| In this thesis, we propose two approaches to using data inference techniques to conserve energy in wireless sensor networks. First, we propose a novel approach for efficiently sensing a remote field by trading off reduced energy usage for reduced accuracy of the data recorded. Our approach, the infer algorithm, puts nodes into sleep mode for a given period of time and uses Bayesian inference to infer the missing data from the nodes in sleep mode. Simulations show that on average our algorithm produces energy savings of 59% while producing results that are accurate to within 7.9%. Second, we solve the problem of inferring per node loss rates using passive end-to-end measurements. We formulate the problem as a Maximum-Likelihood Estimation (MLE) problem and show how it can be efficiently solved using the Expectation-Maximization (EM) algorithm. Finally, we validate our analysis through simulations. |
