Dynamic sensor collaboration via sequential Monte Carlo

We consider the application of sequential Monte Carlo (SMC) methods for Bayesian inference to the problem of information-driven dynamic sensor collaboration in clutter environments for sensor networks. The dynamics of the system under consideration are described by nonlinear sensing models within randomly deployed sensor nodes. The exact solution to this problem is prohibitively complex due to the nonlinear nature of the system. The SMC methods are, therefore, employed to track the probabilistic dynamics of the system and to make the corresponding Bayesian estimates and predictions. To meet the specific requirements inherent in sensor network, such as low-power consumption and collaborative information processing, we propose a novel SMC solution that makes use of the auxiliary particle filter technique for data fusion at densely deployed sensor nodes, and the collapsed kernel representation of the a posteriori distribution for information exchange between sensor nodes. Furthermore, an efficient numerical method is proposed for approximating the entropy-based information utility in sensor selection. It is seen that under the SMC framework, the optimal sensor selection and collaboration can be implemented naturally, and significant improvement is achieved over existing methods in terms of localizing and tracking accuracies.