本論文主要是在為一台具麥卡納姆輪型全向輪及一隻六軸機械手之移動機器人增加自主跟隨人之移動性。基本上就是使機器人 具有物體(人員)辨識、物體(人員)追蹤、物體(人員)跟隨及避障能力。且整合之系統可經由人員語音命令的方式進行操控。
首先本系統使用由兩台網路攝影機自製之3D相機，進行相機校正後，並使用Opencv函式庫計算出兩台相機的像差圖以及換算出實際距離，然後再使用Yolov4及Deep SORT演算法進行物體辨識及追蹤，並使用標籤定位。最後配合本研究開發的路徑規劃演算法以完成人員跟隨及避障功能。
本論文結合了包含語音命令辨識、立體視覺、導航、Socket通訊、標籤定位等模組，及整合現有全向輪機器人、六自由度機器手臂等模組，並將整套系統軟體安裝於筆記型電腦上，使本研究開發之移動機器人可以實現自主人員追隨及避障路徑規劃。

This research mainly enhances a mobile robot platform with autonomous people-following mobility, i.e. with the ability to recognize objects, track people, follow people and avoid obstacles, and the system can be controlled by people's voice commands. The robot platform is with the omni wheels and a six-axis robot arm.
First, this system uses a 3D camera made by two network cameras. After camera calibration, the OpenCV library was used to create a disparity map of the two cameras and compute the real distance. Then Yolov4 and Deep SORT algorithms were introduced for object detection and tracking. Label(marker) positioning was used, too. Finally, they are used and cooperated with the route planning algorithm developed in this research to complete the functions of human following and obstacle avoidance.
This research glues modules including speech recognition, stereo vision, navigation, socket communication, marker positioning and integrates existing omni wheel robots, six-axis robot arm and other modules to serve as a mobile robot that would follow a person and avoid collision with people and objects.

[1] Wikipedia, "Kinect," [online].Availiable:
https://en.wikipedia.org/w/index.php?title=Kinect&oldid=1071410096 [Accessed: 10. Feb,2022]
[2] B. X. Chen, R. Sahdev and J. K. Tsotsos, "Person Following Robot Using Selected Online Ada-Boosting with Stereo Camera," in 2017 14th Conference on Computer and Robot Vision (CRV), 2017, pp. 48-55, doi: 10.1109/CRV.2017.55.
[3] S. Janzon, C. M. Sánchez, M. Zella, and P. J. Marrón, "Person Re-Identification in Human Following Scenarios: An Experience with RGB-D Cameras," in 2020 Fourth IEEE International Conference on Robotic Computing (IRC), 9-11 Nov. 2020 2020, pp. 424-425, doi: 10.1109/IRC.2020.00076.
[4] T. H. Tsai, P. H. Wu, X. H. Lin, and P. T. Chi, "Live Demonstration: Intelligent Voice Wake-up Human-Following Robot," in 2021 IEEE International Symposium on Circuits and Systems (ISCAS), 22-28 May 2021 2021, pp. 1-1, doi:10.1109/ISCAS51556.2021.9401532.
[5] M. Hua, Y. Nan, and S. Lian, "Small Obstacle Avoidance Based on RGB-D Semantic Segmentation," in 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), 27-28 Oct. 2019 2019, pp. 886-894, doi:10.1109/ICCVW.2019.00117.
[6] A. Hussein, P. Marín-Plaza, D. Martín, A. d. l. Escalera, and J. M. Armingol, "Autonomous Off-road Navigation Using Stereo-Vision and Laser-Rangefinder Fusion for Outdoor Obstacles Detection," in 2016 IEEE Intelligent Vehicles Symposium (IV), 19-22 June 2016 2016, pp. 104-109, doi: 10.1109/IVS.2016.7535372.
[7] A. J. Sanchez-Garcia, H. V. Rios-Figueroa, X. Limon-Riaño, J. A. Sanchez-Garcia, and K. Cortes-Verdin, "Towards Learning Obstacles to Avoid Collisions in Autonomous Robot Navigation," in 2019 International Conference on Mechatronics, Electronics and Automotive Engineering (ICMEAE), 26-29 Nov. 2019 2019, pp. 24-27, doi: 10.1109/ICMEAE.2019.00012.
[8] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, "Yolov4: Optimal speed and accuracy of object detection," in 2020 arXiv preprint arXiv:2004.10934
[9] N. Wojke and A. Bewley, "Deep Cosine Metric Learning for Person Re-identification," in 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), 12-15 March 2018 2018, pp. 748-756, doi: 10.1109/WACV.2018.00087.
[10] 洪瑋呈，麥卡倫輪型移動機器人之路徑追蹤控制，國立成功大學工程科學系碩士在職專班碩士論文，2020。
[11] 劉昇宏，六自由度機器手臂設計實作與軌跡規劃，國立成功大學工程科學系碩士在職專班碩士論文，2020。
[12] A. Zhang. "Speech Recognition " [online].Availiable:
https://github.com/Uberi/speech_recognition#readme. [Accessed: 10. Feb, 2022]
[13] Wikipedia, "Pinhole camera," [online].Availiable:
https://en.wikipedia.org/w/index.php?title=Pinhole_camera&oldid=1073322182.[Accessed: 5. Jan, 2022]
[14] Opencv. "Camera Calibration." [online].Availiable:
https://docs.opencv.org/3.4/dc/dbb/tutorial_py_calibration.html [Accessed: 21, Sep, 2021]
[15] H. Hirschmuller, "Stereo Processing by Semiglobal Matching and Mutual Information," in IEEE Transactions on pattern analysis and machine intelligence, vol. 30, no. 2, pp. 328-341, 2007.
[16] F. Uhrweiller. "Stereo-Vision." [online].Availiable:
https://github.com/LearnTechWithUs/Stereo-Vision [Accessed: 12. Sep,2021]
[17] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You Only Look Once: Unified, Real-time Object Detection," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 779-788.
[18] T.-Y. Lin et al., "Microsoft coco: Common objects in context," in European conference on computer vision, 2014: Springer, pp. 740-755.
[19] Wikipedia, "TensorFlow," [online].Availiable: https://en.wikipedia.org/w/index.php?title=TensorFlow&oldid=1075725116 [Accessed: 22. Oct,2021]
[20] 謝忠霖，具有自主開門及通過門道能力之服務型機器人設計與實作，國立成功大學工程科學系碩士在職專班碩士論文，2022