一般來說，銑削零件的製造流程相當複雜，從起始圖檔的繪製到生成加工路徑所需的刀具選擇、工序編排和策略選擇等設計，這些步驟都需要人工來完成規劃和安排。隨著電腦輔助製造軟體的普及，符合品質要求的參數組合並非唯一，這使得不熟悉相關技術知識的人員在面對這些複雜且無標準規範的路徑規劃問題時難以下手。即使是經驗豐富的師傅也未必能在短時間內找到最合適的解決方案，這種情況會對產業造成負面的影響，因此有研究提出了非傳統的製程規劃系統，結合電腦輔助製造(Computer Aided Manufacture, CAM)軟體的二次開發功能與機器學習(Machine Learning, ML)模型，試著要用卷積神經網路(Convolutional Neural Network, CNN)來從零件的影像中提取特徵，藉此進行加工工序、工法、策略與刀具尺寸選擇的參數預測，此系統已於測試中有了不錯的成果，但於模型預測的準確率仍然有進步空間。
本研究提出不同的資料蒐集處理以及不同模型架構的設計，以此來優化整體模型表現，降低模型的失誤率，藉此提升自動化製程規劃系統的可應用性。透過本研究的調整，相較於以往存在低於70%的個別標籤準確率的表現，新的資料集與模型架構有著最低85%的個別標籤準確率，平均準確率來到95%的優良表現。

In this study, a milling process planning system for 2.5D parts based on Convolutional Neural Network (CNN) was developed to infer the operation sequence, tool selection, and cutting parameters. A CAD file was sliced into layers as what is done for the 3D printing process planning. It is expected that the cross-sectional images of these layers can provide enough geometric information to the CNN model to conduct process planning automatically. With a proper design of the output code, the planning of operation sequence, tool size, and tool path pattern is generated integrally. The samples of optimal process planning results obtained by exhaustive method were used as the training data. We tested different structural connection for the CNN’s architecture, and used different feature extraction layer such as Residual Block from Resnet and Inception Block from GoogleNet to develop a CNN model that has the average accuracy of 95%, and the milling process planning system can generate the tool paths and NC code within 5 minutes. This study showed the feasibility of CNN on process planning. It is expected that the system can assist the process planning engineers to highly reduce their loading in the future.

[1] Wu J, Lei R, Peng Y (2022) Manufacturing Feature Recognition Method Based on Subgraph Decomposition. 2022 International Conference on Frontiers of Artificial Intelligence and Machine Learning, Hangzhou, China. 120-126. https://doi.org/10.1109/FAIML57028.2022.00031
[2] Hendrawan YM, Muttaqin R, Pratama A, Harja HB, Rasyid MUHA, Winarno I (2021) Intersection Cylindrical Feature Recognition Algorithm for Counterbore and Countersink Geometry Application. 2021 International Electronics Symposium, Surabaya, Indonesia. 538-544. https://doi.org/10.1109/IES53407.2021.9594026
[3] Yan X, Yamazaki K, Liu J (2000) Recognition of Machining Features and Feature Topologies from NC Programs. Computer-Aided Design. Vol. 32 Issue 10: 605-616. https://doi.org/10.1016/S0010-4485(00)00045-2
[4] Zhang Z, Jaiswal P, Rai R (2018) FeatureNet: Machining Feature Recognition Based on 3D Convolution Neural Network. Computer-Aided Design. Vol. 101: 12-22. https://doi.org/10.1016/j.cad.2018.03.006
[5] Zhang H, Zhang S, Zhang Y, Liang J, Wang Z (2022) Machining Feature Recognition Based on a Novel Multi-task Deep Learning Network. Robotics and Computer-Integrated Manufacturing. Vol. 77. https://doi.org/10.1016/j.rcim.2022.102369
[6] Qi CR, Su H, Mo K, Guibas LJ (2017) PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation. 2017 Computer Vision and Pattern Recognition. https://doi.org/10.48550/arXiv.1612.00593
