Event-triggered fusion estimation under bounded noises

This paper is devoted to solving the event-triggered multi-sensor fusion estimation problem with bounded noises. Under the centralized fusion framework, a compensation strategy is developed to recover the non-triggered data by analyzing the intrinsic properties of Kalman filter. Then, an optimal centralized fusion estimator is derived by minimizing an upper bound of estimation error at each time. Under the distributed fusion framework, the optimal local estimator on each observable subspace is obtained by the ideas similar to those in the centralized fusion framework. Then, two different distributed fusion estimators are respectively designed by means of a spectral radius inequality and a trace inequality, where the former achieves better estimation performance and the latter is more computationally efficient. Particularly, all the proposed centralized and distributed fusion estimators are shown to be stable under moderate conditions. Finally, a classical target tracking system is utilized to verify the effectiveness and advantages of the proposed methods.(c) 2023 Published by Elsevier Ltd.

Automated summary

This paper focuses on solving the event-triggered multi-sensor fusion estimation problem with bounded noises. In the centralized fusion framework, a compensation strategy is developed to recover the non-triggered data, and an optimal estimator is derived by minimizing the estimation error. In the distributed fusion framework, optimal local estimators are obtained on each observable subspace, and two different distributed fusion estimators are designed using spectral radius inequality and trace inequality, respectively.