Study of Techniques For Reliable Data Transmission In Wireless Sensor Networks

This thesis addresses the problem of traffic transfer in wireless sensor networks (WSN). In such networks, the foremost challenge in the design of data communication techniques is that the sensor's transceiver circuitry consumes the major portion of the available power. Thus, due to stringent limitations on the nodes' hardware and power resources in WSN, data transmission must be power-efficient in order to reduce the nodes' power consumption, and hence to maximize the network lifetime while satisfying the required data rate. The transmit power is itself under the influence of data rate and source-destination distance. Thanks to the dense deployment of nodes in WSN, multi-hop communication can be applied to mitigate the transmit power for sending bits of information, i.e., gathered data by the sensor nodes to the destination node (gateway) compared to single-hop scenarios. In our approach, we achieve a reasonable trade-off between power-efficiency and transmission data rate by devising cooperative communication strategies through which the network traffic (i.e. nodes' gathered information) is relayed hop-by-hop to the gateway. In such strategies, the sensor nodes serve as data originator as well as data router, and assist the data transfer from the sensors to the gateway. We develop several data transmission schemes, and we prove their capability in transmitting the data from the sensor nodes at the highest possible rates allowed by the network limitations. In particular, we consider that (i) network has linear or quasi-linear topology, (ii) nodes are equipped with half-duplex radios, implying that they cannot transmit and receive simultaneously, (iii) nodes transmit their traffic at the same average rate. We compute the average data rate corresponding to each proposed strategy. Next, we take an information-theoretic approach and derive an upper bound to the achievable rate of traffic transfer in the networks under consideration, and analyze its tightness. We show that our proposed strategies outperform the conventional multi-hop scheme, and their average achievable rate approaches the upper bound at low levels of signal to noise ratios.