Multipath interference cancelation in GPS time series under changing physical conditions by means of adaptive filtering

Abstract

Multipath, a highly autocorrelated signal is observable phenomena during time periods longer than the sidereal period of Global Positioning System (GPS) satellites in their constellations. Multipath pattern and sidereal repeat time interaction are examined in terms of time series correlations of topocentric coordinates using GPS code and phase observations collected at high and low frequencies. The horizontal and vertical components at 5 and 30 s sampling rates are analyzed to detect and remove multipath classified by their M P 1 levels and baseline lengths. For the selected 11 stations that have 24-hour data from CORS stations in the USA and Turkey, the repeat times of the GPS constellation were identified as a function of maximum cross-correlation of adjacent time series. The multipath interference signal from two-day time series data is removed by an adaptive filter to improve time series of coordinate estimates. After the filtering process, an effective epoch based position tracking has been accomplished, especially in the height component. The position errors disappeared to maximum possible extent in the pseudo-range measurements. Up to 70% improvement for the phase derived coordinates in reflective environment has been achieved over the baselines ranging from 1 km to 175 km. For the phase observables, the filter success directly related to the multipath level which is specific to the station.