top of page

/ 2024 Achievements /

人才培育
Talent Cultivation

本中心延攬專職碩士級與博士級研究人員,共有16位高階研發人才加入團隊與去年相比成長33%,其中包含5位博士後研究員,以及具有豐富產業經歷、累積40年資歷的高階資深產業經理人,培育具國際研發經驗之年輕學者或博士生數共145名,遠超過達成值。並鼓勵學生或博士後參與國內外研討會或相關課程。下列為研究生獲獎紀錄:

The center has recruited full-time master's and doctoral level researchers. A total of 16 senior R&D talents have joined the team, a 33% increase compared to last year, including 5 postdoctoral researchers and senior senior researchers with rich industry experience and accumulated 40 years of qualifications. Industry managers, and cultivated 145 young scholars or doctoral students with international R&D experience, far exceeding the target. Students or postdoctoral fellows are encouraged to participate in domestic and international seminars or related courses. The following are the graduate student award records: 

  • 曾泰明、黃庭緯、曾國賢、王昱文指導老師李維楨老師,獲「第十六屆上銀智慧機器手實作競賽應用組智慧裝疊季軍」。

  • ​Tseng Tai-Ming, Huang Ting-Wei, Tseng Kuo-Hsien, and Wang Yu-Wen, under the guidance of Professor Li Wei-Chen, won 3rd Place in the Application Group of the 16th HIWIN Intelligent Robot Competition.

  • 邱柏誠指導教授李維楨老師,獲「the Best Student Paper Award - 3rd Place, the 21st International Conference on Automation Technology (Automation 2024) 」。

  • ​Chiu Po-Cheng, under the guidance of Professor Li Wei-Chen, won the Best Student Paper Award - 3rd Place at the 21st International Conference on Automation Technology (Automation 2024).

  • 張皓鈞、蔡承佑、吳秉桓指導教授陳羽薰老師,獲「113年台灣精密工程專題與論文獎-大專專題佳作」。

  • ​Chang Hao-Chun, Tsai Cheng-You, and Wu Ping-Huan, under the guidance of Professor Chen Yu-Hsun, received Honorable Mention in Undergraduate Projects in the 2024 TSPE Research Project and Paper Competition.

  • 張鎬宇、蔡承恩、指導教授陳羽薰老師,獲「113年中華民國機構與機器原理學會-ADAMS論文獎第二名」。

  • ​Chang Hao-Yu and Tsai Cheng-En, under the guidance of Professor Chen Yu-Hsun, won 2nd Place in the ADAMS Paper Award from the 2024 The Chinese Society of Mechanism and Machine Theory (CSMMT).

  • 毛姵媞指導教授陳羽薰老師,獲「113年中華民國機構與機器原理學會-優等碩士論文獎」。

  • Mao Pei-Ti, under the guidance of Professor Chen Yu-Hsun, received the Outstanding Master's Thesis Award from the 2024 The Chinese Society of Mechanism and Machine Theory (CSMMT).

  • 王芯苡指導教授陳羽薰老師,獲「6th IFToMM Symposium on Mechanism Design for Robotics (MEDER 2024)- the Gold Best Student Paper Award」。

  • Wang Hsin-Yi, under the guidance of Professor Chen Yu-Hsun, won the Gold Best Student Paper Award at the 6th IFToMM Symposium on Mechanism Design for Robotics (MEDER 2024).

  • 指導教授林柏廷老師,獲「2024 Automation Best Student Honorable Award」。

  • Professor Lin Po-Ting received the 2024 Automation Best Student Honorable Award.

  • 指導教授林柏廷老師,獲「中華民國機構與機器原理學會-優秀年輕機構與機器原理學者獎」。

  • Professor Lin Po-Ting received the Outstanding Young Scholar Award in Mechanism and Machine Theory from The Chinese Society of Mechanism and Machine Theory (CSMMT).

  • 2024執行財團法人研華文教基金會全國競賽林吳叡、Sofia Haydee Borja Alfonso、Fransisca Cindy Mulyani、Josua Jeffrey Handopo指導教授周碩彥老師,獲「nZEB Energy Management System」獲第二名。

  • Lin Wu-Rui, Sofia Haydee Borja Alfonso, Fransisca Cindy Mulyani, and Josua Jeffrey Handopo, under the guidance of Professor Chou Shuo-Yan, won second place in the Advantech Education Foundation’s National Competition for their project, “nZEB Energy Management System.”

