Direct Target Tracking by Distributed Gaussian Particle Filtering Based on Delay and Doppler

For large-scale, geographically-distributed sensor networks, we develop a fully distributed adaptive direct target tracking by Gaussian particle filtering (D-GPF) using time delay and Doppler. Without a dedicated fusion center in this distributed framework, each sensor in the network adaptively tracks the target in a fully distributed manner without estimating the time delay and Doppler of the target. Based on the Adapt-Then-Combine (ATC) diffusion strategy, each sensor updates the target state estimates by fusing its information with diffused localized intermediate state estimates from its direct neighbors with single-hop transmission only. Simulation results validate that the tracking performance of the proposed D-GPF is quite congruent with that of the centralized particle filter (CPF).