Localization of diffusive sources using distributed sequential Bayesian methods in wireless sensor networks

We develop an efficient distributed sequential Bayesian estimation method to localize a diffusive source in wireless sensor networks. Potential applications include security, environmental monitoring, pollution control, and explosives detection. We first derive the physical model of the substance dispersion by solving the diffusion equations under different environmental scenarios. We then integrate the derived dispersion models into the distributed processing technologies, and propose a distributed sequential Bayesian localization technique, in which the state belief is transmitted in the wireless sensor networks and updated using the measurements from the new sensor node. In order to decrease the required communication burden we propose two parameterizable belief approximations: a Gaussian approximation and a new linear combination of polynomial Gaussian approximation. We also apply the idea of information-driven sensor scheduling and select the next sensor node according to certain criterions to reduce the response time and save energy consumption of the sensor network.