Research And Implementation Of Multi-source Fusion Navigation Algorithm Based On Factor Graph

Nowadays,with the continuous advancement and development of technology,the requirements for navigation and positioning systems are getting increasingly demanding,and traditional navigation and positioning systems usually only use a fixed combination of navigation sensors for multi-source fusion navigation and positioning in a single specific environment,while in reality there are various complex and changing environments and various task requirements.For example,the GPS signal can easily not be received in an obstacle-obstructed environment,or the GPS signal may fail due to external interference.In order to solve the above problems,the research on all-source positioning and navigation technology is gradually intensified,which takes the multisource fusion navigation system as the basis and provides a unified method and framework for the fusion of multiple navigation sensors.The study of multi-source fusion navigation algorithms that satisfy the plug-and-play feature is of great significance to the development of all-source positioning and navigation technology.In this context,this thesis addresses the above-mentioned issues in a related research work.Firstly,this thesis makes a detailed study on the principle of integrated navigation system,and studies the common reference coordinate system and how to convert different coordinate systems.The mechanical arrangement principle of strapdown inertial navigation system is studied.Finally,the relevant knowledge and error model of INS,GNSS,CNS,DNS,RA and ADS navigation sensors are analyzed respectively.Secondly,this thesis studied the factor graph model and its algorithm theory,this thesis introduces the relevant concept definition and factors of factor graph model and integrated algorithm theory,based on the maximum a posteriori estimation of factor graph algorithm do the corresponding analysis,the final factor figure incremental smoothing algorithm is studied in detail,lay a theoretical foundation for later algorithm study.Thirdly,this thesis analyzes the system state equation and its system measurement equation of INS/GNSS/CNS/DNS/RA/ADS integrated navigation system,and through the factor graph model and algorithm theory research,proposes a multi-source fusion navigation algorithm based on factor graph,and constructs its navigation algorithm fusion framework.The fusion-reset federated filtering algorithm is compared with this algorithm to verify the simulation.The simulation results show that the accuracy of the navigation data obtained by the two algorithms is the same,but the performance of the former is better when the navigation system works normally.However,when some sensors fail in navigation system,the performance of the multi-source fusion navigation algorithm based on factor graph is obviously better than that of federated filtering algorithm,and the former shows flexible plug-and-play characteristics.Fourthly,this thesis designs a whole algorithm source navigation system research to validate the simulation platform,the platform is based on the Qt software development tools to carry on the design,can be applied to a variety of navigation sensor simulation and research validation of multi-source information fusion navigation algorithms,and introduces the general design of the simulation platform,the main module design and main interface design,Finally,the simulation platform is tested and verified.