Theory And Application Of Navigation And Positioning Based On Magnetic Beacon And Signal Of Opportunity

The Global Navigation Satellite System(GNSS)is used widely for positioning and navigation.However,harsh environments,such as underground,indoors,etc.,might block the GNSS.This shortcoming of the GNSS may be overcome with the use of multi-source navigating systems.The main investigation areas possible for a multi-source navigating system are: potential navigating source,fusion algorithm,and confidence evaluation of the source.The present thesis would be focused on these three aspects.First,to seek the available source upon the GNSS failure,the Opportunistic Navigation(OpNav)algorithms were investigated.In the case of sparse fingerprint condition,Jacobian logarithm and embedded compensation were proposed for improving the fingerprint-matching algorithm.Furthermore,a virtual-source algorithm for OpNav was proposed to resolve the problem of biased or unknown SoO beacon position.The proposed algorithms were verified experimentally.Next,for a stable,accurate,and robust positioning and navigation in harsh environments,a magnetic based navigation system(MBNS)based on low-frequency magnetic field was investigated,and error identification and compensating methods were proposed for the magnetic beacons and sensing the magnetic field.It was established that the MBNS was highly accurate and robust,and could serve as a reliable navigation source in harsh environments.The next step was the construction of the confidence evaluation model of the source based on information entropy.Fusion weight adaptive-estimating method was proposed for the evaluation of the result,following which,an optimal set contributing to a further accurate fusion result could be selected according to the evaluation.Finally,among the multi-source fusion algorithms,Extend Kalman Filter(EKF)-based multi-source information fusion algorithm was investigated.The algorithm was improved in the fusion models which were fused one by one following the evaluation of the result in comparison to the one fused together.In order to overcome the limitation of the EKF algorithm,which requires accurate statistical values for the characteristics of system noise,the Set-Membership Filtering(SMF)-based fusion algorithm was proposed.Owing to the gain matrix,the design of which is based on the upper limit of the noise,the SMF algorithm exhibited good performance when the system had a low upper-limit of noise and large measurement noise,as verified by the simulation.Considering the flexibility and extensibility of the system,the adaptive weighted fusion algorithm was simplified using Factor Graph(FG)to resolve the issue of frequent addition and deletion of SoO.The findings in this thesis were verified using a pedestrian navigation system comprising OpNav,MBNS,MIMU,Magnetometer,and Altimeter.