Dynamic event-triggered cooperative cubature Kalman filter for nonlinear dynamical systems with packet dropout

This study investigated cooperative cubature Kalman filtering for discrete-time nonlinear systems with packet dropout based on a dynamic event-triggered mechanism. Initially, a dynamic event-triggered mechanism was constructed based on the measurement state information to reduce communication burden and energy consumption. Subsequently, we proposed the distributed filter for each sensor node, which was designed to handle random packet dropouts. This filter employed the minimum mean squared error approximation technique and weighted average consensus method under the established data transmission mechanism. A cooperative cubature Kalman filter algorithm with enhanced precision and robustness was then developed. Furthermore, sufficient conditions were established to ensure the boundedness of the prediction error and stochastic stability of the designed filtering algorithm. The findings indicated that the packet loss rate was upper-bounded and contingent on the average communication rate per node, thereby guaranteeing that prediction-error covariance of the local filter remained bounded at every moment. Finally, the proposed algorithm was applied to track maneuvering targets using multiple unmanned aerial vehicles , and simulation results demonstrated its efficacy and practicality. • This study presents an adaptive event-triggered mechanism based on measurement information and a dynamic event-triggered cooperative Cubature Kalman filter for nonlinear systems with packet loss. • Stochastic stability is analyzed, and simulations demonstrate that the proposed algorithm achieves strong robustness and high filtering accuracy.