For Manufacturing (RosettaBox): Delivers "Superhuman Efficiency." By processing real-time edge data, it detects anomalies early, empowering managers to optimize production decisions without disrupting workflows.

Remote expert video collaboration solution uses AR smart glasses to transmit onsite situation to the remote technical experts with high-definition real-time video, and guide the on-site engineer to troubleshoot and repair the fault from the first-person view perspective and free their hands. This is not only improves maintenance efficiency, but also ensures safety. It also achieve the sharing and inheritance of expert resources and high cost of traveling on site. When encountering difficult situations, multiple technical experts can also be invited for consultation. Experts can provide real-time multi-party HD video guidance with the on-site engineer, whether on PC or smartphone, to jointly solve technical issues and reduce downtime losses.

Under the theme “Economy Next,” the Ministry of Economic Affairs presents dual pavilions on AI and net zero, illustrating Taiwan’s vision for dual transformation. The Net Zero Pavilion showcases resilient green energy systems and low-carbon supply chains, while the Industries AI Pavilion highlights advanced technologies and smart applications that delivers concrete solutions across industries in hope of fostering collaboration opportunities with global partners.

The technical solution of Foxtron Electric Bus combines domestic and international resources, each contributing its strengths： 1. It adopts Foxtron's self-developed VCU software, integrates and develops the EEA architecture for electric buses, and utilizes motors, batteries, and chassis suspension systems from international major manufacturers. In collaboration with domestic manufacturers, Foxtron designs and develops dedicated low-floor bus frames for chassis cooperation. 2. In the product design of electric buses, Foxtron identifies pain points in traditional buses and improves upon them. For example, traditional buses often have issues with rollover protection and numerous blind spots, as well as differential wheel rotation caused by wheelbase, leading to frequent traffic accidents due to visibility problems. To address these issues, Foxtron employs Computer-Aided Engineering (CAE) and NVH (Noise, Vibration, and Harshness) analysis to verify the design of various systems and components. This ensures the overall strength of the vehicle body and chassis, enhances passenger safety, and incorporates features such as panoramic image display and an electronic wide-angle rearview mirror with a blind spot detection system to solve driver blind spot issues. Additionally, to alert pedestrians and other road users to the risk of differential wheel rotation, Model T projects prominent light beams onto the ground during turns, marking the danger zone.

"Hualien AI Agent: 24/7 Digital Travel Hub" is a smart solution specifically designed to enhance urban resilience and boost traveler confidence. Built on the LangGraph architecture and integrated with multiple Large Language Models (LLMs), the system utilizes Natural Language Processing (NLP) to precisely understand user needs. It supports over 20 languages, providing authoritative information ranging from local culture to real-time post-disaster traffic updates and safe routing. As the first AI Agent in Taiwan capable of proactively calling multi-source real-time APIs across various government departments, it combines LBS (Location-Based Services) with a professionally maintained RAG authoritative knowledge base. This allows the system to push precise safety alerts and personalized itinerary recommendations to users. Through deep OMO (Online-Merge-Offline) integration, the service extends from digital platforms to major transportation hubs, visitor centers, and local businesses. It not only demonstrates the resilience of digital governance but also fully empowers the tourism industry to accelerate its recovery.

The Taichung Local Tax Bureau has launched the "Local Tax AI Anywhere Door," a 24/7 intelligent service platform powered by AI and Natural Language Processing (NLP). By integrating local, national, and cross-agency data, the system utilizes Deep Learning (BERT) and a specialized tax knowledge base to provide precise, multi-turn conversational consultations. The platform also features online applications, tax simulations, e-payments, and live text/video support, drastically reducing processing time and costs for citizens. To date, it has served over 440,000 cases with 99% accuracy and over 90% satisfaction, effectively lowering transportation-related carbon emissions and demonstrating the power of human-centric digital governance.

