Design and Development of an Autonomous Caged Drone for Leak Detection in HVAC Ducts

A Master of Science thesis in Mechatronics Engineering by Ahmed Khalil entitled, “Design and Development of an Autonomous Caged Drone for Leak Detection in HVAC Ducts”, submitted in October 2021. Thesis advisor is Dr. Mohammad Jaradat and thesis co-advisors are Dr. Shayok Mukhopadhyay and Dr. Mamoun Abdel-Hafez. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).

Abstract

Interest in unmanned aerial vehicles (UAVs) has increased over the past years mainly due to the benefits that it proposes to achieve. Their ability to perform tasks accurately and efficiently in dangerous conditions makes them an optimal choice in various indoor applications. However, the problem of increased safety and precise control arises as these drones have high rotatory components. To avoid the rotary components being exposed, a cage type structure can be used which helps dampen an impact on the drone. For precise control, advanced adaptive control techniques can be used to achieve a stable vehicle path. In this thesis a hybrid rolling and flying multi-rotor drone is developed for inspection and leak detection in Heating, Ventilation, and Air Conditioning (HVAC) systems ducts through a thermal camera. A multi-rotor drone with a passive cage is designed, manufactured and controlled autonomously using Direct Model Reference Adaptive Control (DMRAC). DMRAC is coupled with a structure of PID controllers. Four methods are used and compared to overcome the drifting gain problem; dead-zone modification, sigma modification, epsilon modification and projection based modification. Additionally, more parameters are studied inside these modifications. The simulations show that all methods complete the desired route with some differences, such as gains drifting and error. The experimental multi-rotor platform is equipped with a 2D LiDAR sensor that provides a scan of the distance surrounding the multi-rotor drone in a 2D plane. An IMU sensor is used to project these distance measurements to a plane parallel to the ground along with a thermal camera for detecting leaks. The projected readings from the LiDAR is compared with the ground truth by using an Optitrack motion capture system. The platform is validated with PID and DMRAC control and the output result is compared amongst themselves, a thermal leak is detected and localized autonomously by the proposed system during real time validation.