본문 바로가기 사이트 메뉴 바로가기 주메뉴 바로가기

KRRI Develops World’s First Drone and AI-based Safety Inspection System for Railroad Facilities

KRRI Develops World’s First Drone and AI-based Safety Inspection System for Railroad Facilities

Demonstration of Safety Inspection of a Railroad Bridge Using Autonomous Drones

CHEONGJU, South Korea – MAY 3 – On May 4th, Korea Railroad Research Institute (KRRI) (President: Seok-youn Han) unveiled its drone and AI-based automated safety inspection system designed to inspect difficult to access railroad facilities such as bridges or electric transmission towers.

Previously, most drones have been operated by human pilots but this new system involves autonomous drones.

About 100 people including government officials, researchers, personnel from railroad operating institutions and related associations observed the demonstration to see for themselves how well the system performed on site.

During the demonstration, two drones inspected the Miho River Bridge in Cheongju-si, Chungcheongbuk-do, including the bottom of the bridge deck, bridge piers, and other areas. Being evaluated at a C grade for safety, the Miho River Bridge is a challenge to access as it crosses the river.

This drone system allows for automated inspection of railroad facilities, such as bridges, which have so far been required manual inspection. Other expected benefits of the system include preventing accidents, reducing time and costs of maintenance, and big data-based systematic management of bridge safety.

※ Related research project (National R&D project by the MOLIT)
- Title: Development of Automated Inspection System for Access Vulnerable Railway Facilities based
on Unmanned Aerial Vehicle
- Research period and budget: Jun. 2019 – Jun. 2023 (4 years), KRW 9.2 billion
- Responsible researcher: Hyuk-Jin Yoon, Head of ICT Based Public Transportation Research
Department at KRRI
- Joint research organizations: Korea Railway Corporation (KORAIL), Korea National Railway (KR),
Softgraphy CO., Ltd., Seunghwa Technologies CO., Ltd., Sejong University
- Contract research organization: UCON System Co., Ltd.

This system consists of drones, a mobile ground control station to operate the drones, and damage detection software which automatically detects damage such as cracks by analyzing acquired photographs from the drone.

The Hexacopter drones, tailored to the operational conditions of railroads, feature six propellers so that they can withstand train-induced wind over railway tracks. Those drones are also unaffected by electromagnetic fields created by overhead power lines.

In addition, 3-D autonomous flight technology enables multiple drones to inspect railroad facilities consecutively, operating even in areas with a weak GPS signal, such as the bottom of a bridge deck. Their position is controlled to an accuracy within 30 cm.
The mobile ground control station delivers an inspection scenario to drones through wireless communication and analyzes facility safety using acquired photographs from the drone.

Damage detection software improves accuracy by applying AI algorithms to acquired photographs from the drone. Specifically, the software groups the photographs based on the types of railroad facilities and displays damage in two or three dimensions so an inspector can easily identify the location and extent of damage.

When the mobile ground control station transmits an inspection scenario to drones specific to each railroad facility, the drones automatically perform the entire procedure, from takeoff through inspection to landing. Then the photographs go through processing to diagnose external damages and automatically provide a digital exterior damage map.

KRRI’s new drones passed a performance evaluation at the country’s first performance evaluation facility for autonomous drones, demonstrating their suitability under railroad conditions including train-induced winds and electromagnetic fields.

The drones have gone through a range of tolerance tests, such as cross winds, updrafts, gusts and electromagnetic waves emitted by railway power lines, and demonstrated excellent operation and safety, maintaining position control accuracy to less than 50 cm under extreme railway conditions; up to 10 m/s crosswinds and gusts of 8 m/s.

“This technology significantly reduces time and costs of maintenance for railroad facilities and prevents safety accidents,” said Hyuk-Jin Yoon, Head of the ICT Based Public Transportation Research Department at KRRI.” KRRI is committed to technological development so that this autonomous inspection technology can be applied to different fields as well.”

“The system is a world-class railroad technology that integrates drone and AI technologies as the Fourth Industrial Revolution technologies with railroad technologies,” said KRRI President Seok-youn Han. “KRRI is committed to develop an inspection and maintenance system for railroad facilities which utilizes big data on safety status.”

※ Reference: Drone and AI-based automated inspection system for railroad facilities








Press Contacts

Yongmook KANG
Korea Railroad Research Institute
ymkang@krri.re.kr
+82 31 460 5184