EdgeLink-X is an AI-driven heterogeneous edge-network platform built to provide resilient, low-latency connectivity for intelligent transportation and railway environments. The system forms a multi-layer architecture that unifies field communications, sensing, video capture, and edge analytics into a reliable and adaptive communication framework. At the field level, IP110H and IP503H Lite IP radios deliver instant voice communication and group coordination for frontline personnel. Sensors and cameras are connected through the AG260 AIoT Gateway, while the AS710 Wi-Fi 7 indoor access point provides high-quality Wi-Fi coverage in stations, platforms, and passenger areas to enhance traveler connectivity. Long-range, time-synchronized backhaul is supported by the SP250 Series, ensuring stable, high-throughput wireless links across trackside and outdoor environments. At the core, the zMEC edge-computing platform performs real-time AI video analytics, abnormal-event detection, and rapid data distribution to minimize latency and support mission-critical responses. By integrating Wi-Fi, LTE, 5G, PTP wireless backhaul, and AIoT devices into a single cohesive system, EdgeLink-X enables applications such as slope monitoring, intersection intrusion detection, pantograph spark/wear monitoring, and onboard operational visibility—forming a robust foundation for next-generation safe and intelligent mobility.

This solution, named "Mission Critical Network," is designed to provide enterprises, rescue teams, defense and security organizations, and more with intelligent applications utilizing 5G private networks, mobile satellite communication, drones, and other technologies. It addresses challenges such as remote inspections, equipment patrols, major disaster relief, and national security needs. With its rapid and efficient deployment capabilities and high mobility, the solution is capable of offering a stable, high-speed, and secure private network in situations where communication stability is severely compromised due to difficulties in reaching locations, numerous devices, extreme disasters, or tense regional conditions. This ensures the secure and reliable transmission of critical information.

The innovative parking operations solution integrates AI-based Automatic Number Plate Recognition (ANPR), sensing devices, IoT infrastructure, and a cloud management platform. By streamlining parking management, reducing manual intervention, and providing actionable insights through real-time data, this solution helps parking operators maximize space utilization and enhance customer satisfaction with a transparent, data-driven, brilliant parking service. The system provides real-time visibility into vehicle entry and exit, space occupancy, and on-site traffic flow. It supports ticketless access and license plate–based payment. It also offers real-time parking guidance and multiple payment options. These features reduce labour costs and enhance the user parking experience. A centralized management platform and operational data analytics give operators insights into traffic patterns, space turnover, and revenue performance. This enables informed decisions on pricing strategies and space allocation. As a result, operators can achieve efficient, sustainable smart parking operations.

"EUP Fleet Management System" can easily manage the core needs of vehicle dynamics, task progress, emergency response, and utilize AI to analyze driving behavior, fuel consumption, and effectively control costs. In addition, EUP provides comprehensive safety driving checks, tracks driver driving habits and physiological monitoring, significantly improving driving safety, and assisting companies in building a safe, efficient, and low-carbon transportation fleet.

Prof. Ming-Yen Lin’s Team – A Prediction Model for Type 2 Diabetes Complications Our research group develops a decision-support prediction model for type 2 diabetes complications using data from over 163,000 Taiwanese patients. Integrating age, HbA1c, and other dynamic factors, the model estimates long-term risks of major complications and mortality, visualizes future trends, and supports personalized care planning to improve disease management and outcomes.

Dr. Hung-Lung Lin’s Team – An Auxiliary Assessment System for Coronary Heart Disease Identification In response to population aging and the younger onset of coronary heart disease (CHD), this study develops an AI-based auxiliary CHD screening system integrated into a mobile app. Grounded in traditional Chinese medicine tongue–heart theory, the system analyzes sublingual vascular images using reflection removal, image segmentation, and a multi-head attention deep learning model. Users can obtain accurate, reproducible screening results and health recommendations within 7–20 seconds using only a smartphone, enabling convenient and cost-effective preliminary CHD assessment.

Prof. Ming-Chung Chou’s Team – Intelligent X-ray Imaging System Traditional X-ray imaging depends on experienced radiographers to balance image quality and radiation dose. Automatic exposure control may misjudge patients with poor positioning or metallic implants, increasing radiation by about 20%. This study develops an AI-based X-ray imaging system with non-contact body thickness measurement. By integrating patient height and weight, the system automatically determines optimal exposure parameters, ensuring stable image quality while reducing unnecessary radiation and cancer risk.

-)Mainframe monitoring system, maximally supporting one million spots in real time. -)Smart loading balanbe for AOS -)Remote backup for AOS -)Adjudge amount of AOS as required, based on flexible structure. -)Central security management.

Prof. Chia-Hsin Chen’s Team – An Innovative AI-Powered Tongue Vibration Training System This project develops an AI-driven tongue vibration training system for dysphagia and obstructive sleep apnea. It combines a safe, adjustable tongue device, AI motion tracking with real-time biofeedback, EEG-based cognitive monitoring, and AR-based gamified training to deliver personalized, engaging, and data-driven rehabilitation.

airaTrack is a comprehensive personnel tracking system that utilizes AI algorithms developed by our team to extract features from both the "face and figure." It generates unique biometric IDs with a recognition accuracy of up to 99%, all without the need for registering facial data, addressing privacy concerns. Capable of cross-camera searches and matching 10,000 faces per second, it significantly enhances the efficiency of traditional manual matching processes and increases search efficiency and accuracy by 97% enabling automatic path tracking of personnel in the field. The system also generates investigation reports for evidential preservation.

NETCenter is an integration and management application, which provides various management functions, like: fault, configuration, accounting, performance and security. All connections and network equipment are under control, you won’t disconnect any network connection, then figure out the problem. Our innovative features are: 1. No need to install a front-end application. No add-ons are need. Remotely manage via Web Browser. 2. Provides Auto-Discover and Topology and automatically generates an L2 Topology. 3. Supported Protocols: SNMP v2C/v3, Telnet, FTP, SMTP and ICMP. NETCenter manages all equipment that supports CLI (Command Line Interface). 4. Provides professional customized services to satisfy enterprise IT infrastructure management. Smooth and stable networks are affected by the rate of information exchange, also it’s an important symbol of an enterprise. Almost all enterprises use multiple vendor's equipment. In conclusion, NETCenter can reduce TCO and increase ROI.

To meet the high-speed operational demands of modern businesses, we offer stable dedicated enterprise circuits. Enterprise customers can flexibly choose transmission bandwidth according to their actual needs, with specifications ranging from 100M to 1G, and supporting ultra-fast transmission speeds up to 10G, ensuring uninterrupted flow of business information.