148 student works on display, with AI- and robot-based solutions tackling fraud prevention, emergency medicine and industrial safety

A future where AI flags voice phishing in real time and robots perform CPR at disaster sites is drawing closer. A showcase of creative student research aimed at solving problems in industry and everyday life opened Thursday at Korea University of Technology and Education.

The university is hosting its 32nd Graduation Research Exhibition for the 2026 academic year at Damheon Silhak Hall on campus from Thursday through Friday.

The exhibition is one of the university's flagship engineering education programs, in which third- and fourth-year students design and build technologies and products applicable to real industrial settings, drawing on their major coursework and hands-on training. This year, 148 works submitted by departments including mechanical engineering, electrical and electronic and communications engineering, computer engineering, design engineering, architectural engineering, and energy and new materials engineering are on display.

A standout feature of this year's exhibition is the large number of projects using AI and robotics to address challenges in society and industry.

ScamGuard: A data-driven fraud risk analysis platform. Users submit suspicious materials — text messages, KakaoTalk chat screenshots, PDFs, images or voice recordings — and an AI analyzes them to assess the likelihood of fraud and present evidence-based findings. [Korea University of Technology and Education]
ScamGuard: A data-driven fraud risk analysis platform. Users submit suspicious materials — text messages, KakaoTalk chat screenshots, PDFs, images or voice recordings — and an AI analyzes them to assess the likelihood of fraud and present evidence-based findings. [Korea University of Technology and Education]

Computer engineering students Lee Jun-young, Kim Du-hyeon and Yu Chang-yeon presented ScamGuard, a data-driven fraud risk analysis platform. The service lets users submit suspicious materials — including text messages, KakaoTalk chat screenshots, PDFs, images and voice recordings — and uses AI to assess the probability of fraud and explain the reasoning behind its findings.

The research team built a database of more than 20,000 fraud cases to help prevent voice phishing and online scam losses. The project drew praise as a practical protective tool at a time when phishing crimes exploiting AI are on the rise.

In the emergency medicine category, a project by mechanical engineering students Lee Woo-won and Lee Won-hyeok drew attention: an automated chest compression robot capable of treating multiple patients simultaneously. The robot was designed to address the shortage of CPR personnel in large-scale disaster situations.

It uses deep learning-based contactless heart rate measurement and an eye-state analysis algorithm to assess a patient's condition, then delivers chest compressions at a consistent depth and speed.

A robotic system that detects patient condition using rPPG deep learning algorithms for contactless heart rate (BPM) measurement and an EAR algorithm for eye-state analysis, then performs chest compressions at a consistent depth and speed via a mock circulatory loop (MCL). [Korea University of Technology and Education]
A robotic system that detects patient condition using rPPG deep learning algorithms for contactless heart rate (BPM) measurement and an EAR algorithm for eye-state analysis, then performs chest compressions at a consistent depth and speed via a mock circulatory loop (MCL). [Korea University of Technology and Education]

Research in industrial safety and disaster response also drew considerable interest.

Design engineering students Nam Gwang-hyeon and Moon Seo-jin developed U-AXIC, a universal emergency escape system for public buses that enables rapid evacuation in accidents involving flooding, rollover or fire. The system combines accident-detection sensors, an automatic emergency reporting function and illuminated escape-route guidance to improve response in crisis situations.

From the electrical, electronic and communications engineering department, an AI-powered automatic Braille label maker — which converts image, text and voice information into Braille output — was selected as an outstanding work. The technology was recognized for its potential to improve information accessibility for people with visual impairments.

The architectural engineering department presented research on optimizing patrol routes for quadruped robots monitoring construction sites. The team developed a path-planning algorithm that allows robots to navigate efficiently through construction environments filled with obstacles and uneven surfaces, improving the technology's real-world applicability.

The energy and new materials engineering department developed a hydrogen sulfide gas sensor for smart farm safety monitoring. The research team built a real-time monitoring and alert system linked to a mobile app, offering a way to protect both workers and livestock.

University President Yoo Gil-sang said this year's graduation projects carry special significance because students translated AI and robotics into practical technologies addressing real problems in industry and daily life. "We will actively support students in growing into problem-solving talent that society and industry need," he said.


fact0514@heraldcorp.com