MELETI: A Machine-Learning-Based Embedded System Architecture for Infrastructure Inspection with UAVs

Telecommunication and power infrastructures are located in hard to reach environments and exposed to extreme weather conditions. As such, dedicated personnel is set to conduct inspections on a regular basis to maintain the quality of these critical infrastructure networks. However, these inspections are costly and usually pose significant risks to personnel. Hence, UAVs have found flourishing ground in the area of infrastructure inspection. In this chapter, we introduce UAV inspection requirements and propose MELETI, a system architecture that investigates how UAVs can be used to enable autonomous infrastructure inspection through innovative intelligent systems and machine learning algorithms. Finally, several infrastructure inspection challenges that implement MELETI are introduced on both power and telecommunication infrastructure inspection.