Nonlinear system state estimation mechanism based on kalman filter for wireless sensor networks localization

The article deals with a class of state vector estimation problems of nonlinear systems, which is derived from single node localization in wireless sensor networks. And a new strong adaptive Kalman filter mechanism is implemented by combining the original nonlinear filtering algorithm such as Square-root Cubature Kalman Filter (SCKF) with Kalman filter. Firstly, the mechanism utilizes the state estimation algorithm based on SCKF to estimate and correct the state vector in state-space model. Kalman filter is then performed for further processing due to the linear changes of state equation. Furthermore, the strong adaptive filter mechanism with Extended Kalman Filter (EKF) is established for comparative purposes, and the Cramer-Rao Bound (CRB) based on the nonlinear model is also derived. Finally, to verify the effectiveness of the mechanism, numerical simulation is made. Results analysis illustrates that the proposed mechanism has high location accuracy and is better than that of the original filtering algorithm without strong adaptive recursion. © 2018 Computer Society of the Republic of China. All rights reserved.