Study On Low-cost BDS/DR Integrated Navigation Filter Algorithm

Intelligent transportation is getting more and more attentions from people as a new sunrise industry, it is a systemic engineering that include many technologies, among them, real-time, accurate, continuous high precision navigation and positioning of the vehicles is a key technology of intelligent transportation. In current, the satellite navigation and positioning has become one main method of vehicle navigation and positioning, it can provide the vehicle with real-time and three-dimensional position, so it is referred to as the cornerstone of intelligent transportation system. The beidou navigation system(BDS) in our country is one of the world’s top four satellite navigation system, it is making steady progress according to the national general layout idea, and has the capability of navigation service for nationwide and surrounding areas in the moment. As an emerging industry of the national strategy, it will become a more and more important role in our country’s and even the globe’s vehicle navigation and positioning services market. This article based on low cost car BDS/DR integrated navigation positioning technology as the research object, mainly around the principle of BDS/DR integrated navigation positioning, centralized Kaman filter and combined Kaman filter of information fusion to carry out systemic research. The main innovative points of this paper summarized as follows:1. If the observed value include error, Kaman filtering has problem of filtering divergence when used in navigation and positioning. In order to achieve the goal of so-called robust estimation, traditional robust Kaman filters to limit the gain matrix through iterative calculation by using the properties of predicted residuals. But the process of iterative calculation greatly reduces the computation efficiency of dynamic positioning. This paper put forward a robust EKF algorithm based on Vondrak gross error detection. The new algorithm effectively solves the problems of traditional robust Kaman filter which needs to iterative calculation; it not only improves the efficiency of calculation, but also ensures the resistance effect of gross error.2. In order to make the selection of information distribution coefficient of federal Kaman filter more reasonable, this paper proposes a new method which can adjust the information distribution coefficient of federal Kaman filter adaptive. This method takes the difference of BDS and DR in positioning and velocity and acceleration into consideration, two groups of distribution coefficient was used to determine the information distribution of filter. The experimental results show that the new algorithm has obvious promotion in the navigation accuracy compared to filter without self-adaption, and also improved the fault tolerance of combination system at the same time.