Yang Y.,Tianjin Navigation Instrument Research Institute |
Wu X.-T.,Tianjin Navigation Instrument Research Institute |
Yang J.-L.,Tianjin Military Representative Office |
Gao W.,Tianjin Navigation Instrument Research Institute |
Pei Z.,Tianjin Navigation Instrument Research Institute
Zhongguo Guanxing Jishu Xuebao/Journal of Chinese Inertial Technology | Year: 2014
A distributed Kalman filter alignment algorithm is developed in order to use the parallel computing ability of the distributed architecture gravimeter to solve the non real-time implementation of filtering based on one single processor. Firstly, the error equations of the azimuth strapdown platform gravimeter are deduced, and the state equations and observation equations are built. Secondly, an error covariance analytical method is applied to the filtering equations, and the system is decentralised into two subsystems with the same dimension. In this way we get the initial alignment filter formed by the two subfilters. Finally, the azimuth strapdown platform model is built by using Matlab, and stationary base alignment is implemented by using global Kalman filter and distributed Kalman filter separately. The simulation results show that the distributed filter has the same filtering accuracy and costs only 60% of time compared with the global one, which is favorable to ensure the real-time performance of the algorithm.