Robust Cooperative Tracking for Aerial Maneuvering Target With Faulty Sensors

This article considers the robust cooperative aerial maneuvering target tracking problem based on the wireless sensor network with faulty sensors. The proposed robust $K$-means distributed cubature information filter (K-DCIF) algorithm is designed by three stages, namely, local filter, cluster, and consensus fusion. Each sensor has processing ability, which can be used to complete the local filtering stage individually. During the clustering stage, the $K$-means method is introduced to divide all the sensors in the sensor network into faulty sensors and reliable sensors. Then, the information matrix and the information vector obtained from the reliable sensors constitute information pairs during the consensus fusion stage. Based on the local neighboring interactions in the network, the accurate state information, such as position, velocity, and acceleration of the aerial maneuvering target, can be obtained by each sensor. Furthermore, by introducing a stochastic process, the boundedness of the estimation error of the K-DCIF algorithm with faulty sensors is proved. Finally, numerical simulation and equivalent experiment for maneuvering target tacking are given to validate the performance of the proposed algorithm.