臺灣博碩士論文加值系統

為了減少培訓時間和實現物聯網（IoT）的即時響應。 AIoT系統結構利用動態分散式運算減少訓練時間，通過在霧和雲計算平台中分發訓練數據來縮短訓練時間。經過訓練的模型部署在邊緣計算平台上。 AIoT系統於邊緣計算平台進行推論和雲計算在AI應用中相對於優勢。終端設備通過LoRa，XBee，Wi-Fi將傳感器數據和圖像傳輸到霧計算平台。在本文中，終端設備配置為監控環境數據。先進動態分配方法是一種綜合策略。霧計算平台接收部分訓練數據和測試數據，從雲驅動器傳輸。經過先進的動態分配，霧計算平台通過Node-RED開始訓練模型。通過先進的動態分配，雲將部分數據傳遞到霧計算平台，然後進行訓練模型。主要優點是在減少培訓時間。當霧計算平台和雲計算平台完成模型訓練時，雲將訓練模型轉移到霧計算平台。霧計算平台從雲計算平台接收訓練模型。 Node-RED結合了來自每個分散式設備的訓練模型。 Node-RED將組合模型傳輸到NVIDIA JETSON TX2，並提供TX2以執行數據推斷。 AIoT系統提供傳感器數據以監控事物板。

In order to reduces training time and achieve internet of things (IoT) real-time response. The AIoT system presents an architecture and a dynamic distribution, reducing the training time by distributing training data in fog computing platform and cloud computing platform. The trained model deploys on edge computing platform. The AIoT system presents an architecture which processes the advantages of both edge processing and cloud computing for AI application on sensor data and image recognition. End device transfers sensor data and image to fog computing platform by LoRa, XBee, Wi-Fi. In this thesis, end devices configures to monitor environmental data. The proposed advanced dynamic distribution approach is an integrated strategy. Fog computing platform receives the partial training data and testing data, transmitd from cloud drive. After the advanced dynamic distributing, fog computing platform starts training model by Node-RED. Through the advanced dynamic distribution, cloud passes the partial data to fog computing platform then starts tensorflow training model simultaneously. The main advantage of the proposed strategy is aims to reduce training time. When fog computing platform and cloud computing platform completed model training, cloud transfers trained model to fog computing platform. Fog computing platform receives trained model from cloud computing platform. Node-RED combines trained model from each distributed equipment. Node-RED transfers the combined model to NVIDIA JETSON TX2 and provides TX2 to execute data inference. The AIoT system provides sensor data to monitor on thingsboard.

目錄 I
圖次 II
表次 III
1 Introduction 1
2 Related Works 4
2.1 Related works 4
2.2 Motivation 6
3 Preliminaries 8
3.1 System architecture 8
3.2 Problem formulation 9
3.3 Basic idea 10
4 Dynamic training-dataset distribution approach 12
4.1 End device registration phase 13
4.2 Dynamic training-dataset distribution phase 15
4.3 Dynamic training-dataset training phase 20
4.4 Inference phase 23
4.5 Data visualization phase 25
5 Experimental Results 27
5.1 Total runtime 29
5.2 Accuracy 30
5.3 Inference error rate 32
5.4 Inference efficiency 33
6 Conclusions 35
7 Acknowledgments 36

