Anagrafe della ricerca

Sustainable management of infrastructures such as bridges requires regular and accurate inspections and monitoring, which are crucial to avoid tragic stories such as the Morandi Bridge one. Traditional methods for bridge inspection have relied on operators who exploit snooper trucks or lifts to access the infrastructure and present challenges related to safety, high costs also due to the possible necessity of temporary closures, human factors linked to the experience and subjectivity of the personnel. Mobile inspection robots have been proposed as a tool to aid bridge inspections. Their key advantages include the ability to access areas of a bridge that are otherwise difficult to inspect and to automate the process of collecting and processing data thus increasing repeatability and reliability of inspections. However, costs and coverage per unit time still present challenges. Recently, drones and related sensing technologies such as visual/thermal cameras/Lidars have become a fundamental asset to complement other measurement sources (in situ and spaceborne observations) thus enabling cost-effective maintenance of many structures. Nevertheless, technological gaps remain which hinder the full exploitation of drone potential for intelligent infrastructure inspection: automated visual inspections and reliable flight operations in complex environments under various visibility and weather constraints present challenges such as poor GNSS satellite coverage, magnetic interference due to metal elements, and challenging micro-weather conditions in the proximity of infrastructures. That is why we propose "4IPLAY - Improving Intelligent Infrastructure Inspection by Pushing UAS Level of AutonomY in challenging environments", a project that will demonstrate in relevant environments the concept of repeatable automated visual inspection based on autonomous drones. 4IPLAY builds on the know-how experience of a multi-disciplinary research team and will focus on three main pillars: - Automated acquisition and processing of multi-spectral data using artificial intelligence approaches. - Advanced flight control logics which are tailored for the conditions encountered in the proximity of large infrastructures. - Distributed multi-drone architectures to support real time cooperative navigation and off-line trajectory reconstruction. 1-to-N planning and control will be implemented, also considering tight integration of cooperative navigation and distributed sensing. These concepts will be demonstrated in high fidelity simulation environments, also exploiting advanced tools for generation of synthetic scenarios which represent a digital twin of the considered infrastructure. They will then be tested in flight exploiting experimental facilities already available at the research units. The project will pave the way for repeatable autonomous multi-drone-based inspection and monitoring, with significant industrial impact and societal benefits