Entity

Time filter

Source Type


Juqing Z.,Changan University | Jianliang N.,Geodetic Data Processing Center | Yafang X.,Changan University
Proceedings - 4th International Conference on Intelligent Computation Technology and Automation, ICICTA 2011 | Year: 2011

Due to different levels and characteristic deformation existing in any images, it is difficult to select an appropriate model to correct effectively. Considering BP neural network is a high and nonlinear complicated system which can approach at any precision, we attempt to use it for image geometric correction. In this paper, errors correction of scanning map by BP neural network as an example was discussed. The result shows that the accuracy of scanning map corrected by the BP neural network is higher than by normal function fitting. © 2011 IEEE. Source


Nie J.,Geodetic Data Processing Center | Nie J.,Changan University | Yang Y.,Xian Jiaotong University | Wu F.,Xian Jiaotong University | Wu F.,Zhengzhou University
Acta Geodaetica et Cartographica Sinica | Year: 2010

A modified particle filtering is proposed. The convergence speed of the particle filtering is tried to be improved. The influences of linearization of nonlinear functional models and the non-Gaussian random errors to the results of dynamic precise point positioning will be weakened. In the new procedure, the free-ionosphere ambiguities are fixed at first to reduce the number of parameters in the state vector. The accuracy of the initial positioning results is improved and the convergence of the particle filtering is modified. Kalman filtering as predicted filtering of particle filtering is employed to improve the efficiency of the important sampling of the particle filtering and the precision of the sampling particles, as well as to slow down the degeneracy of the particle. An actual dynamic GPS data set is employed to test the new particle filtering procedure. It is shown that the modified procedure of the particle filtering based on fixing free-ionosphere ambiguities can improve the accuracy of the dynamic precise point positioning. Source


Wu F.,Zhengzhou University | Wu F.,Xian Research Institute of Surveying and Mapping | Nie J.,Geodetic Data Processing Center | He Z.,Changan University
Wuhan Daxue Xuebao (Xinxi Kexue Ban)/Geomatics and Information Science of Wuhan University | Year: 2012

In GPS/INS integrated navigation, the number of observations is usually less than that of the state parameters, and the single adaptive factor is usually applied in Kalman filtering, which can lead to precision loss of indirect observational parameters. A new algorithm of classified adaptive filtering is presented based on predicted residuals and selecting weight filtering, and the corresponding formulas are given. Finally, an actual calculation is given. The new algorithm can not only degrade the influence of the disturbances from the state but also avoid the loss of estimated precision of indirect observational parameters, and improve the accuracy of the navigation further. Source


He Z.,Changan University | Wu F.,Zhengzhou University | Wu F.,Xian Research Institute of Surveying and Mapping | Nie J.,Geodetic Data Processing Center
Wuhan Daxue Xuebao (Xinxi Kexue Ban)/Geomatics and Information Science of Wuhan University | Year: 2011

The influence of the prior covariance errors to the standard dynamic Kalman filtering is discussed. The influence expressions of the prior covariance matrix errors including the covariance matrix of state parameters, dynamical model errors and measurement noises are deduced. A GPS/INS tight integration navigation is performed, and it shows that the unreasonable errors of the covariance matrices of the dynamic model information and measurements will result in biases of the dynamic navigation results. The minus errors of the covariance matrix of dynamical model information will increase the navigation errors. If the errors of the covariance matrix of the predicted states are positive, then the effects of the dynamical model error will be weakened. However, contrary conclusion could be got if only the errors of the covariance matrix of measurement noises are considered. Source


He Z.,Changan University | Nie J.,Geodetic Data Processing Center | Wu F.,Zhengzhou University | Wu F.,Xian Research Institute of Surveying and Mapping | And 2 more authors.
Wuhan Daxue Xuebao (Xinxi Kexue Ban)/Geomatics and Information Science of Wuhan University | Year: 2012

Kalman filter is widely used in the area of kinematic positioning and navigation. However, it doesn't have the ability to resist the influence of measurement outliers, hence its performance is easy impacted by the observation outliers or kinematic state disturbing. In order to guarantee the reliability of the navigation with precise dynamic model, a model set, which contains many different observation models, is established. An improved Kalman filtering, in which the design matrix of the observational model is substituted by its expectation is proposed to control the influences of the measurement outliers. An integrated GPS/INS navigation example is given to show that the modified Kalman filtering algorithm works well. Source

Discover hidden collaborations