Geomatics and Soft Computing Methods for Infrastructure Monitoring

Abstract

Our society is heavily dependent on many interdependent and complex critical infrastructures. Deficiencies in the functionality of the transportation network (e.g., vehicular traffic interruptions or limitations) can cause enormous inconvenience to communities and people. The Italian transport infrastructure heritage and new infrastructure construction is so relevant that the issue of its preservation and safety has become a priority. Specialistic advice is therefore required to understand if the static behaviour of these infrastructure has changed significantly after extraordinary events (e.g., earthquakes, landslides). With the advent of the internet of things (IoT), infrastructures are becoming smart and procedures simpler. In the framework of smart infrastructure development, we implemented an experimental system that integrates soft computing and geomatic methodologies for solving early warning problems. This system, which has been tested on the Petrace bridge (Southern Italy), is able to generate forecasting information on the infrastructure behaviour over time, mainly exploiting geomatic parameters. We built this "early warning/predictive" system through integration of several significant (geometric/structural) infrastructure models, which have been merged into a final "type" model. The results derived from various possible scenarios have been implemented in a neural network. The only system’s input is represented by displacement measurements acquired by sensors placed on the infrastructure, and the output consists in an estimation of different risk levels. Sensor data are then transmitted to a control unit that sends them to a processing server, where the calculation system is hosted. All received data and model results are displayed on the Wordpress platform with colour codes calibrated on the calculated risk thresholds.