Analysis of error models in Precise Point Positioning

Most of the error sources in Precise Point Positioning (PPP) are mitigated using predictive models. One of the factors that limit the accuracy of PPP is the effects of unmodeled errors or incorrect assumptions of predictive models. The objective of this thesis is to develop methods and algorithms to isolate errors in GPS (PPP) for the purpose of error analysis and predictive model testing and verification. Three errors analyzed are troposphere error, site displacement error and receiver phase windup error. The troposphere delay and site displacement varies with location and time and the receiver phase windup varies directly with the antenna orientation. This thesis presents two methods for isolating the troposphere delay using concepts from static relative positioning. If the troposphere delay is known at one of the antenna position then the troposphere delay at the unknown station can be estimated. The first method computes the coordinates of the unknown antenna, along with the troposphere delay, while the second method only computes the troposphere delay. The author presents an algorithm to isolate the site displacement as the difference between the instantaneous coordinates of the GPS antenna at a particular epoch and the reference coordinates, such as the ITRF coordinates, using algorithms similar to satellite single-differenced PPP. This thesis shows how the phase windup due the receiver antenna can be measured using the time difference of the geometry-free single-difference phase measurement generated from two GPS receivers. The result can be used to determine unmodeled errors or incorrect assumptions of predictive models used in PPP.