The study, published in the journal City and Built Environment, outlines a system that utilizes Global Navigation Satellite System (GNSS) technology and artificial intelligence to monitor the positions of elderly drivers and agricultural machinery while simultaneously identifying road defects such as potholes and cracks. The primary objective of this design is to address the rising safety risks associated with an aging population and the deteriorating road infrastructure in regions experiencing demographic decline.
The motivation for this system stems from demographic shifts and the specific vulnerabilities of older drivers. According to the 2022 United Nations World Population Prospects report, the global population is estimated to reach 8.5 billion by 2030 and 10.9 billion by 2100. In South Korea, where 70% of the population resides in urban areas, the proportion of elderly drivers is increasing rapidly. These drivers, along with operators of agricultural machinery, face higher risks of traffic accidents due to reduced physical abilities and slower reaction times.
The core of the proposed system is a dual-function data collection and sharing platform. On one hand, it tracks the real-time locations of vehicles operated by elderly drivers and agricultural machinery. On the other hand, it collects and verifies information on road hazards using AI-equipped devices mounted on delivery vehicles in urban areas and citizen-participation vehicles in low-traffic suburban areas. This comprehensive approach aims to provide general drivers with real-time warnings about potential hazards and the presence of vulnerable road users, thereby preventing secondary accidents and improving overall traffic management.
To achieve high-precision positioning, the system employs GNSS satellite technology combined with Network Real-Time Kinematic (RTK) correction signals. The design utilizes the flächen korrektur parameter (FKP) method, which generates correction information through a set of triangular surfaces or cells applicable to specific areas. This correction data is derived from multiple reference base stations, including three national and two internal reference stations. By leveraging these signals, the system aims to measure vehicle positions in real time with an accuracy of within 5 cm. This level of precision is critical for rapidly moving vehicles such as cars and agricultural machinery, ensuring that the location data transmitted to navigation applications is reliable and timely.
The detection of road hazards relies on the YOLO V4 algorithm, a computer vision model chosen for its balance between inference speed and mean average precision. The researchers selected YOLO V4 over more recent versions, such as YOLO V8, because it achieves excellent detection performance with relatively fewer resources. The algorithm is trained to recognize specific road hazards, including potholes, road cracks, falling objects, and poor road surface markings. This automated detection process is a key component of the platform’s ability to identify infrastructure issues that may require urgent maintenance.
Data collection is structured to cover both high-traffic urban environments and low-density suburban areas. In urban settings, delivery vehicles equipped with AI-equipped video devices and sensors collect road hazard information at regular intervals. In regions with low traffic volumes and risks of population decline, the platform supports the participation of citizen drivers who use data-collection devices mounted on their personal vehicles. This citizen-participation model is designed to ensure comprehensive coverage of road networks that might otherwise be neglected due to reduced traffic. The collected data, which includes location information, road view video, and 3-axis accelerometer readings, is transmitted to a central control center for processing and verification.
The system architecture includes a situation control room that plays a vital role in validating the collected data. The control room is organized into two teams: the Situation Monitoring Team and the Situation Response Team. The Monitoring Team is responsible for real-time oversight of the data construction and distribution system, identifying issues that require action. The Response Team then performs the necessary tasks to verify the data and ensure the continuous operation of the system. Only after this verification process is the data distributed to the navigation applications of general drivers.
The platform also facilitates the prompt communication of urgent maintenance needs to relevant authorities. This feature aims to improve road maintenance, particularly in regions facing population decline where infrastructure management may be lacking. By enabling rapid response to road defects, the system seeks to break the cycle of deteriorating road conditions and enhance the safety of the road network for all users, especially those most at risk.
The study highlights the importance of integrating advanced geospatial technology with community participation to create safer and more efficient road environments. The design goes beyond existing road information construction, which has primarily focused on precision mapping for autonomous vehicles, by incorporating real-time hazard detection and location sharing for vulnerable road users.
Despite the detailed design, the study acknowledges a significant limitation: it lacks a practical evaluation involving real citizens. The current work focuses on the design and piloting of the system, but it has not yet been fully validated in a live environment with actual user participation. The authors note that future research will involve a pilot program in Gimcheon, South Korea, to conduct system validation and assessment.
By sharing real-time data between vulnerable road users, general drivers, and maintenance authorities, the platform aims to reduce the incidence of traffic accidents and improve the overall quality of road infrastructure.
The research contributes to the field of safe mobility by providing a comprehensive design for a data platform that addresses multiple aspects of road safety simultaneously. It combines the precision of satellite-based positioning with the speed of AI-driven hazard detection and the reach of citizen participation. The findings suggest that by enhancing the flow of information between different stakeholders in the road ecosystem, it is possible to create a safer environment for elderly drivers, agricultural machinery operators, and all other road users.
Subject of Research: Design of a data platform for safe mobility using GNSS and AI
Article Title: Design of data platform for safe mobility
Article References: Lee, J., Kang, E., Lee, J., & Hwang, J. (2025). Design of data platform for safe mobility. City and Built Environment, 3(1), Article 14. https://doi.org/10.1007/s44213-024-00039-y
Image Credits: AI Generated
DOI: 10.1007/s44213-024-00039-y
Keywords: safe mobility, GNSS, YOLO V4, road safety, elderly drivers, data platform, urban infrastructure, South Korea, Design, data, platform, safe
News Source: Denise Maddox. (November 12, 2026). Design of data platform for safe mobility. Scienmag.
