請用此 Handle URI 來引用此文件：
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/94661完整後設資料紀錄
| DC 欄位 | 值 | 語言 |
|---|---|---|
| dc.contributor.advisor | 吳安宇 | zh_TW |
| dc.contributor.advisor | An-Yeu Wu | en |
| dc.contributor.author | 楊絜雯 | zh_TW |
| dc.contributor.author | Chieh-Wen Yang | en |
| dc.date.accessioned | 2024-08-16T17:23:15Z | - |
| dc.date.available | 2024-08-17 | - |
| dc.date.copyright | 2024-08-16 | - |
| dc.date.issued | 2024 | - |
| dc.date.submitted | 2024-08-01 | - |
| dc.identifier.citation | [1] C. Resende, et al. "TIP4.0: industrial internet of things platform for predictive maintenance," Sensors, vol.21, no.14, pp.4676, 2021.
[2] Sii Poland, "Industry 4.0: The Industrial Revolution," https://sii.pl/blog/en/industry-4-0-the-industrial-revolution/ [3] Predictive Maintenance Technology L.L.C. "D-I-P-F Curve: Design, Installation, Potential Failure, and Failure," https://predictivemaint.com/ [4] Statista, "Global Data Creation is About to Explode," https://www.statista.com/chart/17727/global-data-creation-forecasts/ [5] Y. Koizumi et al., “Description and discussion on DCASE2020 challenge task2: Unsupervised anomalous sound detection for machine condition monitoring,” in Proceedings of the Detection and Classification of Acoustic Scenes and Events 2020 Workshop (DCASE2020), November 2020, pp. 81–85. [6] R. Giri, et al. “Self-supervised classification for detecting anomalous sounds,” in Proc. Detection and Classification of Acoustic Scenes and Events Workshop (DCASE), pp. 46–502020. [7] H. Hojjati and N. Armanfard, "Self-supervised acoustic anomaly detection via contrastive learning," in Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022, pp. 3253–3257. [8] Y. Liu, J. Guan, Q. Zhu, and W. Wang, "Anomalous sound detection using spectral-temporal information fusion," in ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022, pp. 816–820. [9] H. Chen, L. Ran, X. Sun, and C. Cai, "SW-WAVENET: Learning representation from spectrogram and wavegram using wavenet for anomalous sound detection," in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023, pp. 1–5. [10] J. Guan, et al., “Anomalous sound detection using audio representation with machine ID based contrastive learning pretraining,” in ICASSP 2023 -2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023, pp. 1–5. [11] H. Zhang, et al., "Anomalous sound detection using self-attention-based frequency pattern analysis of machine sounds," arXiv preprint arXiv:2308., 2023. [12] T.-L. Tsai, et al., "An Efficient Anomalous Sound Detection by Robust Processing and Reformation of Objective," Proc. IEEE 6th Int. Conf. Artif. Intell. Circuits Syst. (AICAS), Apr. 2024. [13] J. Holdsworth et al., "Implementing a gammatone filter bank," Annex C of the SVOS Final Report: Part A: The Auditory Filterbank, vol. 1, pp. 1-5, 1988. [14] S. Choi and J.-W. Choi, "Noisy-ArcMix: Additive noisy angular margin loss combined with mixup anomalous sound detection," arXiv preprint arXiv:2310.06364, 2023. [15] H. Zhang et al., "mixup: Beyond empirical risk minimization," in Proc. 6th International Conference on Learning Representations (ICLR), Vancouver, BC, Canada, Apr. 30 - May 3, 2018, Conference Track Proceedings, 2018. [16] Yuma Koizumi et al., “ToyADMOS: A dataset of miniature-machine operating sounds for anomalous sound detection,” in Proceedings of IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA), November 2019, pp. 308–312. [17] Harsh Purohit et al., “MIMII Dataset: Sound dataset for malfunctioning industrial machine investigation and inspection,” in Proceedings of the Detection and Classification of Acoustic Scenes and Events 2019 Workshop (DCASE2019), November 2019, pp. 209–213. [18] E. Liberis and N. D. Lane, "Differentiable Neural Network Pruning to Enable Smart Applications on Microcontrollers," Proc. ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, vol. 6, no. 4, pp. 1-19, 2023. [19] S. Chen et al., "Mobilefacenets: Efficient CNNs for accurate real-time face verification on mobile devices," in Biometric Recognition: 13th Chinese Conference, CCBR 2018, Urumqi, China, August 11-12, 2018, Proceedings, vol. 13, Springer International Publishing, 2018. [20] M. Sandler et al., "Mobilenetv2: Inverted residuals and linear bottlenecks," in Proc. IEEE Conference on Computer Vision and Pattern Recognition, 2018. [21] L. Chen et al., "Knowledge from the original network: restore a better pruned network with knowledge distillation," Complex & Intelligent Systems, vol. 8, no. 2, pp. 709-718, 2022. [22] Y. He et al., "AMC: AutoML for model compression and acceleration on mobile devices," in Proc. European Conference on Computer Vision (ECCV), 2018. [23] T. P. Lillicrap et al., "Continuous control with deep reinforcement learning," arXiv preprint arXiv:1509.02971, 2015. [24] N. Ketkar and N. Ketkar, "Introduction to Keras," in Deep Learning with Python: A Hands-on Introduction, 2017, pp. 97-111. [25] S. Imambi, K. B. Prakash, and G. R. Kanagachidambaresan, "PyTorch," in Programming with TensorFlow: Solution for Edge Computing Applications, 2021, pp. 87-104. [26] J. Bai, F. Lu, K. Zhang, et al., "ONNX: Open Neural Network Exchange," 2019, https://github.com/onnx/onnx. [27] PINTO0309,"onnx2tf," GitHub repository, https://github.com/PINTO0309/onnx2tf. [28] E. Liberis, "tflite-tools," GitHub repository, https://github.com/eliberis/tflite-tools. [29] D. David, R. Robert, et al., "TensorFlow Lite Micro: Embedded Machine Learning for TinyML Systems," in Proc. Machine Learning and Systems, vol. 3, pp. 800-811, 2021. [30] L. Lai, N. Suda, and V. Chandra, "CMSIS-NN: Efficient neural network kernels for ARM Cortex-M CPUs," arXiv preprint arXiv:1801.06601, 2018. [31] ARM Software, "CMSIS-DSP," GitHub repositor | - |
| dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/94661 | - |
| dc.description.abstract | 在工業4.0中，預測性維護將人工智慧（AI）和物聯網（IoT）整合在一起，形成工業物聯網（IIoT）。預測性維護的核心目標是及早發現潛在的機器故障，以減少計劃外停機時間並將中斷降到最低。異常聲音檢測（Anomalous sound detection）系統非常重要，因為它們能夠偵測出問題機械所產生的異常聲音，從而在問題發展成重大故障之前進行預警。因此，開發能夠在邊緣設備上運行的高效且可靠的異常聲音檢測系統，有助於提升預測性維護策略的可靠性和效能。
然而，在邊緣設備上部署高性能異常聲音檢測系統面臨兩大挑戰。首先，在吵雜的環境中準確預測異常聲音具有高運算複雜度。其次，在微控制器 (MCU) 部署異常聲音檢測系統需克服嚴格的記憶體限制。為了打造輕量且高效的異常聲音檢測系統，我們在模型訓練階段導入了不同的資料處理方法並且加入新的訓練策略。此外，為了符合微控制器的嚴格記憶體限制，我們設計了一種記憶體導向的自動模型壓縮(Memory-oriented automated model compression)算法，該算法使用強化學習(Reinforcement learning)方法來計算最佳壓縮比。最後，我們詳細介紹了在微控制器上完整部署異常聲音檢測系統的過程。 透過解決有限的異常數據、環境中的噪音干擾和微控制器上記憶體資源不足等問題，我們提供了一種適用於邊緣裝置的即時異常聲音檢測方案。這種方法提升了工業4.0中預測性維護系統的可靠性和效率。 | zh_TW |
| dc.description.abstract | Predictive maintenance in Industry 4.0 integrates Artificial Intelligence (AI) and the Internet of Things (IoT), forming the Industrial Internet of Things (IIoT). A crucial aspect of predictive maintenance is the early detection of potential machine failures to reduce downtimes and minimize disruptions. Anomalous sound detection (ASD) systems are essential because they detect irregular sounds that signal mechanical issues before they develop into major problems. Therefore, developing an efficient and robust ASD system capable of operating on edge devices is essential for enhancing the reliability and effectiveness of predictive maintenance strategies.
