Filtering Techniques to Detect Millimetric-Scale Dynamic Displacements

Ricardo E. Schaal and Ana Paula C. Larocca

Abstract:	Global Positioning System (GPS) can provide reliable information about dynamic behavior of large structures, particularly bridges that have deflections in the centimeters range [Roberts et al., 2001, 2002; Ogaja et al., 2001, Larocca, 2004a,b; Larocca and Schaal, 2005]. These bridges which have superstructure type of truss, box or I girder (steel or steel-concrete composite), orthotropic deck, for example, must have high rigidity just for preventing excessive deflection of girders and consequently failure. Measurements of millimetric-scale oscillations using GPS has not been extensively study by community due to constrains given by receiver noise, multipath and satellite constellation geometry, particularly in the vertical direction [Meng et al.,2002]. The purpose of this article is to present different mathematical tools to quantify millimetric-scale measuring of vertical dynamic displacement with good amplitude confidence, measured with L1 phase double-difference data. Most of experiments in the literature rely on the kinematics solution to obtain the displacements of the rover antenna attached to structure under test [Ogaja C., et al., 2000, 2001; Roberts, et al., 2005; Liu and Yao, 2007]. In this work it is analyzed the residuals from L1 phase double difference static solution with the purpose of testing the method for measurement millimetric amplitude dynamic displacements [Schaal and Larocca, 2001,2002.]. The residuals contain all uncorrelated phase measurements among satellites and receivers. The residuals contain all uncorrelated phase measurements among satellites and receivers. Choosing the reference satellite close to horizon in the phase double difference static processing, a vertical displacement of the antenna will show as a phase shift in the residual of a satellite close to the zenith. The Figure 1 shows the L1 double difference phase residuals (DDPR), converted to millimeter, in a 125 seconds session, with a pair of receivers Novatel Propack 3151RE with Garmin L1antenna, collecting 20 epochs per second, spaced in a very short baseline.. A 10 mm sinusoidal vertical displacement was imposed on the rover antenna during a period of one minute, starting at the thirty fifth second of the session. The controlled movements were applied using an antenna electro mechanical oscillator (EMO) illustrates on Figure 2. The residuals show very clear the uncorrelated random noise, spanning over 10 mm, the slow varying contribution due to multipath, ranging more than 15 mm and the one minute periodic displacement starting at 35th second. Most of the millimeter displacements are masked by the noise and multipath, requiring filtering tools to determine its amplitude and frequency with reliable standard deviations. Frequency determination is more straightforward because GPS time measurements are based on the on board atomic frequency standards. Amplitude is more cumbersome due to receivers and atmosphere phase noise. Figure 3 shows two DDPR obtained at a zero-baseline, one with the two satellites over 75 degrees of elevation and the other with the satellites below 15 degrees. The contribution of the atmosphere is very clear implying different approaches for different observing conditions.
Published in:	Proceedings of the 20th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2007) September 25 - 28, 2007 Fort Worth Convention Center Fort Worth, TX
Pages:	2178 - 2186
Cite this article:	Schaal, Ricardo E., Larocca, Ana Paula C., "Filtering Techniques to Detect Millimetric-Scale Dynamic Displacements," Proceedings of the 20th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2007), Fort Worth, TX, September 2007, pp. 2178-2186.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In