The Research And Application Of Three-dimensional Positioning Navigation Base On Interactive Multi-model Algorithm

With the increasing traffic problem,in order to avoid the traffic congestion problem,reduce the traffic pressure,we often build overpass at the intersection,the traditional two-dimensional positioning navigation has been unable to use in this three-dimensional complex path positioning continuously,sometimes when we are driving on the bridge,the navigation system tell us we are under the bridge.Therefore,the height position of the vehicle needs to be accurately positioned.For this reason,we could add one kind height measurement tools,then through the appropriate filtering algorithm for data analysis and processing to improve the positioning accuracy of the vehicle.In this paper,the GPS receiver provide the height position and the atmospheric pressure altimeter offer the relative height as the systematic observation of the height position of the vehicle.The Unscented Ka lman Filter is used to filter the height position information of the vehicle under the nonlinear system model.It is found that the Unscented Kalman filter can be used to estimate the height of the vehicle.However,when the state parameters of the vehicle c hanges,such as acceleration,turning,etc.,there will be increased positioning error,the height difference between the bridge is relatively small,we need higher positioning accuracy,this can not meet the requirements.In order to solve this problem,the I nteracting Multiple Model algorithm is used to combining with the Unscented Kalman Filtering algorithm.It is an adaptive algorithm.Through the experiments,it can be seen that the positioning accuracy of Interacting Multiple Model Unscented Kalman Filtering algorithm is obviously better than Unscented Kalman Filtering algorithm,it can be described the states of the system well when the system states changed.In the many overpass scenarios simulated in this paper,we can realize the height information of the vehicle in real time and locate the vehicle in a certain bridge.The Compared to Unscented Kalman filter,the overall accuracy can be increased by about 30%.At the same time,also solved the special vehicles(such as traffic patrol)on the problem of loca tion,speed real-time monitoring,then it is easier to manage and al ocate resource.In this study I also found that the overpass often use the section cross model,from the same section but different lane,leading to the road of different directions and height.At this point,the height of the positioning accuracy can not distinguish the problem of bifurcation path selection,which is often result in lost,detour problem in a giant interchange scene.Therefore,based on the high-precision positioning of the three-dimensional position before simulation,this paper combines the lane-level GIS to solve the problem of how to locate the vehicle to a certain lane,so as to solve the choice of road bifurcation from the source.The research of this paper can provide a re ference for the vehicle positioning in the complex three-dimensional path,and how to determine the position and speed of the vehicle in the intelligent transportation system.