The Research Of Adaptive Unscented Kalman Filtering In Integrated Navigation Systems For Land Vehicles

That micro inertial navigation systems which based on MEMS technology are integratedwith other navigation systems (such as satellite navigation systems,celestial navigationsystems and so on) can avoid only one navigation system bringing drawbacks and improveintegrated navigation systems’ navigation accuracy. Further needs is put forward forintegrated navigation systems against carriers’ high maneuvering, high dynamic and satellitesnavigation’s signals temporarily missing in urban areas, forest areas, tunnel and so on.Therefore, it is very important how to design and improve integrated navigation systems’navigation algorithms, improve navigation information’s effect of integration for enhancingintegrated navigation systems’ navigation accuracy and extending micro inertial navigationsystems’ practical fields.This thesis’s research objects are the land vehicles. The vehicles’ kinematic models anddynamic models are set up for vehicles’ translational and rotational by theoretical knowledgeof kinematics and dynamics. Micro inertial sensors’ sources are divided by analyzing theIMU’s (Inertial Measurement Unit) structure in the integrated navigation systems. Resources’(especially the stochastic error sources) models are established in the micro inertial sensors ofthe land vehicles’ integrated navigation systems.Adaptive unscented kalman filter navigation algorithm is brought forward throughanalysing current information fusion technology and researching integrated navigationsystems’ integrated navigation algorithms in the integrated navigation systems in landvehicles. But for the land vehicles motion characteristics and GPS’s signals frequentlymissing, only one simple navigation algorithm is not satisfied for the navigation performancerequirements. A new integrated navigation algorithm-adaptive unscented kalman filteralgorithm based on this problem is proposed. This new algorithm can fuse the aboveimproved kalman filter algorithms’ advantages, overcome land vehicles’ motion propertiesand remain higher navigation accuracy in case of the satellites’ signals’ missing.Based on the land vehicles’ kinematics and dynamics’ models’ founded and the microinertial devices’ error models’ established, adaptive unscented kalman filter is analysed, set upand tested simulated in the vehicles’ navigation systems. Results demonstrate that adaptiveunscented kalman filter algorithm has higher navigation accuracy and better performancecompared with alone improved kalman filtering algorithms in the navigation systems.