by DRONELIFE Staff Author Ian J. McNabb
A team of researchers led by Associate Professor Satoshi Suzuki from Chiba University recently announced that they’d developed an modern recent low-cost UAV designed for powerline inspection. While most inspection-focused UAVs use expensive and sophisticated sensor systems to make sure accuracy, the researchers claim that their recent drone, using a low-cost GNSS system, RGB camera and a millimeter wave radar, can achieve equally or more accurate results at close range, providing a more accessible alternative.
Using the comparatively easy sensor package, the drone identifies power lines and implements a knowledge-based line ID system, which employs a method called Hough Transform to remain in extremely close proximity without colliding. As well as, a heading correction module counters the consequences of magnetic fields on the UAV’s internal compass, calculating the specified direction via specified start and end points and mapping accordingly. Due to the comparatively low accuracy of the GNSS system, the researchers implemented a waypoint correction system to deal with the mismatch between the plotted and actual positions of the drone. Finally, environmental aspects affecting flight path (like strong winds) are countered via an auxiliary controller.
After testing on a 10kv line, the proposed method effectively addressed magnetic interference and hewed closely to the plotted route, but researchers were careful to spotlight that improved control algorithms could improve the already impressive results even further. Dr. Suzuki said, “By addressing the challenge of distant inspections, this method can ensure an uninterrupted electricity supply, contributing to the sustainable development goal of ensuring access to scrub and inexpensive energy. Furthermore, this method also holds the potential for carrying out automated maintenance work.”
The complete paper, including more detailed methodology and technology information, is obtainable here. The study was first published in .
Read more: