The Theory Of Real-time Kinematic Positioning With Browser/Server Model Based On Long-Range Reference Stations Network

This thesis deals with the new generation technique of RTK, named the network RTK technique that is applied widely., which integrates multiple continuously operating reference stations into one combined network to support navigation and positioning services, especially for centimeter-level carrier phase-based positioning on medium range. Not only the construction cost of a continuously operating reference station is very expensive, but also the requirement of the location is special, for example:good circumstances of observing satellites, perfect communication facilities and no electromagnetic that influences the satellite signal and so on. Then in order not to lose the positioning accuracy, if the long-range network RTK technology can be carried out, it can be multiplied to reduce the siting number of continuously operating reference stations and save construction and maintenance costs; and the problem of Network RTK system construction in the region where we have difficulty in stations construction can also be resolved.Currently, all Network RTK system softwares use the client/server mode, Mobile positioning terminal needs to install specific client software, such as the most common difference positioning algorithm. The more installed algorithm is, the higher requirements the receiver’s OEM boards and processor are, which result into the high prices of the positioning terminal. When an algorithm needs to be improved, the receivers are upgraded one by one; the process is complex and redundant. If the client software can run on the server and the client only needs the results and do not care about the procedure, the receiver achieves the reunification and zero maintenance, which is only accompanied by data capture, forwarding and remote monitoring. The server runs all network RTK data processing software, establishes a services platform based on location information and can extend services. This is the dynamic positioning services system with browser/server mode.The theory of Network RTK and methods of processing all kinds of errors were introduced and researched fully in this dissertation, and a system of beneficiating and quality analysis of data are developed on reference station. On the basis of this work, an algithim of kalman filter used to resolve ambiguities between long-range with estimating atmospheric parameter is firstly researched and completed. Then the high-precision model of regional errors was established and the approach of generating virtual reference station observation is provided. And the method of dynamic ambiguities resolution on rovers at single epoch was researched. In addition to assisting the NRTK technology, the long distance RTK algorithm with single reference station as one of the goals of this paper is studied. Finally, based on the reference station network, a dynamic positioning service platform with B/S model is putted forward.(1) The GNSS foundational observations are intruduced such as GNSS carrier phase and pseudo range, and the commonly used combinations of GNSS observation are discussed. Then mathematical equations of differential positioning are deduced and the characteristics and the processing methods of all kinds of GNSS positioning errors are analyzed in detail, the corresponding error statistical information and prior parameters are obtained. On this basis, the software is developed and runs on reference station, which beneficiating and quality analysis of data.(2) Double difference ionospheric delay is analyzed in detail and a parameter called the relative ionospheric zenith delay between long-range stations is used to estimate it. The algorithm is validated by actual measurement, that double differential ionospheric delay can be reduced to less5cm with the relative ionospheric zenith delay estimation and improve the ambiguity resolution efficiency.(3) A kalman filter algorithm is firstly researched and completed, which used to resolve ambiguities between long range stations with estimating atmospheric parameter. The algorithm needs dual-frequency observation. The range of baseline used to test is from48km to364km. If the observed condition is good, it only needs one epoch to fix integer ambiguity. Once the cycle slips or reacquisition of lost lock satellite occurs, it needs an epoch to resolve the integer ambiguity again.(4) The regional modeling principle of distance-dependent error is compared and analysized. It mainly contains:linear combination algorithm, distance-based linear interpolation algorithm, linear interpolation algorithm, low-order Surface algorithms and least squares collocation. The precision of above-mentioned algorithms interpolating double differential tropospheric delay and double differential ionospheric delay is validated.(5) In addition to assisting the NRTK technology, a kalman filter algorithm for long-range real-time kinematic positioning with one reference station is studied. The experimental results show that, the mean square error of one epoch dybamic positioning on48km baseline is3.06cm,2.5cm and3.3cm in the north, east and upper directions respectively. The same as above is1.43cm,1.05cm and4.26cm on100km baseline between static stations.(6) The ambiguity resolution method for rovers with single epoch was proposed. Firstly the wide-lane ambiguities are resolved by the geometric distance or combination equation of phase and pseudo range. Secondly, a proposed linear combination model of L1and L2carrier phase is used to resolve L1and L2ambiguities. The mean square error of positioning is1.7cm,1.9cm and3.4cm in the north, eastl and upper directions respectively.(7) The dynamic positioning services platform with browser/server mode is introduced, based on reference station network, extending service functions of the server, simplifing and standarding of the receiver, saving production and upgrade cost. The work flow and communication protocols and data format of B/S model is described.The algorithms in this dissertation were programmed and the software was developed by programming language of C#.net. Then the algorithms test and the results analysis were implemented by observation data of CORS network. The software which transfers data is developed by programming language of C++6.0. Finally, the theory of real-time kinematic positioning with browser/server model based on long-range reference stations network has been completed, which is highly reliable and available and provides new alternatives for CORS network construction.