隨著當今資訊及電腦科技的發展與進步，越來越多具自走操作的無人搬運系統出現於日常工作環境。自走無人化是近幾年快速發展的新興技術，根據資科領域的即時定位及地圖建構理論，並結合計算視覺系統輔助辨別能力，可達成自走導航及意外處理的功能。本論文針對文獻中已發表之即時區域地圖技術進行基本研究，對適用於無人物流自走搬運車之法則進行篩選，再結合計算視覺進行辨別路徑障礙物或行人，完成自走路徑規劃。對於地圖共享與群體互用計算資源，以去中心化架構完成群體間信息交流及分享，建構自走定位、路徑規劃、導航及群體信息交流的自走物流搬運車系統。
本論文首先進行即時區域地圖建構技術文獻回顧，其中包括使用傳統傳感器的即時區域地圖建構技術與使用視覺法的即時區域地圖建構技術，並對其基本定義與習知技術內涵，選擇其中合適方法測試於本研究之應用。進而針對文獻上習知之計算機視覺中的特徵點法與當下熱門的深度學習辨識法進行分析比較，選擇快速且準確之計算法則。最終使用機器人操作系統整合上述應用於自走搬運車之軟硬體系統，在此基礎上加入去中心化之構想與初步方法並進行相關技術之驗證，期望能提供自走物流搬運車之研究發展重要參考。

With the development and progress of information and computer technology, more and more unmanned handling systems with self-propelled operation appear in daily working environment. According to the theory of real-time positioning and map construction in the field of information science, combined with the auxiliary discrimination ability of computational vision system, it can achieve the functions of self-propelled navigation and accident handling. In this paper, based on the published real-time area map technology, the rules suitable for unmanned logistics self-propelled vehicles are screened, and then combined with computational vision to identify obstacles or pedestrians, the self-propelled path planning is completed. For map sharing and group interactive computing resources, information exchange and sharing among groups are completed by decentralized architecture, and a self-propelled logistics carrier system with self-propelled positioning, path planning, navigation and group information exchange is constructed.
Firstly, this paper reviews the literature of real-time regional map construction technology, including the real-time regional map construction technology using traditional sensors and the real-time regional map construction technology using visual method. Then, the basic definition and technical connotation of the real-time regional map construction technology are studied, and the appropriate methods are selected to test the application of this study. We compare the feature point method of computer vision in literature with the popular deep learning identification method, and choose the fast and accurate calculation method. Finally, the robot operating system is used to integrate the above software and hardware systems applied to self-propelled transporters. On this basis, the concept and preliminary method of decentralization are added, and the related technologies are verified. It is expected that this paper can provide an important reference for the research and development of self-propelled logistics carriers.

摘 要----------------------------I
Abstract------------------------II
目 錄---------------------------IV
圖目錄---------------------------VII
表目錄---------------------------XI
第一章 簡介-----------------------1
1-1 研究背景----------------------1
1-2 研究目的----------------------2
1-3 文獻回顧----------------------3
1-3-1即時定位及地圖建構技術--------4
1-3-2 機器人操作系統---------------5
1-3-3 目標檢測系統----------------6
1-3-4 去中心化技術----------------7
第二章 基礎理論-------------------10
2-1 即時定位及地圖建構技術---------10
2-1-1 ORB-SLAM-------------------13
2-1-2 LSD-SLAM-------------------15
2-1-3 Gmapping-------------------17
2-2 ROS操作系統-------------------19
2-2-1 ROS------------------------19
2-2-2 ROS 2----------------------19
2-3目標檢測方法-------------------20
2-3-1 特徵點法--------------------21
2-3-2 YOLO-----------------------22
2-3-3 R-CNN----------------------24
2-3-4 Tensor RT推斷方法-----------25
2-4去中心化技術-------------------26
2-4-1 區塊鏈與聯盟學習-------------26
2-4-2 集群機器人技術---------------27
第三章 系統模擬分析----------------41
3-1 模擬平台與資料介紹-------------41
3-2 即時定位及地圖建構模擬----------42
3-2-1 ORB-SLAM的模擬--------------42
3-2-2 LSD-SLAM的模擬--------------43
3-2-3 Gmapping的模擬--------------44
3-3 目標檢測法的模擬---------------45
3-3-1 特徵點法的模擬---------------45
3-3-2 YOLOv3的模擬----------------46
3-3-3 Faster R-CNN的模擬 ---------46
3-3-4 TensorRT Inference的模擬----46
3-4 去中心化模擬------------------47
3-4-1 區塊鏈模擬------------------47
3-4-2 ROS 2去中心化模擬-----------48
3-4-3 集群機器人模擬--------------48
3-5 模擬分析結論------------------49
第四章 自走搬運車設計與驗證---------62
4-1 實驗架構與流程-----------------62
4-2 硬體設備介紹-------------------62
4-3 SLAM的實驗與整合---------------63
4-4 目標檢測技術實驗與整合----------64
4-5自走車設計與製造----------------66
4-6自走車驗證實驗------------------67
第五章 結論與討論------------------82
5-1結論---------------------------82
5-2 未來展望----------------------84
參考文獻--------------------------86

[1] C. Kerl, J. Sturm and D. Cremers, "Dense visual SLAM for RGB-D cameras," 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems, Tokyo, pp. 2100-2106, 2013.
