The Intelligent Automation and Robotics Center of Tamkang University, in collaboration with Prefactor Tech, has been selected for the Ministry of Economic Affairs (MOEA) and NVIDIA's Taipei-1 OVX Computing Power Program (5th Call for Proposals). Leveraging the computational capabilities of the Taipei-1 AI supercomputer, the team will carry out the project, "Resilience Verification of AI-Driven AMR Dispatching Decisions in a Digital Twin Hospital Environment." Tamkang University is the only private university selected for an OVX computing project in this funding round, joining leading research teams from organizations such as the Industrial Technology Research Institute (ITRI) and National Tsing Hua University.
Vice President for Academic Affairs Hui-Huang Hsu expressed great confidence in the project's potential impact. He noted that Tamkang University's robotics research team has been developing robotics technologies for more than two decades. In the past, robots were constrained by technological limitations and could perform only relatively simple tasks. "With the emergence of large AI models, robots are now capable of making far more sophisticated decisions about their surroundings and can even communicate naturally with people through spoken language. This represents a significant leap forward for robotics." He thanked the Ministry of Economic Affairs for continuously promoting the Leading Enterprise R&D Deep Cultivation Program, which encourages industry–academia collaboration through the Taipei-1 computing initiative. He also expressed gratitude to NVIDIA for providing access to the Taipei-1 AI supercomputer, enabling the team to accelerate the development and validation of large language models and large-scale digital twin hospital environments. In addition, he acknowledged the assistance provided by the Industrial Technology Research Institute (ITRI) throughout the proposal preparation, project execution, and dissemination of project outcomes.
Project advisor Professor Ching-Chang Wong stated that the collaboration has given students valuable opportunities to utilize the computational power of the Taipei-1 supercomputer. He noted that the intelligent dispatching system developed through the project has enormous potential beyond hospital logistics, with future applications in industrial park delivery services, campus patrol systems, smart factories, semiconductor manufacturing facilities, and other autonomous logistics environments.
Using hospital logistics as its primary application scenario, the project establishes an intelligent dispatching simulation and validation platform that integrates digital twin technology, Autonomous Mobile Robots (AMRs), and Large Language Models (LLMs). Advanced Intelligent Technologies Inc. serves as the lead organization, responsible for constructing the hospital digital twin environment, developing Visual Components 3D models, defining AMR operational scenarios, and designing the data architecture for the intelligent dispatching system. The Tamkang University research team utilizes NVIDIA Isaac Sim to place multiple AMRs within a virtual hospital environment, where they can repeatedly simulate dispatching strategies and navigation routes. This virtual testing environment not only ensures operational safety but also substantially reduces development and testing costs.
The team employs Meta Llama 4, a state-of-the-art large language model, as the "brain" of the dispatching system. By understanding contextual information much like a human operator, the system can instantly generate optimal dispatching decisions based on factors such as task urgency, robot battery levels, and real-time corridor traffic conditions, thereby enabling more intelligent and resilient fleet management. A demonstration video of the project is available at: https://youtu.be/KVEhv4SsQfg
