Research Of Algorithms On Improving GPS Single Point Positional Precision

This article takes the application of GPS positioning system as a background, takes the enhancement of GPS absolute and relative pointing accuracy and the ant jamming virility as the goal, it conducts the research separately in the modern non-linear Kalman filter technology as well as the phase smooth technology's application to enhance the GPS absolute and relative pointing accuracy.In GPS absolute single point position, this paper uses one kind new precise GPS simple point localization model and one kind Sigma Point Kalman Filtering algorithm based on weighting statistics linear regression in view of the flaw in the commonly used simple localization model and extended Kalman filtering algorithm. This algorithm carries on the on-line auto-adapted adjustment to the random variable statistical property, so it obtains a distinct enhancement in the filter accuracy of nonlinear system. Then aims at the information fusion algorithm's flaw because of the feed back causing the bad robust ability in the federal Kalman filtering for GPS/DR integrated navigation system. It proposes one kind information fusion factor method to improve the filter effect based on the failure detection function. Afterwards it obtains good effect by using the Sigma Point Kalman filtering algorithm in this application.In the GPS relative localization, this paper studies the pseudorange and phase smooth technology's pointing accuracy and its rule, then proposes one kind of improvement real-time false distance phase smooth technology. This technology uses partitions of the pseudorange and its the freeze weighting factor to smooth, the simulation experiment indicates that using the partition can obtain the quite ideal real-time locating result by choosing suitable freeze factor. Moreover in carrier phase difference localization application, this paper introduces the basic principles of both the integer least square method and the LAMBDA method for resolving the integer ambiguity solution. Through a short baseline example and compared with the unlimited least square method, the results show that, LAMBDA method considers the ambiguity's integer characteristic and also decorrelates ambiguity's covariance, so it meliorates the covariance field of ambiguity, avoid the covariance's discontinuity. So the LAMBDA method improves the search area and the related positioning precision apparently.