However, deploying high-performance ASD systems on edge devices faces two main challenges. First, the high complexity of accurately predicting anomaly sounds amidst noisy environmental disturbances. Second, deploying ASD systems within the tight memory constraints inherent in microcontrollers (MCUs). To create a lightweight and efficient ASD system, we implemented robust processing methods and reformulated the training strategy during the model training stage. Additionally, to comply with the strict memory limitations of MCUs, we designed a memory-oriented automatic model compression algorithm that uses a reinforcement learning agent to determine the optimal compression ratio for model deployment. Finally, we detailed the process of deploying the ASD system on MCUs to achieve end-to-end deployment. By solving challenges such as limited abnormal data, environmental interference, and MCU memory constraints, we provide a practical solution for real-time anomaly sound detection systems on resource-constrained edge devices. This approach enhances the reliability and efficiency of predictive maintenance systems in Industry 4.0. | en |
| dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2024-08-16T17:23:15Z No. of bitstreams: 0 | en |
| dc.description.provenance | Made available in DSpace on 2024-08-16T17:23:15Z (GMT). No. of bitstreams: 0 | en |
| dc.description.tableofcontents | 誌謝 v
摘要 vii ABSTRACT ix CONTENTS xi LIST OF FIGURES xv LIST OF TABLES xix Chapter 1 Introduction 1 1.1 Background 1 1.1.1 Predictive Maintenance in Industry 4.0 1 1.1.2 Machine Anomaly Detection (MAD) 3 1.1.3 Edge Intelligence 4 1.2 Motivation and Objective 5 1.2.1 Accurate Prediction of Anomalous Sound Detection 5 1.2.2 Model Miniaturization 6 1.2.3 Deployment on Edge Devices 7 1.3 Thesis Target 9 1.4 Thesis Organization 9 Chapter 2 Review of Anomalous Sound Detection System 11 2.1 Anomalous Sound Detection System 11 2.2 Related Works of Anomalous Sound Detection System 15 2.2.1 Multiple-Domain Feature Approach 15 2.2.2 Methods of Related Works 16 2.3 Challenges of the Prior Works 17 2.4 Summary 17 Chapter 3 Lightweight and Robust Anomalous Sound Detection system 19 3.1 Robust Feature Extraction by Dynamic Signal Processing 19 3.1.1 Motivation and Challenge 19 3.1.2 Analyze Dataset 20 3.1.3 Dynamic Signal Processing 21 3.1.4 Redefinition of Operation Region 23 3.2 Optimization of Training Strategy 24 3.2.1 Motivation and Challenge 24 3.2.2 Mix-up technique 24 3.3 Experiment 25 3.3.1 Audio Dataset 25 3.3.2 Experimental Results 26 3.4 Summary 28 Chapter 4 Hardware-Aware Model Compression 29 4.1 Microcontroller Constraints 30 4.1.1 Flash Memory and SRAM on MCU 30 4.1.2 Mapping Neural Network Execution on MCU 30 4.1.3 Flash Constraints for ASD System 32 4.2 Model Information 33 4.2.1 Model Architecture 33 4.2.2 Estimated Peak Memory Usage 34 4.2.3 MobileFaceNet Model Size and Peak Memory Usage 36 4.3 Memory-Oriented Channel Pruning 37 4.3.1 Model Pruning Techniques 37 4.3.2 Related Works: AutoML for Model Compression (AMC) 38 4.3.3 Automated Compression with Reinforcement Learning 39 4.3.4 Memory-Oriented Search Protocols 40 4.4 Experimental Results 42 4.5 Summary 44 Chapter 5 Microcontroller Demonstration 45 5.1 Overview of MCU Deployment Flow 45 5.2 Model Conversion Process 46 5.2.1 Convert Neural Network via ONNX 47 5.2.2 Convert to Tensorflow Lite 50 5.3 Integrating DSP and NN on MCUs 52 5.3.1 Introduce to CMSIS 52 5.3.2 Implementation of DSP Module 54 5.4 Deployment Result 55 5.5 Summary 56 Chapter 6 Conclusions and Future Directions 57 6.1 Main Contributions 57 6.2 Future Directions 59 Bibliography 61 | - |
| dc.language.iso | en | - |
| dc.subject | 模型壓縮 | zh_TW |
| dc.subject | 異常聲音偵測 | zh_TW |
| dc.subject | 邊緣計算 | zh_TW |
| dc.subject | 微控制器單元 | zh_TW |
| dc.subject | Microcontroller Units | en |
| dc.subject | Model Compression | en |
| dc.subject | Edge Computing | en |
| dc.subject | Anomalous Sound Detection | en |
| dc.title | 利用微控制器實現高效率異常聲音檢測系統 | zh_TW |
| dc.title | An Efficient Anomalous Sound Detection System for Microcontrollers | en |
| dc.type | Thesis | - |
| dc.date.schoolyear | 112-2 | - |
| dc.description.degree | 碩士 | - |
| dc.contributor.oralexamcommittee | 馬咏治;沈中安;盧奕璋 | zh_TW |
| dc.contributor.oralexamcommittee | Win-Ken Beh;Chung-An Shen;Yi-Chang Lu | en |
| dc.subject.keyword | 異常聲音偵測,模型壓縮,邊緣計算,微控制器單元, | zh_TW |
| dc.subject.keyword | Anomalous Sound Detection,Model Compression,Edge Computing,Microcontroller Units, | en |
| dc.relation.page | 63 | - |
| dc.identifier.doi | 10.6342/NTU202402622 | - |
| dc.rights.note | 未授權 | - |
| dc.date.accepted | 2024-08-03 | - |
| dc.contributor.author-college | 電機資訊學院 | - |
| dc.contributor.author-dept | 電子工程學研究所 | - |
| 顯示於系所單位： | 電子工程學研究所 | |
文件中的檔案：
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| ntu-112-2.pdf 未授權公開取用 | 22.14 MB | Adobe PDF |
系統中的文件，除了特別指名其著作權條款之外，均受到著作權保護，並且保留所有的權利。