Research On GNSS Kinematic Relative Positioning Algorithms

GNSS Kinematic Relative Positioning is a new kind of positioning technology based on the static relative positioning technology and quasi-kinetic relative positioning technology, which can be applied to aerial refueling, spacecraft docking, carrier-based aircraft landing and so on. It’s a hot research direction in the study of GNSS.This paper studies the GNSS relative positioning algorithms, including observation data processing, error correction, ambiguity searching and data process model. Main work and innovations are summarized as follows:1. Taking into account the dynamic relative GNSS applications in space, this paper analyzes carrier Terrestrial Service Volume, Space Service Volumn visibility problems from the perspective of the signal’s launch angle, and the signal power of the earth shelter, and derived in a variety of space visibility under formula.2. A Study on the GNSS Kinematic Relative Positioning observation model, error model to analyze the characteristics of the main error sources, effects and common ways to correct. For relatively dynamic features of GNSS, formulas are derived on the varying degrees of impacts on GNSS relative positioning caused by satellite ephemeris error, ionospheric and tropospheric delay error, and errors with distance, height and other factors changes.3. In dynamic conditions, a study on the problem of ill-conditioned variance matrix caused by high data sampling rate and inadequate observed data from two aspects: looking for more prior information and looking for more data means. After comparing the truncated singular value decompsiton, ridge estimation, regularization approach, this paper found these methods are of not much difference. And after the comparison of increasing more priori information and more data means in decreasing the ill condition, this paper come to a conclusion that taking advantage of prior conditions is a better way than through data means.4. On the basis of the original angle function method, a new and easier way in determining Carrier double difference observation equations priori covariance matrix is proposed and proven more effective than the angle function method after the comparative study on other ways in determining double difference prior covariance matrix5. In GNSS Kinematic Relative Positioning process, the Carrier’s filtering model will be easiely influenced by colored noise and ill-conditioned matrix because of inadequate prior information and high dynamics. Thus wrong judgements will be made in ambiguity which can not be tested by Ratio value and OVT method. This paper proposed a new ambiguity recognition criteria based on variability of Ratio value, reducing ambiguity confirmed bug rate, proven effective.6. A study on problems in detecting and repairing cycle slip in kinematic positioning. The detection and repairing of cycle slip is one of the key skills in kinematic positioning. Through the analysis of conditions and causes of cycle slip, this paper finds that unrelated geometric volume is one of the most appropriate test method in detecting cycle slip.7. A study on feasibility of Doppler method in kinematic positioning. There is no unified view on the accuracy in Doppler observation. The author compared datas got from Different types of receivers on the zero baseline. A conclusion that the accuracy of Doppler observation is mainly related to the types of receivers is drawn. Doppler method can be used in kinematic positioning but definite kinestate and observation accuracy must be offered.8. A study on the principle and characteristics of methods in detecting cycle slip, such as Doppler method, combination of pseudo range and carrier phase observation, the ionosphere residual method and triple frequency combination observation. And the effectiveness of these methods are tested. This paper finds that all of the methods can be used in detecting and repairing cycle slip, but with their limited range of application. So they should be used coordinatedly. And a conclusion that the triple frequeny combination observation is most suitable in detecting and repairing cycle slip is drawn.9. Aiming at characteristics of dynamic relative positioning, carrier body dynamic model and filter model were investigated, the dynamic model of vehicle that you want to opt for different dynamic model of the process, not the more complex models as possible, on the filter model the study shows that in complex dynamic conditions, filter model to be able to suppress divergence, to take advantage of the state of motion vectors and other prior information, can inhibit motor acceleration and observational data the impact of outliers, concluded that the anti- poor adaptation filtering satisfy the above conditions.