Fuzzy Federated Kalman Filter and Its Application to Integrated Navigation System
The most important property of federated Kalman filter is information-sharing principle. That is sharing the information in the sub-filters without changing the amount of total information. For the fraction values of local filters in the federated Kalman filter are set forehand, It cant reflect the situation in time, which decreases convergence rate of filter. On the other hand, when using low-order model for decreasing the burden of computer or having little information of accurate model, the performance of filter decreases greatly. Therefore, the effectiveness of filter decreases. In this paper, the ideal of fuzzy filter, which can adjust the fraction values of local filters on line, is introduced. The fuzzy federated Kalman filter can also adjust the process noise information in the filter on-line. According to the ideal introduced above, a simulation is taken on an INS/GPS/Doppler integrated navigation system.
Huang Hao Deng Zhenglong
Dept. Of Control Engineering,, Harbin Institute of Technology
国际会议
北京
英文
63-66
2007-11-04（万方平台首次上网日期，不代表论文的发表时间）