Event-triggered consensus adaptive filters for target localization

Distributed filters show strong robustness by using certain resources of communications and calculations to collaboratively estimate or track an unknown dynamic process of interest over a sensor network. In this paper, an event-triggered mechanism (ETM) is introduced for least mean square (LMS)-based consensus adaptive filters to deal with applications with communication resource constraints. An upper bound of the estimation errors for the proposed event-triggered consensus adaptive filters is established under a cooperative information condition without independent or stationary signal assumptions. To verify the effectiveness and resource saving properties of the proposed event-triggered consensus LMS-based filters, numerical simulations for target localization using bearing-only measurements of multiple unmanned aerial vehicles are provided. It is proved that the ETM provides settable thresholds to artificially adjust the proportions between the estimation accuracy and the resource consumption. Finally, experimental results are given to further show the performance and applicability of the proposed algorithm in practical engineering.