The world＇s first AI-embedded CNC controller was developed, enabling AI model inference on both CPU and NPU. Model pruning (compression) was applied for efficient deployment, and data can be uploaded to the cloud for model refinement. The AI technologies include axial thermal deformation prediction, tool wear image detection, and interpolation and servo parameter optimization, aiming for high speed, precision, quality, and low cost.

This technology features four major innovations: 1. Hardware/software integration with Plug-In, NPU, sensor data capture, model management, and cloud modules. 2. Axial thermal deformation prediction with a model failure mechanism and transfer learning for model refinement. 3. Tool wear detection using a hybrid of continual and semi-supervised learning. 4. Interpolation and servo parameter optimization to minimize machining time within fixed contour error and jerk limits.

In collaboration with a domestic CNC controller manufacturer, this technology led to the development of the world’s first AI-embedded CNC controller. It integrates motion control and AI inference on a single chip, reducing costs while enabling real-time compensation. Three key AI functions greatly enhance high-speed, high-precision, and high-quality machining. Over a thousand units have been shipped, marking a significant contribution to the industry.