INS aided fuzzy quality control for precise RTK GPS positioning

The global positioning system (GPS) has not only revolutionized navigation. Processing the signal`s carrier phase, its real-time kinematic (RTK) mode allows to guide machines with cm-level accuracy, to survey traffic routes precisely while driving along them, to set out points efficiently. Engineering geodesy has become unthinkable without RTK GPS. However, objects like buildings, trees or construction machines close to the antenna cause signal distortions which can hardly be modelled. These effects may bias the estimated position by several centimetres. They are extremely difficult to detect and mitigate while measuring kinematically. Consequently, RTK GPS is of limited use in critical applications, and its potential accuracy is not fully exploited. Additional sources of information are required to solve the problem. This project aims at investigating the capability of low-cost inertial navigation systems (INS) to provide this information. INS perform almost independently of the environment, and their properties are complementary to GPS. Originally developed for the guidance of missiles, they are now the core component of many navigation systems, typically integrated with other sensors. INS/GPS systems provide high long-term stability of the results and are robust with respect to short GPS outages. Different authors have shown that so-called GPS cycle-slips can be detected and repaired reliably by INS aiding. In this investigation, the quality control will be extended to multipath, diffraction effects, and outliers. Using MATLAB, an error model of low-cost INS shall be created and applied in a statistical analysis of minimum- detectable biases based on Kalman filtering. The algorithms will be embedded into a strategy for robust estimation and quality control using fuzzy logic, which the applicant has developed as part of his dissertation.