AI Weekly Malaysia

Back to items Summaries

APU students develop AI-powered smart evacuation system

ID

10619

Status

summarized

Published

04 Aug 2026, 4:12 PM

Fetched

04 Aug 2026, 4:18 PM

Provider

Digital News Asia

Category

malaysia-tech

Original URL

https://www.digitalnewsasia.com/digital-economy/apu-students-develop-ai-powered-smart-evacuation-system

Source URL

https://www.digitalnewsasia.com/rss.xml

Excerpt

AI-powered system adapts evacuation routes in real time to improve safety during emergencies Award-winning student project combines smart sensing, IoT and edge AI to protect building occupants Ever wondered what would happen if an emergency exit sign could do more than point to a fixed doorway? Three students from the School of Computing (SoC) and School of Technology (SoT) at Asia Pacific University of Technology & Innovation (APU) asked that question and turned it into an award-winning solution. Their invention, the Lumina Smart Evacuation System, earned first runner-up at the Tech 4 Good Challenge 2026 Grand Finale, held in Batu Kawan, Penang, on 25 July 2026. The team advanced through preliminary screenings and prototype showcases to become one of 24 finalists from nearly 190 participating teams. The team comprises Lai Zi Huey, an Information Technology student specialising in Business Information Systems; Low Wei Ling, a Software Engineering student; and Woo May Eng, an Information Technology student specialising in Cloud Engineering. Traditional exit signs indicate predefined evacuation routes but cannot respond when those routes become unsafe because of fire, smoke, congestion or other hazards. Lumina addresses this limitation through a decentralised system capable of assessing conditions and adapting evacuation guidance in real time. Using RGB cameras, thermal infrared and acoustic sensors, together with edge artificial intelligence, dynamic routing and floor-projected guidance, the system identifies hazards and redirects occupants towards safer evacuation routes. Designed with people at its centre, Lumina also incorporates multi-sensory guidance to support elderly occupants and individuals with sensory impairments. Jointly organised by ViTrox Technologies and a local university, the Tech 4 Good Challenge encourages young innovators to apply technologies such as artificial intelligence and the Internet of Things to solve real-world problems. Guided by Dr Kamalanathan Shanmugam, senior lecturer at SoT, and Assoc Prof Dr Muhammad Ehsan Rana, associate professor at SoC, the students refined their technical architecture, tested their prototype through simulations and strengthened their ability to communicate complex engineering concepts. Lai said, "Building Lumina enabled us to transform classroom knowledge in system architecture, interface design and programming into a functional prototype. "We applied real-time processing, YOLO-based computer vision and dynamic A* routing to identify safer evacuation paths, while integrating sensors, controllers and alerts. Our mentors guided us through the IoT learning curve, helping us refine the Detect-Decide-Guide concept and unite the software, dashboard, routing logic and hardware into one coherent system." Testing conducted by the team showed that Lumina achieved a 58.6% reduction in 90th percentile evacuation time. Its IoT-enabled "Pull Policy" stop-line mechanism was also designed to manage cr