Distributed Divided Difference Filtering Based on Covariance Intersection over Sensor Network

This paper explores the distributed state estimation (DSE) for the nonlinear systems over a sensor network. Divided difference filtering (DDF) is one of the well-known sigma-point Kalman filtering algorithms, which can be categorized into the first-order divided difference (DD1) filter and the second-order divided difference (DD2) filter. Compared with the extended Kalman filter (EKF), DDF possesses numerous advantages. However, a corresponding distributed filtering algorithm has yet to be developed. Therefore, this paper combines the covariance intersection (CI) fusion algorithm and DDF to propose two distributed DDF type consensus algorithms, called the covariance intersection-based first-order divided difference (CI-DD1) filter and the covariance intersection-based second-order divided difference (CI-DD2) filter, which retain the advantages of distributed filters and the DDF. The effectiveness of the two proposed algorithms is demonstrated via a typical air traffic control scenario.