This product mainly provides intelligent assistance for the "walking" part of the visually impaired. The assistance can be divided into two parts. The first part uses the camera on the intelligent white cane combined with the obstacle (including motorcycles, bicycles, sundries, etc.) image recognition to solve the problems that the visually impaired may encounter on the sidewalk. when crossing the road, identify pedestrian traffic lights and pedestrian crossing and intelligently detect moving objects (cyclists, pedestrians, cats, dogs, etc.); the other part, mainly The intelligent navigation technology is equipped with a Chinese dictionary for the visually impaired and the vibration module of the intelligent white cane provides the visually impaired with accurate orientation guidance when they go to their destination. Ultimately, it is hoped that this product can help the visually impaired to "walking" independently in a world full of sunshine.

This project develops an innovative, integrated service to help fish farmers transition toward intelligent and sustainable pond management. The system’s core is a cascade filtration tank employing anaerobic ammonium-oxidizing (anammox) bacteria and maifan stone ceramic rings to convert harmful nitrogen compounds into harmless nitrogen gas. White shrimp are introduced to absorb excess nutrients, reducing sludge formation, while carbon-sequestering species such as Corbicula fluminea, selected crops, microalgae, and beneficial microbes enhance water quality and promote ecological synergy. An environmental monitoring module continuously tracks parameters including temperature, pH, and ammonia nitrogen to support data-driven management decisions. Finally, a “Green-Gold Digital Twin” integrates mathematical modeling and artificial intelligence to simulate and predict system performance, creating a comprehensive and intelligent framework for circular aquaculture management.

This study develops an AI-powered knowledge management platform to enhance knowledge sharing within the Ministry of Environment. By leveraging generative AI, keyword extraction, knowledge association networks, and topic-based ontology structures, the system transforms large volumes of reports, regulations, and case studies into accessible and verifiable knowledge. Focusing on soil and groundwater management platform enables users to obtain accurate, context-based insights for decision-making. Results show that integrating ontology networks with large language models improves response accuracy, minimizes irrelevant outputs, and provides source verification. Testing by government officials confirmed that all responses were based on report content, demonstrating the platform’s potential as a scientific tool for training, education, and strategic environmental planning.

The AI-powered Video-based Internal Educational Dialogue System is designed to enhance organizational learning through interactive and intelligent training experiences. By integrating real-time inquiry, process guidance, procedural explanations, and video playback control, the system enables employees to engage directly with training materials in a conversational manner. This approach fosters immersive and scenario-based learning, improving operational understanding, retention, and efficiency across internal education programs. The platform demonstrates how AI-driven dialogue technologies can transform traditional video training into a dynamic, adaptive knowledge environment.