Tree Based Data Gathering from Sensors:Topology Management Sustaining QoS.
No Thumbnail Available
maintain both the connectivity and the sensing coverage in the network. While the rst contribution focuses only on the connectivity aspect, the second contribution of the thesis considers both the connectivity and the coverage maintenance as the design objectives. Considering irregular terrain property and optimal positioning of the sink, the energy depletion rate gradually decreases from the sink towards the terrain periphery. So, both the connectivity and the sensing coverage are a ected as the nodes near the sink die out of energy sooner than the leaves of the tree rooted at the sink, thereby creating network holes. For an improved network lifetime, a gradient based node deployment strategy has been proposed that also satis es the initial connectivity and the coverage criteria. The density of the deployed nodes follows a gradient, which is estimated as the amount of energy dissipation at any intermediate node to that of the leaf nodes in its rooted subtree. The proposed theory has been justi ed through the worst case analysis of the sensor network calculus. As unbalanced data gathering tree escalates the problem of uneven energy depletion in the network, a load-balanced distributed BFS tree construction scheme has been proposed. Further, to handle arbitrary node failure a localized and coste ective tree maintenance scheme has been introduced. Finally, the characteristics and the design objectives of tree based data gathering with failure support have been explored considering two di erent applications, one for road tra c monitoring, and the other for critical infrastructure monitoring. The inherent challenges in distribution and management of sensor network along the road require an application-speci c protocol support for the network connectivity, sensing coverage, reliable data gathering and the network lifetime improvement. The next contribution of the thesis introduces the concept of k-strip length coverage along the road, that ensures a better sensing coverage for the detection of moving vehicles, compared to the conventional barrier coverage and full area coverage, in terms of the availability of su cient information for statistical processing as well as the number of sensors required to be active. To extend the network lifetime, every sensor follows a sleep-wakeup schedule maintaining the network connectivity and the k-strip length coverage. This scheduling problem is modeled as a graph optimization, the NP-hardness of which motivates to design a centralized heuristic, providing an approximate solution. The properties of the proposed centralized heuristic are then explored to design a per-node solution based on local information.
Supervisor: Sushanta Karmakar
COMPUTER SCIENCE AND ENGINEERING