Korea Expressway Corporation is accelerating efforts to reduce traffic accident casualties by introducing artificial intelligence (AI) and digital technology across highway safety management. The key is moving beyond simple campaigns and manpower-based inspections to build precise, data-based safety infrastructure. Thanks to these efforts, highway traffic accident deaths on the government-financed highways managed by the corporation fell by 24 over five years, from 171 in 2021 to 147 last year.
Digital Road Safety Diagnosis Introduced, Along With AI-Based Automatic Pothole Detection
Road inspections have long relied mainly on managers' skilled experience and visual observation. As a result, there were limits to accurately identifying minute surface irregularities or poor drainage in environments where vehicles travel at more than 100 kilometers per hour. To overcome this, the corporation introduced "digital road safety diagnosis" technology, which combines LiDAR sensors with digital transformation (DX) technology.
A vehicle equipped with LiDAR sensors precisely scans the road's shape, and the collected data is used to create a digital model that reproduces the road's condition. When this reveals causes of rainy-road accidents such as rainwater backflow and water pooling, follow-up measures such as installing early drainage facilities or repaving are carried out. The corporation preemptively applied the technology to 10 accident-risk sections where similar accidents repeatedly occur, and plans to expand it to about 60 sections this year.
On the highway sections managed by 59 branch offices nationwide, the "AI-based automatic pothole detection" system is used to scan regularly twice a month. The method covers a vast stretch of 216,000 kilometers annually, quickly identifying potholes in urgent need of repair to eliminate accident causes. As a result, pothole damage compensation last year fell to 3.486 billion won, down 16% from 4.153 billion won in 2024.
Big Data-Based Drowsiness Index (DDI) Introduced
Management of drowsy driving among cargo trucks was also strengthened. Over the past five years, the leading cause of highway traffic accident deaths was drowsiness and inattentive driving, with cargo truck fatalities accounting for 82% of the total (341 of 414). This is because cargo truck drivers are easily exposed to drowsiness due to long hours and long-distance driving. Based on traffic big data, the corporation developed a "Drowsiness Index (DDI)" that analyzes 18 variables such as accidents, driving behavior and environment, dividing drowsiness risk into four grades.
The drowsiness index was applied to 20 major routes that take cargo truck driving characteristics into account. When a driver travels continuously for two hours in a high-risk Grade A section, a drowsiness-prevention voice guidance and music play from the cargo truck navigation app. During safety patrols, location information is linked so that a siren and prevention voice are automatically transmitted upon entering a drowsiness-risk section. The cargo truck navigation safety score rose to 92 points last year, 14% higher than the previous year's 81 points. The corporation plans to expand the drowsiness index to all routes going forward and to broaden prevention services in cooperation with logistics and transportation companies.
AI CCTV Monitoring System and AI Patrol Introduced
As important as accident prevention is how quickly sudden incidents can be recognized and responded to. On highways, every second is directly linked to additional casualties from secondary accidents. The corporation is pushing forward with an "AI traffic control construction project" through the end of this year, which merges closed-circuit television (CCTV) on highways nationwide with AI to detect sudden incidents in real time. AI will be applied to about 4,000 of the roughly 9,500 CCTVs nationwide, automating the entire process from recognizing sudden incidents to providing information. Rather than simply displaying screens, it applies a multi-stage analysis structure to automatically identify stopped or wrong-way vehicles and pedestrians. Through this, the corporation aims to raise sudden-incident detection accuracy from the previous level of 50% to more than 96%. It is also providing a public-private cooperation service since July that merges collected incident information with KakaoNavi data to inform drivers of accident information in real time.
An "AI Patrol" for on-site response was also introduced. Personnel dispatched to accident scenes share real-time information with the control room through smart safety helmets, reducing report-writing time and allowing them to focus on accident handling. As a result of a pilot introduction at three branch offices through April this year, traffic accident casualties in the sections under those branches' jurisdiction fell by 22%.
Using AI to Crack Down on Illegal Vehicles That Cause Falling-Object Accidents
The crackdown on illegal vehicles that disrupt highway order and cause large falling-object accidents has been improved from officials' visual methods to AI methods. The "AI regulation violation crackdown system" uses smartphones in safety patrol cars to detect vehicles violating regulations, such as improper loading and designated-lane violations, in real time during patrols. It then automatically generates the report content and reports it to the Safety Report platform.
The "AI automatic improper-loading identification system," aimed at cracking down on improperly loaded vehicles that could lead to large falling-object accidents, has been in place since 2020 and will be expanded to all lanes (359 toll offices and 474 lanes) by the end of this year. This system photographs vehicles entering tollgates to continuously and automatically determine whether loading is improper, improving inspection efficiency and minimizing screening wait times. As a result of its introduction, the average number of inspections per worker per day rose from 300 to 1,230, quadrupling work efficiency.
"By preemptively introducing and actively utilizing AI technology, we will strengthen road management and illegal-activity monitoring," a corporation official said. "If AI and digital technology are introduced across highway management, casualties from traffic accidents can also be greatly reduced."