[7] Ji W, Wang L, Haghighi A, Givehchi M, Liu X (2016) A Reachability Based Approach for Machining Feature Sequencing. Journal of Manufacturing Systems. Vol. 40 Part 1: 96-104. https://doi.org/10.1016/j.jmsy.2016.06.006
[8] Hashimoto M, Nakamoto K (2019) A Neural Network Based Process Planning System to Infer Tool Path Pattern for Complicated Surface Machining. International Journal of Automation Technology Vol. 13 No. 1: 67-73. https://doi.org/10.20965/ijat.2019.p0067
[9] Hashimoto M, Nakamoto K (2021) Process Planning for Die and Mold Machining Based on Pattern Recognition and Deep Learning. Journal of Advanced Mechanical Design, Systems, and Manufacturing. Vol. 15 Issue 2. https://doi.org/10.1299/jamdsm.2021jamdsm0015
[10] Cicek O, Abdulkadir A, Lienkamp SS, Brox T, Ronneberger O (2016) 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation. 2016 Medical Image Computing and Computer-Assisted Intervention. 424-432. https://doi.org/10.1007/978-3-319-46723-8_49
[11] Chen ZC, Dong Z, Vickers GW (2003) Automated Surface Subdivision and Tool Path Generation for 31212-Axis CNC Machining of Sculptured Parts. Computers in Industry. Vol. 50 Issue 3: 319-331. https://doi.org/10.1016/S0166-3615(03)00019-8
[12] Inui M, Taguchi S, Umezu N (2023) Graphical Assistance for Determining Cutter Axis Directions in 3+2-Axis Machining. Computer-Aided Design and Applications. Vol. 20 No. 4: 689-703. https://doi.org/10.14733/cadaps.2023.689-703
[13] Peddireddy D, Fu X, Shankar A, Wang H, Joung BG Aggarwal V, Sutherland JW, Jun MBG (2021) Identifying Manufacturability and Machining Processes Using Deep 3D Convolutional Networks. Journal of Manufacturing Processes. Vol. 64: 1336-1348. https://doi.org/10.1016/j.jmapro.2021.02.034
[14] 張宏銘 (2022) 使用卷積神經網路開發 2.5D 零件自動化銑削製程規劃系統, 國立成功大學機械工程學系碩士學位論文.
[15] 李佳政 (2019) 以 NX 開發被夾持零件之自動刀具路徑規劃系統, 國立台灣科技大學機械工程學系碩士學位論文.
[16] 黃柏松 (2023) 透過緊密連接卷積神經網路開發自動化銑削製程規劃系統, 國立成功大學機械工程學系碩士學位論文.
[17] Huang G, Liu Z, Maaten LVD, Weinberger KQ (2017) Densely Connected Convolutional Networks. 2017 IEEE Conference on Computer Vision and Pattern Recognition. 4700-4708. https://doi.org/10.48550/arXiv.1608.06993
[18] 震虎精密科技股份有限公司 (2019) 鎢鋼鑽頭型錄 Carbide Drills Catalogue. https://www.speedtiger.com.tw/document/ Accessed 19 June 2024
[19] 震虎精密科技股份有限公司 (2019) 震虎綜合型錄 General Catalogue. https://www.speedtiger.com.tw/document/ Accessed 19 June 2024
[20] Block HD (1970) A Review of "Perceptrons: An Introduction to Computational Geometry". Information and Control. Vol. 17 Issue 5: 501-522. https://doi.org/10.1016/S0019-9958(70)90409-2.
[21] Freund Y, Schapire RE (1999) Large Margin Classification Using the Perceptron Algorithm. Machine Learning. Vol. 37: 277-296. https://doi.org/10.1023/A:1007662407062
[22] Lecun Y, Boser B, Denker JS, Henderson D, Howard RE, Hubbard W, Jackel LD (1990) Handwritten Digit Recognition with a Back-Propagation Network. Advances in neural information processing systems 2. Morgan Kaufmann Publishers Inc. San Francisco, CA, United States. 396-404.
[23] Bromley J, Guyon I, Lecun Y, Sackinger E, Shah R (1993) Signature Verification Using a "Siamese" Time Delay Neural Network. Proceedings of the 6th International Conference on Neural Information Processing Systems. 737-744.
