Research On Dual MIMU/Magnetometer Technology Of Pedestrian Navigation System

PNS(Pedestrian Navigation system)is kind of equipment whose main function is to provide real-time positioning service or autonomous navigation for rapid response force in unknown or known environment.In the procedure of responding to emergencies,precise navigation and positioning is of great significance to the successful implementation of the task under unpredictable conditions of dynamic surrounding.In the development process of research on improving navigation accuracy,the main idea is to combine chief device with other auxiliaries.As the representative of high-precision satellite navigation system,GPS(Global Positioning System)become one of the ideal choices in PNS naturally.In addition WiFi monocular or map matching can also effectively work.But whether GPS WiFi or cameras are restricted to external conditions,that is to say when a pedestrian is in rough environment which is sheltered or poorly lighted,a drawback will be exposed in corresponding navigation scenario which results in large fluctuations of localization quality.Thus in order to make application environment unrestricted,to meet different needs of PNS,the study of autonomous navigation technology is up to much.MIMU(Micro Inertial Measurement Unit)based on MEMS technology is mostly used in fully autonomous navigation system as the main sensor,which is often mounted on shoes.In this way we can use ZUPT(Zero Velocity Updates)to conduct sensor error correction,but there is defection during ZUPT procedure.That gyroscope in up direction is of poor observability affects accuracy of heading and finally results in inferior localization.In this paper we introduce an error reduction method called MIMU/magnetometer based on QSF(Quasi Static Field)to offset the shortage of measurement information during ZUPT.We construct the observed variables by comparing the difference between the changing rate and the predicted value of output along the magnetometer triaxial at QSF time when the carrier moves in order to provide extra information.And finally realize the purpose of improving positioning accuracy with estimating the gyroscope error parameters.In this way the excogitation based on MIMU works more reliably and practically without the signal of GPS or other auxiliary devices.Form the viewpoint of increasing the number of sensor,we propose a pedestrian navigation program which conduct mutual correction between double MIMU through sensor fusion because of the inferior in autonomous method compared with non-autonomous ones.That is to say a pedestrian will equip two foot-mounted MIMU(include MEMS gyroscope,MEMS accelerometers,MEMS magnetometer).Due to the physical space description of the two systems,a second order nonlinear constraint equation can be obtained.The suggested information fusion method is expected to make use of nonlinear equality constrained optimization Kalman Filtering,in this way the secondary navigation error correction is done and ultimately improve the positioning accuracy.Build a true navigation system and design corresponding positioning experiment like straight line and closed rectangle trajectories.The result shows good localization effect in a short time without outside aid information.Thus the effectiveness and feasibility of MIMU based pedestrian navigation scheme are demonstrated finally.