to estimate the channel. Simulations results show that our proposed methods significantly outperform previous identification methods based on embedded signaling and hard decoding. Indeed, gains of approximately 1.5 dB and 2 dB are obtained at high signal to noise ratio when the non-hierarchical scheme is used respectively for 8 Phase Shift Keying and 16PSK modulations. Simulations show also that the non-hierarchical scheme always outperforms the hierarchical scheme.