The Information Fusion Technology In Vehicle Integrated Navigation System

Vehicle integrated navigation technology is an important part of the Internet of Vehicle and it plays an irreplaceable role in city traffic. With the GPS technology widely used in vehicle navigation,it brings people a great convenience and safety.Then due to poor reliability of GPS and low positioning accuracy and bad adaptation in the environment,It is often unable to accurately locate vehicles in some special cases. DR has a relatively strong autonomous navigation and high positioning accuracy in a short time, people begin to look for a data fusion method in combination of two navigation systems. With the computer using on Information fusion technology that will be multiple sensors together the data collected and analyzed by the weight of each data value, the optimal state of the system by making and estimation. Therefore, by using multi-sensor data fusion technology we can combine the GPS with DR to a integrated navigations system and this is the development trend of the navigation.Firstly,this paper introduces the development of vehicle integrated navigation and its application, and makes presentation of current foreign research. Then the advantages and disadvantages of GPS and DR positioning principle are analyzed.Secondly,this paper studys on the application of kalman filtering in integrated navigation system and derives discrete linear and nonlinear systems of equations Kalman filtering. The Beidou navigation system(BD) is introduced into the integrated navigation system, the data on the GPS / BD / DR Integrated Navigation Fusion technology to do a study of the state of the system is derived equation and measurement equation. For separate DR systems and GPS / DR integrated navigation system.we do a filter estimation and simulate in Matlab. The GPS/BD/DR integrated navigation system is simulated in Matlab too.In the end, this paper proposed PF for nonlinear system based on KF. Because of the problem of degradation of the particles set,resampling is proposed to improve it.Aiming at the problems of PF we propose UT transform theory and UPF filtering,and a new filter based on STPF strong tracking filter. Matlab simulation results improved particle filter in terms of accuracy and error localization is better than ordinary PF.