GPS Multipath Effect Mitigation Algorithm Based on Empirical Mode Decomposition

For short baseline GPS positioning, the multipath effect is the major error in GPS data since it can't be reduced by traditional double differencing technique. To deal with this problem, a GPS multipath effect mitigation algorithm is proposed in the paper based on the empirical mode decomposition (EMD) method utilizing the repetition characteristics of GPS satellites in every sidereal day. More specifically, decompose first the GPS data in day 1 and day 2 separately by EMD into intrinsic mode functions (IMFs). The correlation coefficient of sum of the last a few IMFs of day 1 and day 2 are calculated and the signal with maximum correlation coefficient value represents the multipath effect. After removing the multipath effect from the original GPS data the measurement accuracy can be highly improved. Measurements from a calibration test are adopted to verify the proposed algorithm. By comparing the actual displacement of the shaking table with the calculated one, it is found that: (1) The repetition phenomena of GPS satellite can be used to reduce the multipath effect. The proposed algorithm is a reliable method for reducing multipath effect in GPS data. For all the test cases, the measurement accuracy has been improved to some extent. (2) For sine wave with low vibration amplitude and frequency higher than 2Hz, the improvement of accuracy by the proposed method is not significant.