The 5G private network technology, characterized by high-speed transmission and low latency, effectively enhances communication stability and supports business operations. However, due to the susceptibility of 5G millimeter-wave signals to obstruction, small cell base stations are required to strengthen coverage.
This project aims to verify the feasibility of 5G private networks for remote collaboration through industry–academia cooperation, while comparing the coverage, transmission speed, and stability of 5G and Wi-Fi networks within a campus setting. The study also evaluates their performance at various distances and in CloudXR streaming applications.
During the verification process, the 5G private network employs indoor-type Radio Units (RUs) with transmission power comparable to household Wi-Fi. A 5GHz wireless network card on a laptop is used to simulate differences between outdoor RUs and Wi-Fi access points, assessing the impact on transmission distance.
In the next phase, the project will focus on facial image reconstruction technology in Extended Reality (XR) applications. By restoring facial visuals, users wearing XR headsets can maintain social connection with the real world, reducing isolation and enhancing the realism of immersive experiences. This research further explores the potential of 5G technology in mixed reality (MR) applications.

ABC-ATEC has collaborated with National Taipei University in an industry–academia project to develop the “Metaverse AI Multimodal Cross-Language Task-Oriented Dialogue System.” This system integrates artificial intelligence (AI), natural language processing (NLP), and deep learning neural network models to create a multilingual and multimodal dialogue platform. Initially applied in the healthcare sector, the system is being expanded to enterprise and educational applications, with the ultimate goal of deployment in manufacturing sites in Malaysia—aiming to reduce labor demands and enhance operational efficiency.
The project underwent extensive discussions—from conceptual planning to real-world implementation—addressing key issues such as multilingual communication requirements, humanized voice interaction, data source expansion, and large language model selection. Given Malaysia’s multilingual workforce, the system has been specifically optimized for cross-language communication in production environments, helping employees overcome language barriers, reduce miscommunication, and streamline manufacturing processes.

▲One Tree, One Mountain: Plant a Tree, Walk Home