拉麵大師機器人 RamenBot 2026/07/14

簡介 Introductions

工研院與光聚晶電攜手研發全臺首創拉麵大師機器人，運用 VLA (Vision-Language-Action) 模型與數位雙生技術，成功突破軟質配料（如溏心蛋、筍乾）形狀與位置不固定的夾取難題。僅透過視覺學習即可智慧判斷、客製化配置小料，並以職人級水準精緻擺盤，實現「職人技藝 × 標準化量產」的智慧餐飲新模式。

ITRI, in collaboration with Star Fusion Group, has developed Taiwan’s first AI RamenBot. Leveraging the VLA model and digital twin technology, the RamenBot overcomes the challenge of handling soft, irregularly shaped ingredients such as soft-boiled eggs and bamboo shoots. Using pure visual intelligence, it precisely customizes toppings and delivers elegant, chef-quality plating — seamlessly blending master craftsmanship with reliable standardization.

特色與創新 Characters and Innovations

• 全臺首創 VLA 智慧餐飲：機器手臂透過攝影機即時看見畫面，自主推理並執行精準動作。
• 雙核心學習技術：結合模仿學習與強化學習，即使配料角度與訓練時不同，仍能透過即時視覺靈活應對。
• 極簡機構設計：純視覺辨識與先進演算法完成配料作業，減少機械結構與廚房空間需求，提升部署彈性。

• Taiwan’s First VLA Smart Dining Solution: The robot instantly sees, reasons, and acts with precision.
• Dual-Core Learning (Imitation + Reinforcement): Adapts instantly to varying ingredient positions and angles.
• Minimalist Design: Pure vision and AI, no complex mechanisms, saving significant kitchen space.

應用與效益 Applications and Benefits

• 解決餐飲業缺工與品質不穩問題。
• 提升效率、降低營運成本。
• 維持產品一致性與職人級視覺美感。

• Solving labor shortages and quality inconsistency.
• Boosting efficiency and reducing costs.
• Maintaining superior consistency and visual appeal.