Research On Cycle Slip Detection And Correction Of Gnss Real-Time Dynamic Data

One of the important factors affecting the precision of GNSS carrier phase positioning is cycle-slip, a reliable detector for jumps is particularly significant in real-time dynamic GNSS positioning. With two different frequency signals, it systematically analyzes and confirms the effects and shortages of some detectors in the thesis. By improving the traditional detectors, it is possible to detect and estimate cycle-slips for real-time processing.Firstly, based in the three frequency carrier phase data after GPS modernization, it illustrates the general forms of linear combinations and analyzes the noise and characteristics of combined observations as well as ionospheric refraction and so on. Meanwhile, the thesis also investigates the source and characteristic of cycle-slip after introducing the theory based in GNSS carrier phase data. How to build a detector and several common detectors are both presented in the thesis.Secondly, with single frequency the paper discusses the performances and applications of non-difference methods such as high-order difference and Kalman filter after experiments and so as that approaches with dual-frequency signals such as geometry free combination, pseudorange minus phase linear combination and Melbourne-Wubbena combination.Owing to the shortages of standard Kalman filter, it takes benefit of the priori information estimated with the third-order polynomial smoothing and the data to be estimated by the improved Kalman filtering making the detection and correction of cycle slips in real-time to be possible. After testing static and dynamic data with improved Kalman filtering and geometry free combination respectively, the results indicates that jumps of more than3cycles can be declared by the improved Kalman filtering. Indeed, compared with the standard filtering, the improved algorithm increases the sensibility of detector such as the successive cycle slips and allows making a reasonable enough detector for more applications. The experiments show that the improved approach is effective.A general detector for detecting and repairing cycle-slip in real-time data processing is developed, which combined with the improved Kalman filtering and geometry free combination after analyzing their characteristics respectively. The estimating by improved Kalman filtering can be performed parallel to the detection. A real-time simulation test is implemented using low dynamic GPS vehicle measurement data. The results show that the real-time detecting algorithm can identify cycle-slips of more than1cycle; moreover, the estimates are right in general. The real-time detecting algorithm can provide an effective and reliable detection, also for small jumps, multi-jumps, successive jumps and some special jumps that could not be detected with geometry free combination.