Low-complexity energy optimization of wireless sensor networks

Wireless sensor networks (WSNs), which consist of numerous devices that take measurements of a physical phenomenon, are becoming a popular area of research. Since the sensor nodes are typically battery powered, energy optimization and efficiency is extremely important in WSNs. However, optimizing power proves to be a non-trivial task since decreasing the transmission power will result in degraded signal and unsuccessful transmission. In this thesis, we look at the physical layer and propose two schemes based on channel codes that could be employed for optimization of transmission power of WSNs. First we consider the fact that the phenomena being observed by the sensor nodes are commonly correlated in space. Therefore, we devise a low-complexity coding scheme for correlated sources based on Slepian-Wolf compression, and analyze its performance in terms of diversity order. The main idea of this scheme is to use the correlated measurements as a substitute for relay links. Although we show that the asymptotic diversity order is limited by the constant correlation factor, we give experimental results that show excellent performance over practical ranges of SNR. In the second part of the thesis, we consider the fact that there may be many potential relays within radio range of a source; similarly, there may be many potential sources seeking to use relays. Allocating these resources is a non-trivial optimization problem. We consider fractional cooperation, where each potential relay only allocates a fraction of its resources to relaying. It is shown that linear programming can be used to optimally allocate resources in multi-source, multi-relay networks, where the relays use the demodulate-and-forward (DemF) or the decode-and-forward (DF) strategy, and where the transmissions are protected by low-density parity-check (LDPC) codes. Compared with existing optimization schemes, this method is particularly suitable for very large networks with numerous sources and relays. Simulation results are presented to demonstrate the performance of this scheme.