Nonuniformity in wireless sensor networks

Abstract

Deemed as suitable for various challenging sensing tasks, wireless sensor networks evolve as candidates for deployment to different environments to perform the given task for as long as possible. Those networks operate redundantly in a distributed manner, requiring little intervention. However, various factors such as the sink location, node deployment characteristics or external impacts such as intentional destructions can severely limit the lifetime of the overall network. The nonuniformity of the networks caused by such factors may result in underutilized network deployments where some parts of the network is still alive yet unable to reach the sink due to disconnections in the intermediate sections. In this thesis, the effects of deployment nonuniformity is analyzed and methods for mitigating the problems are presented. A sink placement algorithm is presented to find a safe and optimal location for the sink node. Methods to find the bottleneck nodes and sensing holes inside the network are provided. Redeployment based hole mitigation techniques are proposed to prolong the network lifetime under strict quality requiring situations such as border surveillance. An analytical quality metric is also presented to understand the sensing quality under lossy assumptions. The presented methods and models are tested using simulations and the results for different parameter sets are given to see the suitability of each.