With 5G private network system integration provided by Wave-In Communication, Inventec is transforming its Taoyuan Factory into a 5G smart factory, creating highly intelligent and automated production lines with AI and IoT capabilities.

This project integrates cable TV systems, Civil IoT, and AI video recognition technology to build smart disaster prevention and site safety services. Centered on set-top boxes (STB) and intelligent sensing, the solution focuses on bridging the gap in real-time information and safety protection for elderly and rural populations. Through its wired network infrastructure, the system developed under this project is capable of delivering real-time disaster prevention information, including earthquakes, typhoons, and heavy rainfall alerts. In addition, AI-based recognition capabilities have been integrated into the system. Currently, the system has completed event recognition functions such as parcel placement and retrieval detection. It also performs analysis of human movement trajectories and abnormal group behavior detection. In the future, the system will be further expanded to identify dangerous situations such as violent behavior. Through this project, Taiwan Optical Platform (TOP) demonstrates the potential of integrating cable television networks with AI and IoT technologies to develop a platform for smart disaster prevention and public safety applications.

Live crowd movement information is crucial in busy public spaces. Manual checks and basic sensors often lack accuracy and fail to keep up with real-time changes. The People Counting System uses AI video analytics and multi-sensor fusion. It delivers accurate, real-time crowd data. The system automatically counts people, tracks direction, monitors density, and provides instant alerts and reports. Deep learning ensures accuracy in low-light or crowded conditions. Multi-target tracking prevents double-counting. Edge computing and cloud-based management enable on-site analysis and long-term monitoring. Integrating cameras, infrared, Wi-Fi, and Bluetooth data improves detection and increases coverage. The system generates real-time heat maps, flow analyses, and historical records. These support layout optimization, entry regulation, evacuation planning, and efficiency. The system serves malls, transit hubs, campuses, parks, and public buildings. It is vital for smart cities. It replaces manual counting with intelligent management for better safety, responsiveness, and adaptability.

In the era of automation and digital transformation, the challenge for managers is no longer just hardware installation, but how to integrate fragmented subsystems into a connected network with a “brain.” The SOYAL WEB PLC (AR-401-PLC-1616R) was created precisely for this purpose as a cross-domain integration core. It breaks through the limitations of traditional PLCs, with a built-in web management interface and powerful dual communication channels (CH1 / CH2), enabling simultaneous coverage of both “identity recognition” and “physical facility monitoring.” With full support for the Modbus TCP protocol, on-site data can seamlessly connect to cloud platforms or SCADA central monitoring systems, making remote control and data analysis easily accessible.

Integrated production management system covering 4M1E. Features real-time data collection, digital Kanban visualization, and full traceability. Enables paperless operations and complete process transparency. 1. Real-time Monitoring: Connects equipment signals to display WIP, utilization rates, and yield data instantly. 2. Traceability: Records comprehensive production lot numbers and process parameters for quality auditing. 3. Smart Scheduling: Optimizes production plans based on capacity constraints and order requirements.

To enhance facility management efficiency and visitor service quality in Nantou County’s scenic areas, the Nantou County Scenic Area Administration Office is implementing a Smart Management System. The system applies digital technologies to strengthen operations, safety management, and data-driven governance. It integrates maintenance and inspection records for key facilities such as trails and suspension bridges, establishing a centralized database. Through data analysis, authorities can monitor facility conditions, identify maintenance needs, and implement preventive maintenance strategies, reducing costs and improving operational stability. CCTV and smart sensors are gradually deployed in high-risk areas to enable real-time monitoring and anomaly alerts. During extreme events such as typhoons or earthquakes, managers can quickly assess infrastructure safety, improve emergency response efficiency, and reduce manual inspection workload. A QR Code-based smart reporting function allows visitors to instantly report facility issues. Reports are automatically recorded and assigned for follow-up, accelerating response time and enhancing visitor safety and service experience. Through smart management and data-driven decision-making, Nantou County is building a safer, more efficient, and sustainable scenic area management model, supporting smart tourism and long-term development.