A Distributed Algorithm for Bootstrapping a Wireless Visual Sensor Network

Transmission of visual data can be quite demandingin terms of energy and bandwidth. Therefore, it is important thatall the sensors in a Wireless Visual Sensor Network get a goodsignal quality for their transmissions when they come online forthe first time. The purpose of this report is to design, simulate andevaluate an algorithm that the sensors can use to perform thisinitialization, i.e. bootstrapping. First the bootstrapping processis modeled as an NP-hard optimization problem. Then, by takinginto account what information is directly available to a sensor,the algorithm is designed to be distributed in order to saveenergy and achieve scalability of the network. To avoid excessivecomputing times due to the NP-hardness of the problem, thealgorithm is designed to approximate the theoretical solution.The algorithm is implemented and tested in a software simulationenvironment that was built as part of the project. The tests showthat the algorithm performs close to optimal for small networksand retains a good approximation ratio for medium to largenetworks