摘要
A novel neural network based on iterated unscented Kalman filter （IUKF） algorithm is established to model and com- pensate for the fiber optic gyro （FOG） bias drift caused by temperature. In the network, FOG temperature and its gradient are set as input and the FOG bias drift is set as the expected output. A 2-5-1 network trained with IUKF algorithm is established. The IUKF algorithm is developed on the basis of the unscented Kalman filter （UKF）. The weight and bias vectors of the hidden layer are set as the state of the UKF and its process and measurement equations are deduced according to the network architecture. To solve the unavoidable estimation deviation of the mean and covariance of the states in the UKF algorithm, iterative computation is introduced into the UKF after the measurement update. While the measure- ment noise R is extended into the state vectors before iteration in order to meet the statistic orthogonality of estimate and mea- surement noise. The IUKF algorithm can provide the optimized estimation for the neural network because of its state expansion and iteration. Temperature rise （-20-20℃） and drop （70-20℃） tests for FOG are carried out in an attemperator. The temperature drift model is built with neural network, and it is trained respectively with BP, UKF and IUKF algorithms. The results prove that the proposed model has higher precision compared with the back- propagation （BP） and UKF network models.
A novel neural network based on iterated unscented Kalman filter （IUKF） algorithm is established to model and com- pensate for the fiber optic gyro （FOG） bias drift caused by temperature. In the network, FOG temperature and its gradient are set as input and the FOG bias drift is set as the expected output. A 2-5-1 network trained with IUKF algorithm is established. The IUKF algorithm is developed on the basis of the unscented Kalman filter （UKF）. The weight and bias vectors of the hidden layer are set as the state of the UKF and its process and measurement equations are deduced according to the network architecture. To solve the unavoidable estimation deviation of the mean and covariance of the states in the UKF algorithm, iterative computation is introduced into the UKF after the measurement update. While the measure- ment noise R is extended into the state vectors before iteration in order to meet the statistic orthogonality of estimate and mea- surement noise. The IUKF algorithm can provide the optimized estimation for the neural network because of its state expansion and iteration. Temperature rise （-20-20℃） and drop （70-20℃） tests for FOG are carried out in an attemperator. The temperature drift model is built with neural network, and it is trained respectively with BP, UKF and IUKF algorithms. The results prove that the proposed model has higher precision compared with the back- propagation （BP） and UKF network models.
基金
supported by the National Natural Science Foundation of China(61104184
40904018)
the National Key Scientific Instrument and Equipment Development Project(2011YQ12004502)
the Research Foundation of General Armament Department(201300000008)
the Doctor Innovation Fund of Naval University of Engineering(HGBSCXJJ2011008)
the Youth Natural Science Foundation of Naval University of Engineering(HGDQNJJ12028)
作者简介
Corresponding author.Feng Zha was born in 1984. He received his B.S. degree from the First Aeronautic College of Airforce in 2006. He received his M.S. and Ph.D. degrees from Naval University of Engineering in 2（X）8 and 2012 respectively. Currently he is a lecturer in Navigation Engineering Department of Naval Uni- versity of Engineering. His research interests are optic gyro and application in rotating inertial naviga-tion system. E-mail： zha_feng@ 126.comJiangning Xu was born in 1963. He received his B.S. degree from the Northwestern Polytech- nical University in 1983 and M.S. degree from Naval University of Engineering in 1986. He re- ceived his Ph.D. degree from Southeast University in 2002. Currently he is a professor in Navigation Engineering Department of Naval University of Engineering. His research interests are inertial technol-ogy and application. E-mail： xujiangning @hoimail.comJingshu Li was born in 1985. He received his B.S. and M.S. degrees from the Airforce Aviation University in 2008 and 2012 respectively. Currently he is a Ph.D. candidate in Naval University of En- gineering. His research interests are inertial technology and application. E-mail： elvisstef@ 163.tomHongyang He was born in 1990. He received his B,S. degree from Northwestern Polytechnical Uni- versity in 2011. Currently he is working towards an M.S. degree in Naval University of Engineering. His research interests are inertial technology and application. E-mail： elvisstef@ 163.com
