According to data from the "2021 Machine Tool Industry White Paper", Taiwan is the fifth largest exporter of machine tools in the world, which the industry plays an important role on the development of Taiwan's manufacturing industry, and has shown a market value growth trend from 2020 to 2024. In this background, Taiwan needs to actively develop next-generation smart manufacturing technologies to maintain market competitiveness. However, as AI technology advances rapidly, preventing Taiwan from falling behind other advanced countries is the main focus in this project. Our research directions are shown as follows:

1. Develop advanced control technology: Through intelligent compensation systems, server and interpolation parameters, curve smoothing and other technologies, in order to improve processing the accuracy and efficiency of industries such as machine tools, electrical discharge machining, and mold stamping.

2. AI embedded CNC controller: Combining AI technology and controller design to improve the performance of domestic controllers.

3. Online machine learning technology: Develop machine learning technology suitable for smart manufacturing plants, such as online automatic feature labeling, transfer learning, continuous learning, etc.

4. High-speed network applications: Apply 5G and Wi-Fi 6 technologies to build communication systems for sensors, controllers and edge computing platforms.

5. Measurement and testing modules: Develop common measurement and testing modules to improve the accuracy of semiconductor and electronic process equipment.

In addition, the plan also pays special attention to energy-saving and carbon-reducing technologies, such as the development of low-energy-consuming cutting processing technology, to achieve environmental protection goals.

Until August 2024, the project team has demonstrated the research results, where the participants in this exhibition, the teams of Professor Chao-Ching Ho (National Taiwan University of Science and Technology), Professor Meng-Hsiun Tsai (National Taiwan University), will be honored to demonstrate their inventions. Prof. Ho focuses on semiconductor critical dimension measurement, using spatial filters placed on the pupil surface to effectively screen the reflection zero-order spectrum, further improving the system measurement accuracy, with a milestone of correlation coefficient (>0.95), and critical dimensions being all less than 1.6%. Prof. Tsai collaborates with the industry to develop embedded AI CNC controllers, focusing on human-machine interfaces, model management systems, ECB edge computing modules and NPU inference acceleration. This project currently integrates thermal displacement prediction technology for AI inference, and also through using the U-Net model to analyze blade wear, the amount of calculation can be reduced by up to 96.5%.

The two teams look forward to exchanging experience and at this exhibition, creating more opportunities for interdisciplinary cooperation, transforming academics into value for local industries, accelerating industrial upgrading, and promoting Taiwan’s smart manufacturing industry.