Data Fusion In Multi-sensor Integrated Navigation System

With the development of maritime industry, higher and higher precision, reliability and fault tolerant ability of navigation systems are required for accurate and reliable ship positioning. In the past few years, navigation system has been developed from single sensor system to multi-sensor integrated system. The processing methods have also been developed from single system processing methods to multi-sensor data fusion methods. In this way, the precision and reliability of navigation system are improved significantly. An integrated navigation system composed of four navigation systems including GPS, GLONASS, Beidou and Loran_C navigation systems is studied in this thesis. Data fusion algorithms are studied to improve the precision and reliability of the integrated navigation system.Firstly, the basic principles, general models and applications of data fusion are briefly introduced. The basic composition and positioning principles of the four navigation systems are also separately introduced. Secondly, centralized and distributed data fusion systems are investigated separately. In both systems, the weighted average algorithm,α-βfilter tracking algorithm and federal filtering algorithm are appliedand studied. Finally, these two data fusion systems are simulated on the MFC platform of VC++ software and compared. The simulation results show that, in centralized system, through weighted average, the precision is improved and throughα-βfilter tracking, both the reliability and the precision are improved. The results also show that in the distributed system, because federal filter is applied, the precision, reliability and the fault tolerant ability are all improved.