[24] Lecun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-Based Learning Applied to Document Recognition. Proceedings of the IEEE. Vol. 86 Issue 11: 2278-2324. https://doi.org/10.1109/5.726791
[25] Ciresan D, Meier U, Schmidhuber J (2012) Multi-Column Deep Neural Networks for Image Classification. 2012 Computer Vision and Pattern Recognition. 3642-3649. https://doi.org/10.48550/arXiv.1202.2745
[26] Sanchez J, Perronnin F (2011) High-Dimensional Signature Compression for Large-Scale Image Classification. 2011 Computer Vision and Pattern Recognition. 1665-1672. https://doi.org/10.1109/CVPR.2011.5995504
[27] Krizhevsky A, Sutskever I, Hinton GE (2017) ImageNet Classification with Deep Convolutional Neural Networks. Communications of the ACM. Vol. 60 Issue 6: 84-90. https://doi.org/10.1145/3065386
[28] Simonyan K, Zisserman A (2015) Very Deep Convolutional Networks for Large-Scale Image Recognition. 2015 International Conference on Learning Representations. https://doi.org/10.48550/arXiv.1409.1556
[29] Liu S, Deng W (2015) Very Deep Convolutional Neural Network Based Image Classification Using Small Training Sample Size. 2015 3rd IAPR Asian Conference on Pattern Recognition. 730-734. https://doi.org/10.1109/ACPR.2015.7486599
[30] Girshick R, Donahue J, Darrell T, Malik J (2014) Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation. 2014 IEEE Conference on Computer Vision and Pattern Recognition. 580-587. https://doi.org/10.1109/CVPR.2014.81
[31] Girshick R (2015) Fast R-CNN. 2015 IEEE International Conference on Computer Vision. 1440-1448. https://doi.org/10.1109/ICCV.2015.169
[32] Long J, Shelhamer E, Darrell T (2015) Fully Convolutional Networks for Semantic Segmentation. 2015 IEEE Conference on Computer Vision and Pattern Recognition. 3431-3440. https://doi.org/10.1109/CVPR.2015.7298965
[33] He K, Gkioxari G, Dollar P, Girshick R (2018) Mask R-CNN. 2017 IEEE International Conference on Computer Vision. 2980-2988. https://doi.org/10.1109/ICCV.2017.322
[34] Zheng H, Lin F, Feng X, Chen Y (2020) A Hybrid Deep Learning Model With Attention-Based Conv-LSTM Networks for Short-Term Traffic Flow Prediction. IEEE Transactions on Intelligent Transportation Systems. Vol. 22 Issue 11: 6910-6920. https://doi.org/10.1109/TITS.2020.2997352
[35] Sinha S, Sunil P, Bharill N, Patel OP (2023) HybridSeqNet: A Multimodal Approach Incorporating Convolutional and Long Short-Term Memory Networks for Comprehensive Structural Protein Classification. 2023 IEEE/ACIS 8th International Conference on Big Data, Cloud Computing, and Data Science. 289-294. https://doi.org/10.1109/BCD57833.2023.10466300
[36] He K, Zhang X, Ren S, Sun J (2016) Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition. 770-778. https://doi.org/10.1109/CVPR.2016.90
[37] Szegedy C, Liu W, Jia Y, Sermanet P, Reed S, Anguelov D, Erhan D, Vanhoucke V, Rabinovich A (2015) Going Deeper with Convolutions. 2015 IEEE Conference on Computer Vision and Pattern Recognition. 1-9. https://doi.org/10.1109/CVPR.2015.7298594
[38] Lecun Y (1989) Generalization and Network Design Strategies. R. Pfeifer, Z. Schreter, F. Fogelman, & L. Steels (Eds.), Connectionism in perspective Elsevier.
[39] Jarrett K, Kavukcuoglu K, Ranzato M, Lecun Y (2009) What is the Best Multi-Stage Architecture for Object Recognition? 2009 IEEE 12th International Conference on Computer Vision. 2146-2153. https://doi.org/10.1109/ICCV.2009.5459469