[2] A. Diosi, G. Taylor and L. Kleeman, "Interactive SLAM using Laser and Advanced Sonar," Proceedings of the 2005 IEEE International Conference on Robotics and Automation, Barcelona, Spain, pp. 1103-1108, 2005.
[3] S. Jingwei, W. Jun, Z. Liang, et al. "MIS-SLAM: Real-time Large Scale Dense Deformable SLAM System in Minimal Invasive Surgery Based on Heterogeneous Computing". IEEE Robotics & Automation Letters :1-1, 2018
[4] M. Quigley, K. Conley, P. Gerkey, J. Faust, T. Foote, J. Leibs, R. Wheeler, and Andrew Y. Ng. "Ros: an open-source robot operating system. " In ICRA Workshop on Open Source Software, 2009.
[5] S. Kohlbrecher, O. von Stryk, J. Meyer, and U. Klingauf , "A flexible and scalable SLAM system with full 3D motion estimation. " In 2011 IEEE International Symposium on Safety, Security, and Rescue Robotics, pages 155–160, Nov., 2011.
[6] J. M. Santos, D. Portugal and R. P. Rocha, "An evaluation of 2D SLAM techniques available in Robot Operating System," 2013 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), Linkoping, pp. 1-6, 2013.
[7] P. Benavidez, M. Muppidi, P. Rad, J. J. Prevost, M. Jamshidi and L. Brown, "Cloud-based realtime robotic Visual SLAM," 2015 Annual IEEE Systems Conference (SysCon) Proceedings, Vancouver, BC, pp. 773-777, 2015.
[8] 劉维一, 盛益强,”特徵點匹配法實現汽車牌照的快速試别”,光電子激光(3):273-276, 2002.
[9] B. Herbert, E. Andreas, T. Tinne, "Speeded-Up Robust Features". computer vision & image understanding, 110(3):404-417, 2008.
[10] H. Jiang and E. Learned-Miller, "Face Detection with the Faster R-CNN," 2017 12th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2017), Washington, DC, pp. 650-657, 2017.
[11] D. Gordon, A. Kembhavi, M. Rastegari, J. Redmon, D. Fox, and A. Farhadi, "Iqa: Visual question answering in interactive en- vironments, " In the IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), June 2018.
[12] D. R. Rehak, P. Dodds, and L. Lannom, "A model and infras- tructure for federated learning content repositories. " In WWW 2005, 2005.
[13] J. Konecny´, H. B. McMahan, Felix X. Yu, Peter Richt´arik, Ananda Theertha Suresh, and D. Bacon, "Federated learning: Strategies for improving communication efficiency. "CoRR, abs/1610.05492, 2016.
[14] M. Reke M, D. Peter, J. Schulte-Tigges, et al, "A Self-Driving Car Architecture in ROS2" .2020 International SAUPEC/RobMech/PRASA Conference, 2020.
[15] N. L. Husni, A. S. Handayani, S. Nurmaini, and I. Yani, "Cooperative searching strategy for swarm robot. " In2017 International Conference on Electrical Engineering and Computer Science (ICECOS), pages 92–97, 2017.
