Estimation du canal pour les réseaux de capteurs sans-fil utilisant les techniques de diversité coopérative
Yi Zhang · HAL (Le Centre pour la Communication Scientifique Directe) · 2012
This thesis considers channel estimation in amplify-and-forward wireless relay networks when coherent distributed space-time block coding (DSTBC) is performed at the relays in the aim of benefiting from distributed diversity. The studied network consists in one source and one destination node, and several relay nodes distributed between them. Given the practical limitations to acquire full channel state information (CSI), we consider the case where only partial CSI is available at the destination node. Here, for the case of pilot-only-based channel estimation, we explain the limited receiver performance when pilot overheadshould be kept to a minimum for the reasons of network energy efficiency. To overcome this problem, we firstly propose to use semi-blind (SB) channel estimation based on the expectation maximization (EM) algorithm. At a first step, we consider orthogonal DSTBC at the relays and minimum mean-square-error signal detection at the destination. In this case, starting by the classical formulation of EM for our system (that we call CB-EM for Classical Biased EM), we further propose a modified implementation of EM in view of obtaining an unbiased channel estimate at the receiver. We show the interest of the proposed scheme, called UL-EM (standing for Unbiased Linearly-combined EM), especially for relatively large signal constellations. At a second step, we consider non-orthogonal DSTBC at the relays, and propose to use an iterative CB-EM-based channel estimation and parallel interference cancelation for signal detection. This scheme allows an important performance improvement, as compared to the classical scheme. Finally, for the case of relatively fast fading conditions, instead of using SB estimation, we propose to use an improved signal detection scheme that takes the channel estimation errors into account and reduces the impact of channeluncertainty on the receiver performance. We illustrate the interest of this scheme for relatively large signal constellations.