Abstract:
Early Warning System (EWS) for monitoring megastructures deformation, natural hazards, earthquakes, and landslides can prevent economic and life losses. Nowadays, Real-Time Precise Point Positioning (RT-PPP) plays a vital role in this domain since it relies on precise real-time measurements derived from a single receiver, provides real-time monitoring and global coverage. Nevertheless, RT-PPP measurements and methodology is very sensitive to outliers in products, latencies and changes in the constellation geometry. Consequently, there are long initialization periods, losses of convergence and different noise sources, with a high impact on the warning system's availability or even led out to initiate false warnings. This study presents the first experiment to propose a methodology that can help the decision-makers confirm the warning based on the probability of the detected movement by using machine learning classification models. For this, in the firs experiment, a laser engraving machine device was modified to simulate deformations. A control unit will be designed based on open-source software, Python libraries are implemented, and the G programming language used to control the device motions. All this research will be the background on which the early warning service will be developed.