[16] D. Q. Bao and I. Zelinka, "Obstacle avoidance for swarm robot based on self-organizing migrating algorithm." Procedia Computer Science,150:425–432, 01 2019.
[17] J. Luo, G. Oubong, "A comparison of sift, pca-sift and surf. " International Journal of Image Processing, 2009.
[18] R. Mur-Artal, J. M. M. Montiel, and J. D. Tard´os, "Orb-slam: A versatile and accurate monocular slam system. " IEEE Transactions on Robotics, 31(5):1147– 1163, Oct 2015.
[19] R. Mur-Artal and J. D. Tard´os, "Orb-slam2: An open-source slam system for monocular, stereo, and rgb-d cameras. " IEEE Transactions on Robotics, 33(5):1255–1262, Oct 2017.
[20] E. Rublee, V. Rabaud, K. Konolige, and G. Bradski, "Orb: An efficient alternative to sift or surf. " In 2011 International Conference on Computer Vision, pages 2564–2571, Nov 2011.
[21] M. Calonder, V. Lepetit, C. Strecha, and P. Fua, "Brief: Binary robust independent elementary features. " In Kostas Daniilidis, Petros Maragos, and Nikos Paragios, editors, Computer Vision – ECCV 2010, pages 778–792, Berlin, Heidelberg, Springer Berlin Heidelberg, 2010.
[22] J. Engel, T. Sch¨ops, and D. Cremers, "Lsd-slam: Large-scale direct monocular slam. "In David Fleet, Tomas Pajdla, Bernt Schiele, and Tinne Tuytelaars, editors, Computer Vision – ECCV 2014, pages 834–849, Cham, Springer International Publishing, 2014.
[23] G. Grisetti, C. Stachniss, and W. Burgard, "Improved techniques for grid map- ping with rao-blackwellized particle filters. " IEEE Transactions on Robotics, 23(1):34–46, Feb 2007.
[24] Y. Maruyama, S. Kato, T. Azumi, "Exploring the performance of ROS2." The 13th International Conference. IEEE, 2016.
[25] 楊曉敏、吳炜、卿鄰波,"圖像特徵點提取及匹配技術",光學精密工程,017 (009):2276-2282, 2009.
[26] J. Redmon, S. K. Divvala, R. B. Girshick, and A. Farhadi. "You only look once: Unified, real-time object detection. " CoRR, abs/1506.02640, 2015.
[27] J. Redmon and A. Farhadi, "Yolov3: An incremental improvement. " CoRR, abs/1804.02767, 2018.
[28] R. Girshick, J. Donahue, T. Darrell, and J. Malik, "Rich feature hierarchies for accurate object detection and semantic segmentation. " In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2014.
[29] R. Girshick, "Fast r-cnn. " In The IEEE International Conference on Computer Vision (ICCV), December 2015.
[30] S. Q. Ren, K. M. He, R. Girshick, and J. Sun, "Faster r-cnn: Towards real-time object detection with region proposal networks. " In C. Cortes, N. D. Lawrence, D. D. Lee, M. Sugiyama, and R. Garnett, editors, Advances in Neural Information Processing Systems 28, pages 91–99. Curran Associates, Inc., 2015.
[31] Q. Yang, Y. Liu, T. Chen, and Y.X. Tong, "Federated machine learning: Concept and applications." ACM Trans. Intell. Syst.Technol., 10(2), January 2019
[32] 王偉嘉,鄭雅婷,林國政,等.集群機器人研究綜述.機器人(2):232-256, 2020.
[33] Ngai Kou, C. Peng, H. Ma, T. Kumar, and S. Koenig, "Idletime optimization for target assignment and path finding in sortation centers. " Proceedings of the AAAI Conference on Artificial Intelligence, 34:9925–9932, 04 2020.
[34] A. Carballo, S. Seiya, J. Lambert, H. Darweesh, P. Narksri, L. Y. Morales, N. Akai, E. Takeuchi, and K. Takeda, "End-to-end autonomous mobile robot navigation with model-based system support." Journal of Robotics and Mechatronics, 30(4):563–583, 2018.