The Korea Institute of Civil Engineering and Building Technology’s Ground Research Headquarters develops AI inspection drone technology that can assist with safety inspections of aging tunnels. The appearance of the AI drone inspecting the tunnel. /Courtesy of Korea Institute of Civil Engineering and Building Technology

An artificial intelligence (AI) drone has been developed to detect damage and cracks in urban underground expressways and tunnels.

The Ground Research Institute of the Korea Construction Technology Institute noted on the 15th that it has developed AI inspection drone technology that can assist in the safety inspection of aging tunnels.

As of the end of 2023, there are 2,892 tunnels in South Korea. Among them, 172 are aging tunnels over 30 years old, accounting for 5.9% of the total. This number is expected to increase to 740 by 2033, representing 25.6% of the total.

The most critical part of tunnel inspection is the ceiling. Currently, visual inspections are mainly conducted using high-altitude work vehicles. This method occupies the road, hindering vehicle flow and posing safety issues for inspection workers.

The research team has overcome the limitations of existing inspection methods by integrating AI and autonomous driving technology into the drone technology. The newly developed drone utilizes a 200-meter long-range indoor positioning sensor to navigate freely within the tunnel with an error margin of 20 cm. It can also perform tunnel inspections and vehicle operations simultaneously, thereby minimizing inconvenience for citizens and ensuring safety during inspection tasks.

The research team highlighted that the most significant feature of the newly developed drone is its ability to autonomously synthesize unusual concrete damage scenes that can only be observed in aging infrastructure. Previously, cracks were detected by processing data collected on-site, but the newly developed generative AI can create data so meticulously that it is indistinguishable from actual footage.

Park Seon-kyu, the director of the Korea Construction Technology Institute, said, “This generative AI will solve the limitations of field application due to a lack of training data in concrete inspection using artificial intelligence, and it will further enhance the safety of aging infrastructure.”