Topology Control in an Indoor Wireless Sensor Network using Realistic Propagation Models by Gagan Goel

Topology Control in an Indoor Wireless Sensor Network using Realistic Propagation Models Gagan Goel Master of Applied Science Graduate Department of Electrical and Computer Engineering University of Toronto 2014 This thesis considers the problem of designing an energy-efficient communication backbone for a wireless sensor network deployed in an indoor environment. We propose and implement a Topology Control scheme to determine optimal sensor node transmission power levels that maintain a desired level of network connectivity. Power optimized singly connected network is constructed using a Minimum Spanning Tree based formulation. To achieve fault-tolerance, we employ a Matroid Intersection based approach to construct a Minimum Cost Rooted k-Outconnected Spanning Subgraph, basis for the computation of a Minimum Power k-Connected Spanning Subgraph, k > 1. Realistic wireless channel assumptions are made to construct the optimized network topologies. Field measurements are performed in an indoor office setup to characterize the channel parameters and hence evaluate the accuracy of various path loss models suitable for such propagation environments. Network simulations validate the necessity of exercising Topology Control to achieve network longevity.