Cooperative Diversity for Energy Efficient Wireless Sensor Networks

Abstract

This thesis investigates the feasibility of deploying cooperative diversity as a practical energy saving technique for extending the operational lifetime of a wireless sensor network. Wireless sensor networks consist of many resource-constrained distributed nodes, powered by small batteries which are typically never replaced. In cooperative communication a partner node overhears and repeats a source node's message to the destination receiver, where the two independently faded signals are combined. The resulting diversity benefit of achieving reliable communication at a much lower transmission energy cost reduces the energy consumption of the sensor nodes. A novel low-complexity distributed cooperation strategy for energy-constrained wireless sensor networks is presented in this thesis. The proposed cooperation strategy is named ECO-OP in reference to the energy-conserving cooperation it facilitates among autonomous sensor nodes. ECO-OP is based on simple yet robust power allocation and partner choice heuristics which enable individual sensor nodes to independently make energy efficient cooperation decisions. The heuristics are computationally efficient and based solely on measurements of average channel path loss, making them suitable for practical implementation on resource-constrained sensor nodes. Importantly, ECO-OP thereby enables simple sensor nodes to cooperate autonomously to extend the lifetime of the wireless sensor network as a whole. To the best knowledge of the author, ECO-OP is the first distributed cooperation strategy for coordinating cooperative communication in an energy-constrained wireless network. The behaviour and performance of network-wide cooperation using ECO-OP is analysed in three representative wireless sensor network topologies, with respect to the resulting energy conservation and network lifetime extension. Importantly, it is demonstrated that altruistic cooperation among autonomous resource-constrained sensor nodes using ECO-OP significantly extends the lifetime of the wireless sensor network as a whole. This applies to both sparse and dense networks, with both random and directed data flow. Moreover, it is shown that the low-complexity distributed ECO-OP cooperation strategy proposed in this thesis overall achieves a comparable network lifetime improvement to a computationally intensive centralised cooperation algorithm from the literature.