Experimental assessment of the accuracy of GPS and RTS for the determination of the parameters of oscillation of major structures

Based on experimental evidence we explore the possibility of using GPS and robotic total stations (RTS) for measurements of oscillations of relatively rigid structures (modal frequencies up to 3-4 Hz). Our strategy was to compare uni-axial oscillations of known characteristics with simultaneous recordings of both GPS and RTS, and analyze obtained time series to determine amplitude and frequency of oscillations. The conclusion of this study is that GPS can record oscillations up to 4 Hz with a minimum amplitude of 5-10 mm with an accuracy of a few millimeters, and that RTS can record peaks of oscillations with submillimeters to a few millimeters accuracy, but at high frequencies some cycles were lost. Based on recordings of both the instruments frequencies of oscillations were also accurately determined, though noise seems to increase with increasing frequency. Spectral analysis was based on the least-square-based software which permits one to analyze discontinuous, short, and non-equispaced time series. The latter derive either from GPS signal outages/imperfections and hardware/software imperfections, or from a non-constant rate of sampling for RTS. The overall conclusion is that the adopted computational approaches permit us to overcome some main disadvantages of these instruments and to use conventional GPS and RTS instrumentation for a wide range of cases of structural monitoring, especially if displacements relative to an independent coordinate system are required.