Staggered scheduling of estimation and fusion in long-haul sensor networks

In long-haul sensor networks, sensors are remotely deployed over a large geographical area to perform certain tasks. We study a class of such networks where sensors take measurements of one or more dynamic targets and send state estimates of the target(s) to a fusion center via long-haul satellite links. The severe loss and delay over the satellite channels can easily reduce the chance that an estimate is successfully received by the fusion center, thereby limiting the potential information fusion gain and resulting in suboptimal accuracy performance of the fused estimates. In this work, starting with the temporal-domain staggered estimation for an individual sensor, we explore the impact of the so-called intrastate prediction and retrodiction on estimation errors. We also investigate the effect of such estimation scheduling across different sensors on the spatial-domain fusion performance, where the sensors retain the same estimation frequency, but with possibly asynchronous estimation instants staggered over time. In particular, the impact of communication delay and loss on such scheduling is explored by means of numerical and simulation studies that demonstrate the validity of our analysis.