Innovation-based adaptive Kalman filter with sliding window for integrated navigation

Considering that it is difficult to select window width of the innovation-based adaptive Kalman filter for the reason that tracking accuracy and tracking sensitivity of noise are contradictory to each other, an innovation-based adaptive Kalman filter with sliding window for integrated navigation is proposed. This method designs the gradient detection function of noise statistical characteristic to sense real-time variation of noise statistical characteristics. The window width obtained in real time by using window adaptation function makes the window slide in the preset range to adapt to the actual noise. Simulation results show that the innovation-based adaptive Kalman filter with sliding window for integrated navigation can track real-time variation of noise statistical characteristics effectively, and adaptive tracking accuracy and tracking sensitivity is considered simultaneously. © 2017, Editorial Office of Systems Engineering and Electronics. All right reserved.