[1] S. M. Kim, H. S. Choi, and W. S. Rhee, “IoT Home Gateway for Auto-Configuration and Management of MQTT Devices,” in Proceedings of IEEE Conference on Wireless Sensors(ICWiSe 2015), pp. 12–17, Melaka, Malaysia, Jul. 2015.
[2] MQTT, MQTT, http://mqtt.org/.
[3] thingsboard, thingsboard, https://thingsboard.io/.
[4] Node-RED, Node-RED, https://nodered.org/.
[5] H. Kim, J. Park, J. Jang, and S. Yoon, “DeepSpark: Spark-Based Deep Learning Supporting Asynchronous Updates and Caffe Compatibility,” arXiv preprint arXiv:1602.08191, Oct. 2016.
[6] Y. Ren, X. Wu, L. Zhang, Y. Wang, W. Zhang, Z. Wang, M. Hack, and S. Jiang, “iRDMA: Efficient Use of RDMA in Distributed Deep Learning Systems,” in Proceedings of IEEE 19th International Conference on High Performance Computing and Communications (HPCC 2017), pp. 231–238, Bangkok, Thailand, Feb. 2017.
[7] P. Selonen and A. Taivalsaari, “Kiuas – IoT Cloud Environment for Enabling the Programmable World,” in Proceedings of Euromicro Conference on Software Engineering and Advanced Applications (SEAA 2016), pp. 250–257, Limassol, Cyprus, Aug. 2016.
[8] R. K. Kodali and S. Soratkal, “MQTT Based Home Automation System using ESP8266,” in Proceedings of IEEE Region 10 Humanitarian Technology Conference (R10-HTC 2016), pp. 1–5, Agra, India, Dec. 2016.
[9] T. Cao-hoang and C. N. Duy, “Environment Monitoring System for Agricultural Application Based on Wireless Sensor Network,” in Proceedings of Seventh International Conference on Information Science and Technology (ICIST 2017), pp. 99–102, Da Nang, Vietnam, April 2017.
[10] R. K. Kodali and S. Soratkal, “IoT Sensor Integration to Node-RED Platform,” in Proceedings of International Symposium INFOTEH-JAHORINA (INFOTEH 2018), pp. 1–5, East Sarajevo, Bosnia-Herzegovina, March 2018.
[11] J. Rakesh and T. Samir, “Cloud to Edge: Distributed Deployment of Process-Aware IoT Applications,” in Proceedings of IEEE International Conference on Edge Computing (EDGE 2017), pp. 182–189, Honolulu, HI, USA, Sep. 2017.
[12] Y. H. Lee, Y. Hang, and S. Nair, “A Smart Gateway Framework for IOT Services,” in Proceedings of IEEE International Conference on Internet of Things (iThings 2016), pp. 107–114, Chengdu, China, May 2016.
[13] G. Palaiokrassas, I. Karlis, A. Litke, V. Charlaftis, and T. Varvarigou, “An IoT Architecture for Personalized Recommendations over Big Data Oriented Applications,” in Proceedings of Computer Software and Applications Conference (COMPSAC 2017), pp. 475–480, Turin, Italy, July 2017.
[14] E. Yang, S. H. Kim, and T. W. Kim, “An Adaptive Batch-Orchestration Algorithm for the Heterogeneous GPU Cluster Environment in Distributed Deep Learning System,” in Proceedings of IEEE International Conference on Big Data and Smart Computing (BigComp 2018)), Shanghai, China, May 2018.
15] A. Khumoyun, Y. Cui, and L. Hanku, “Spark Based Distributed Deep Learning Framework for Big Data Applications,” in Proceedings of IEEE International Conference on Information Science and Communications Technologies (ICISCT 2016), pp. 1–5, Tashkent, Uzbekistan, Dec. 2016.
[16] M. Lee, S. Shin, S. Hong, and S. kwang Song, “BAIPAS: Distributed Deep Learning Platform with Data Locality and Shuffling,” in Proceedings of European Conference on Electrical Engineering and Computer Science (EECS 2017), pp. 5–8, Bern, Switzerland, Nov. 2017.
[17] S. Shams, S.Goswami, K. Lee, S. Yang, and S.-J. Park, “Towards Distributed Cyberinfrastructure for Smart Cities Using Big Data and Deep Learning Technologies,” in Proceedings of International Conference on Distributed Computing Systems (ICDCS 2018), pp. 1276–1283, Vienna, Austria, July 2018.
[18] S. B. Calo, A. Touna, and D. C. Verma, “Edge Computing Architecture for Applying AI to IoT,” in Proceedings of IEEE International Conference on Big Data (Big Data 2017), pp. 3012–3016, Boston, MA, USA, Jan. 2017.
[19] K. T. A. Sheth, U. Jaimini and T. Banerjee, “Augmented Personalized /health: How Smart Data with IoTs and AI Is About to Change Healthcare,” in Proceedings of 2017 IEEE 3rd International Forum on Research and Technologies for Society and Industry (RTSI 2017), pp. 1–6, Modena, Italy, Sept. 2017.
[20] P. Patel, M. I. Ali, and A. Sheth, “From Raw Data to Smart Manufacturing: AI and Semantic Web of Things for Industry 4.0,” in Proceedings of 2018 IEEE Intelligent Systems (RTSI 2017), pp. 79–86, Boston, MA, USA„ April 2018.