  • 2024年全國技專校院學生實務專題製作競賽暨成果展,指導教授姚智原老師,第一名為國立臺灣科技大學「資工通訊群」的「AI走入生活:建模技術的革新」,該研究作品整合人工智慧和圖學領域的先進技術,以提升建模技術,且該系統生成模型品質高,可廣泛應用於遊戲、動畫、AR、VR等,兼具創新性與商品化獲獎。

  • At the 2024 National Technical and Vocational Colleges and Universities Student Practical Project Production Competition and Exhibition, under the guidance of Professor Yao Chih-Yuan, the National Taiwan University of Science and Technology "Computer Science and Communications Group" won first place with their project, “AI in Daily Life: A Revolution in Modeling Technology.” This project integrates advanced AI and graphics techniques to enhance modeling technology, producing high-quality generative models widely applicable to games, animations, AR, VR, and more, demonstrating both innovation and commercialization potential.

  • 2024台灣鍍膜科技協會指導教授王丞浩老師,獲「口頭論文競賽」佳作。

  • At the 2024 Taiwan Association for Coating and Thin Film Technology (TACT), under the guidance of Professor Wang Chen-Hao, won Honorable Mention in the Oral Paper Competition.

  • 2024台灣鍍膜科技協會指導教授王丞浩老師,獲「海報論文競賽」特優。

  • At the 2024 Taiwan Association for Coating and Thin Film Technology (TACT), under the guidance of Professor Wang Chen-Hao, won the Excellence Award in the Poster Paper Competition.

  • 2024第十九屆全國氫能與燃料電池學術研討會,指導教授王丞浩老師,獲「海報論文競賽」第一名與特等獎。​

  • ​At the 2024 National Conference on Hydrogen Energy and Fuel Cells, under the guidance of Professor Wang Chen-Hao, won first place and the Special Excellence Award in the Poster Paper Competition.

 光機電技術研發中心研究成果 
Research Achievements of Opto-Mechatronics Technology Center

六軸機械手臂輔助中大型工件智慧化表面拋光系統

Six-Axis Robotic Arm Assisted Intelligent Surface Polishing System for Medium and Large Workpieces

  • 光機電中心113年研製六軸機械手臂輔助中大型工件表面拋光系統,可針對不同外形曲率中大型工件STAVAX不銹模具鋼智慧化選用拋光柱或拋光球,表面進行工件表面定力拋光,藉此改善工件之表面粗糙度至Ra 20奈米。德國Fraunhofer Institute for Production Technology, IPT 也在研發此技術應用於汽車模具。

  • ​A six-axis robotic arm-assisted surface polishing system was developed to intelligently select polishing rods or balls for force-controlled polishing of medium to large workpieces, such as STAVAX stainless mold steel, improving surface roughness to Ra 20 nanometers. The Fraunhofer Institute for Production Technology (IPT) in Germany is also applying this technology to automotive molds.

 智慧型機器人研究中心研究成果 
Research Achievements of Center for Intelligent Robotics

六軸機械手臂整合研究

Six-Axis Robotic Arm Integration Research

  • 顏家鈺校長使用達明六軸協作型機械手臂進行腫瘤消融手術的引導,透過醫學影像及簡易的人機介面,能準確的定位腫瘤,縮短醫生手術的時間,並提高手術準確性。使用機械手臂第六軸上的攝像機,進行影像的定位,搭配landmark定位片,可得知影像空間中的座標系。開發MATLAB APP設計一套簡易的人機介面,讓操作者能容易且快速的上手,縮短手術前的準備時間,透過系統的整合,使此項技術能有效的提升手術的準確度。本年度目標為降低前一年度影像到機械手臂座標系間的誤差,縮短了醫生在對準肺臟腫瘤上的時間,並提供簡易的介面,供操作者使用。

  • Principal Jia-Yush Yen uses the Daman six-axis collaborative robot arm for guiding tumor ablation surgery. Through medical imaging and a simple human-machine interface, it accurately locates the tumor, shortens the surgeon's operation time, and improves surgical accuracy. A camera on the sixth axis of the robotic arm is used for image localization, combined with landmark positioning plates to determine the coordinates in the image space. A MATLAB app was developed to design a simple human-machine interface, allowing operators to quickly and easily get started, reducing the preparation time before surgery. Through system integration, this technology effectively enhances the accuracy of the surgery. The goal for this year is to reduce the error between the previous year's image and the robotic arm's coordinate system, shorten the time spent by the surgeon in aligning the lung tumor, and provide an easy-to-use interface for the operator.

