Research On Information Fusion Algorithm For Integrated Navigation Of GPS&IRS Based On QAR Data

It is mainly focuses in this dissertation on the research of fusion algorithm of GPS/IRS integrated navigation based on QAR data which includes transformation of different coordinate system, data processing model of GPS/IRS integrated navigation, improved high order UKF, and the applications of the PCA and FastICA. The main works and new standpoints are listed as follows:Firstly, the principles of GPS/IRS integrated navigation are summarized. An overview on transformations in different navigation coordination is provided. The data processing in earth-fixed coordinate system frame and the filter models of GPS/INS integration are also provided.Secondly, high order UKF in tracing and estimating an airplane ignores the influence and analysis of noise. To solve this problem, improved high order UKF is used in the application of calculating QAR data. After transformed by the system function and measurement function,the prior probability distribution of prediction value is obtained. The observation then is added by the algorithm to solve the optimal result. Simulations based QAR data show that improved high order UKF has better performance as compared with normal.Thirdly, to improve the accuracy of GPS/IRS when GPS outages occur, the idea of blind source separation is introduced. Comparative study of PCA method, and FastICA method, the results showed that: FastICA and PCA algorithm have the advantage over retained linear features and details, while the PCA algorithm has the advantage over noise filtering than FastICA. When GPS is unavailable, the two BSS algorithm is used to aid IRS, the accuracy of GPS/IRS can be improved.The experimental results based on QAR data proved that the algorithms presented in this dissertation are valid and the integrated navigation system designed is practicable.