A Scalable BP Method for Joint Localization and Synchronization in Dense Wireless Sensor Networks

In this paper, we develop a joint cooperative localization and synchronization scheme for dense mobile wireless sensor networks (WSNs) using a scalable belief propagation (BP) based method. We consider a distributed time-varying WSN with mobile devices, where message packets are propagated among all devices starting from temporal and spatial anchors. To account for the nonlinear system models and to compute the belief at each device while maintaining low communication and computation complexity, we propose an efficient scalable BP scheme, where a temporary posterior belief is calculated and updated sequentially so that the dimension of measurement covariance matrices is fixed instead of the unlimited dimension augmentation and batch computation in sigma point belief propagation (SPBP). Simulation results demonstrate a significant enhancement on the robustness of the algorithm and reduction of the computational complexity compared to the baseline scheme.