  • 陳羽薰教授提出一應用傘齒輪與多連桿之重力補償機構,此機構透過彈簧在機構運動過程中利用彈力位能補償負載之重力位能的變化以達成靜平衡狀態,使機構在運動過程中減少能量消耗,進而提高機機構的有效負載。透過電腦輔助設計模擬軟體Inventor進行本機構靜力平衡之驗證。本研究後續將透過原型機實作與測試,驗證此外裝式機構的負載能力。

  • Professor Yu-Hsun Chen proposed a gravity compensation mechanism using bevel gears and a multi-bar linkage. This mechanism compensates for changes in the gravitational potential energy of the load during motion by utilizing the elastic potential energy from a spring, achieving a static equilibrium state. This reduces energy consumption during motion and improves the effective load capacity of the mechanism. The static equilibrium of the mechanism was verified using the computer-aided design simulation software Inventor. The subsequent phase of the research will involve prototype implementation and testing to validate the load capacity of the external-mounted mechanism.

​數位微影定義 PCB 線路圖案自動化定位

Digital Lithography-Based Automated PCB Circuit Pattern Alignment

  • 林其禹教授成功使用接觸式量測儀對機器手臂的絕對誤差進行校正,並通過實驗進行來驗證機器手臂在局部空間移動的準確度。通過測量機器手臂在不同位置上的誤差,並使用內插法計算每移動1毫米所產生的誤差,我們得以對控制器進行補償,可使機器手臂的精度顯著提高。由於本計畫將機器手臂的絕對誤差視為空間數據,因此誤差建模和補償不依賴於任何類型的運動學模型,此方法具有較好的通用性。此外,由於誤差補償不需要修改機器人控制系統,因此所提方法可用於對具有封閉控制系統的機器人進行校正。實驗結果顯示,機器手臂經過校正後在空間中的誤差從5.468mm降低至0.659mm共降低了約88%,此方法能夠有效補償機器人的絕對誤差。所提方法具有良好的應用價值,適合用於機器人方向變化較小的應用,如飛機機翼、機器人鑽孔等。

  • Professor Chyi-Yeu Lin successfully used a contact measurement device to calibrate the absolute error of a robotic arm and verified its accuracy in local space movement through experiments. By measuring the errors at different positions of the robotic arm and using interpolation to calculate the error generated for each millimeter of movement, we were able to compensate for the controller, significantly improving the arm's precision. Since the absolute error of the robotic arm is considered as spatial data, error modeling and compensation do not rely on any type of kinematic model, making this method highly versatile. Additionally, since error compensation does not require modifications to the robotic control system, this approach can be used to calibrate robots with closed-loop control systems. Experimental results show that after calibration, the error in the robotic arm's movement decreased from 5.468 mm to 0.659 mm, a reduction of approximately 88%. This method effectively compensates for the robot's absolute error. The proposed method has good application value and is suitable for applications where the robot's directional change is minimal, such as in aircraft wings and robotic drilling..

​攀爬機器人性能優化與機器視覺應用

Performance Optimization and Machine Vision Applications for Climbing Robots

  • 李維楨教授於113年度成功開發出一台適用於直徑200 mm鋼纜、可跨越5 mm凸檻、攀爬距離可達100 m、續航時間達4小時並具備緩降機制、且能以1080p解析度記錄完整鋼纜表面影像的爬纜機器人,並實際於大直橋上測試成功。機器人在無動力狀態下滑,原先速度為71 mm/s,安裝旋轉阻尼器後速度變為17 mm/s,減少76%的落下速度。根據力學分析,將從動輪由原先的3吋輪置換為4吋輪,減少機器人上推的阻力,進而提升機器人跨檻的穩定度。

  • Professor Wei-chen Lee successfully developed a cable-climbing robot designed for 200 mm diameter steel cables, capable of overcoming a 5 mm bump, climbing up to 100 meters, with a 4-hour battery life and a descent mechanism. The robot can also record complete surface images of the steel cable at a 1080p resolution and has been successfully tested on the Dazhi Bridge. When the robot slides without power, its original speed was 71 mm/s, but after installing a rotary damper, the speed was reduced to 17 mm/s, resulting in a 76% reduction in descent speed. Based on mechanical analysis, the driving wheel was replaced from a 3-inch wheel to a 4-inch wheel, reducing the resistance to upward movement and improving the robot's stability when crossing bumps.

高負載自主移動機器人

High PayLoad Autonomous Mobile Robot (HAMR)

  • 林柏廷教授發展高負載自主移動機器人的研究,以ROS做為平台,除了整合多種異質感測器(包含慣性量測單元、光達、2D/3D相機、電子皮膚等),還導入避障路徑規劃技術,發展2D/3D人員及物件辨識技術,在數位空間計算機器人運動路徑中是否會與周遭人員及物件碰撞,若可能產生碰撞,則啟動最佳化方法計算安全避障路徑,目前已經能在200 ms內完成計算,達成機器人動態避障路徑規劃的目標。

  • Professor Po Ting Lin has developed research on high-load autonomous mobile robots using ROS as the platform. In addition to integrating various heterogeneous sensors (including inertial measurement units, LiDAR, 2D/3D cameras, electronic skin, etc.), the system also incorporates obstacle avoidance path planning technology and develops 2D/3D human and object recognition techniques. By computing the robot's motion path in a digital space, the system checks whether it may collide with surrounding people or objects. If a potential collision is detected, an optimization method is activated to calculate a safe avoidance path. The system is currently able to complete the calculations within 200 ms, achieving the goal of dynamic obstacle avoidance path planning for the robot.

image.png

此圖為使用DIC進行高負載自主移動機器人位移量的實驗架設

This figure shows the experimental setup for measuring displacement using DIC in the high payload autonomous mobile robot.

