A regularized least-squares approach to event-based distributed robust filtering over sensor networks

In this paper, a new distributed robust filtering problem is investigated for a class of discrete-time systems subject to parameter uncertainty over sensor networks under the event-based transmission mechanism, where the uncertainty in the target plant is norm-bounded. To save limited network resources, an event-based communication strategy is presented to arrange data scheduling, where the next triggering time of each node can be predicted, and thus continuous listening is avoided. By employing the robust regularized least -squares estimation approach, a local robust recursive algorithm is first developed by minimizing the given quadratic cost index. Then, a fully distributed information fusion scheme is implemented by exchanging information between adjacent nodes, where the local information spreads to the entire network after a finite-step iteration under the connected topology. Furthermore, by utilizing the matrix inequality technique and the mathematical induction approach, some sufficient conditions are derived to ensure the uniform boundedness of the weighting matrices. Finally, an illustrative example is given to validate the developed distributed robust filtering algorithm. (c) 2024 Elsevier Ltd. All rights reserved.