A Semi-Autonomous UAV with Human Supervisory Control for Non-Destructive Inspections in Interaction with Concrete Structures

Abstract

Inspection and assessment of large concrete structures is the primary method of monitoring their status and is conventionally conducted by trained inspectors and climbers with specialized equipment, which can be expensive and ineffective at times. Nondestructive inspection solutions emerge as a good candidate for facilitating such evaluations. In this article, we present a human supervisory control to automate both the approach and interaction stages of the measurement process employing a multirotor featuring tiltable rotors, characterized by a streamlined kinematic model. Our software seamlessly integrates with the PX4 autopilot firmware, thereby harnessing all the capabilities of the flight controller. Furthermore, this integration enabled us to leverage all compatible tools such as QGroundControl, log review functionalities, and others, optimizing efficiency and functionality. Thanks to that, it has been possible to test our system without an experienced and/or certified drone pilot on the concrete pillars of a highway bridge, reducing the time and the costs needed to deploy and validate such a robot.