基於數位憑證的資安防護機制

Cybersecurity Protection Mechanism Based on Digital Certificates

  • 林柏廷教授提出一『基於工業控制系統提出的網絡安全模型』,基於『數位憑證』的資安防護機制,確實能夠 止系統數據被篡改。無數位憑證保護的情況下,地址解析協議攻擊後的通信延遲顯著增加,平均延遲從正常的6.92ms增至26.93ms,同時丟包率從0%增加到13%。這些數據充分說明了地址解析協議攻擊對網絡性能和穩定性帶來的負面影響。而在有數位憑證保護的情況下,結合公鑰基礎設施和信任鏈技術的多層次安全架構,能夠有效提升自主移動機器人系統的數據傳輸安全性和身份驗證可靠性。

  • Professor Po Ting Lin proposed a "Network Security Model Based on Industrial Control Systems," which utilizes a "digital certificate" security protection mechanism to effectively prevent the tampering of system data. Without digital certificate protection, communication delay significantly increases after a DNS (Domain Name System) attack, with the average delay rising from the normal 6.92 ms to 26.93 ms, while the packet loss rate increases from 0% to 13%. These data clearly demonstrate the negative impact of DNS attacks on network performance and stability. However, with digital certificate protection, the combination of public key infrastructure and trust chain technology in a multi-layered security architecture effectively enhances the data transmission security and authentication reliability of autonomous mobile robot systems.

​電動車即時車況監測及故障診斷

Real-Time Vehicle Condition Monitoring and Fault Diagnosis for Electric Vehicles

  • 劉孟昆教授將原先電動車之空架構上,進行相關實驗設備之設計與安裝,以達成後續須上路擷取資料之實驗要求,其範圍包括:方向盤與轉向機、動力系統、剎車系統、電力系統、擷取系統及內裝等。為使車輛達得以快速開發、技術驗證以及本地運算目的,使用了Labview以進行車輛控制以及資料擷取;並以Python來進行資料處理、訓練及分析,判別目前車輛是否故障。在機器學習方面,本研究使用隨機森林 (Random Forest, RF) 、支持向量機 (Support Vector Machine, SVM) 及一維卷積神經網路 (One-dimensional Convolutional Neural Network, 1D-CNN) 來驗證間接輪胎壓力監測系統,以識別車輛輪胎的輪胎壓力是否不足。另外亦使用相同之感測器進行馬達高阻連接之故障監診,在進行超參數最佳化並訓練分類器模型後,比較了三種分類器在感測器之篩選前與後之準確率。

  • Professor Meng-Kun Liu designed and installed experimental equipment on the original electric vehicle chassis to meet the requirements for subsequent road data collection experiments. The scope of the work includes the steering wheel and steering mechanism, power system, braking system, electrical system, data acquisition system, and interior components. To facilitate rapid vehicle development, technology validation, and local computation, Labview was used for vehicle control and data acquisition, while Python was employed for data processing, training, and analysis to determine if the vehicle has any faults. In terms of machine learning, this research uses Random Forest (RF), Support Vector Machine (SVM), and One-dimensional Convolutional Neural Network (1D-CNN) to validate the indirect tire pressure monitoring system and identify whether the vehicle's tire pressure is insufficient. Additionally, the same sensors were used for monitoring motor high-resistance connections. After performing hyperparameter optimization and training the classifier model, the accuracy of the three classifiers was compared before and after sensor selection.

Prof. Shun-Feng Su

Director of Intelligent Manufacturing Innovation Center

蘇順豐 講座教授 (智慧製造創新中心主任)

Chen-Pei Chien 

Full-Time Assistant

簡貞佩 專任助理

智慧製造創新中心位置(工業4.0中心)

​(106)台北市大安區羅斯福路四段113巷25號

臺灣科技大學公館校區

No. 25, Ln. 113, Sec. 4, Roosevelt Rd., Da’an Dist, Taipei City 106, Taiwan (R.O.C.)

Tel: +886-227333141 #5207

/ Contact /

© 2023 by Intelligent Manufacturing Innovation Center  All Rights Reserved.

bottom of page