隨著AI人工智慧技術的發展,機器人應用於眾多領域，人工智慧對於機器人領域的重要性不言而喻，尤其在對理解自然語言、機器學習、人工神經網絡、機器視覺等，AI人工智慧的研究中，使機器人在許多領域中廣泛的應用。
而目前市售的服務型機器人均有以下的限制，如功能性單一，無法多工、影像處理技術並不完善、自然語言技術、缺乏高集成度、統一感知協調的中控系統。因此能將目前服務型機器人的缺點加以改善，定能減少家庭的負擔以及增加生活品質。
本文所採用之物體辨識方法為：You only look once (YOLO)、定位精準、辨識迅速等優勢，但此方法需事先製作物體的訓練以獲得樣本參數，即可對新影像中進行特徵比對，從而達到在未知的環近中尋找物體。
執行平台使用NVIDIA JETSON AGX XAVIER作為本論文系統核心，並搭配機器人作業系統(ROS)；底盤輪子採用麥克納姆輪使機器人可做全向運動；深度攝影機採用Microsoft Xbox One Kinect V2使機器人擁有深度感知能力；並於正前方配置光學雷達(LIDAR)使用型號為YDLIDAR，協助機器人建構地圖以及室內定位；語音輸入則採用Logitech 羅技 H600 2.4G 無線耳機麥克風；移動機器人上擁有以6顆AI馬達所自製之六軸機械手臂，以利於能夠幫助使用者拿取物品。當使用者語音指令輸入且機器人成功辨識後，機器人自主導航與避障前往至使用者所指定之目的地，到達目的地後，自主辨識使用者所指定之物品，成功辨識後，朝該物體方向前進，直到機器人手臂可拿取之範圍，即啟動六軸機械手臂拿取該物體。

With the significant growing of AI technology, the importance of AI is increasing in the research field of robots. The advantage of the use of AI is obviously, such as understanding of naturallanguage, machine learning, artificial neural networks, machine vision, etc.
However, it is worth noting that although AI technology has been developed for years, its development is still in its early stages. For example, the family service robots with AI artificial intelligence are juststaring. Currently, the most of the commercially available family service robots focus on the use of companionship and entertainment. Without practical service functions, the development of family service robots is facing the market bottleneck today.
In order to improve this issue, our project develops a family service robot which is based on the AI technology to provide the practical service functions. With the features of image recognition, object picking operation, automatic speech recognition, object depth identification technology.

摘要...i
Abstract...ii
誌謝...iii
目錄...iv
表目錄...vii
圖目錄...viii
第一章 緒論...1
1.1 研究背景...1
1.2 研究動機與目的...1
1.3 研究回顧...2
1.3.1 即時定位與地圖建構(SLAM)...2
1.3.2 路徑規劃...3
1.4 問題描述...3
1.5 論文架構說明...3
第二章 文獻探討與分析...4
2.1 文獻探討一...4
2.2 文獻探討二...5
2.3 文獻分析...5
第三章 硬體設計與架構...6
3.1 硬體設計概述...6
3.2 機械結構設計...7
3.3 底盤結構設計...8
3.4 動力機構設計...10
3.5 機械手臂設計...11
3.6 硬體系統架構概述...13
3.7 上層系統OS...14
3.7.1 第一代機器人與第二代機器人核心...14
3.7.2 最終機器人系統核心...15
3.8 下層執行單元...16
3.8.1 環境偵測系統...16
3.8.2 影像辨識系統...18
3.8.3 底盤驅動單元...19
3.8.4 手臂驅動單元...19
3.8.5 電力系統...20
第四章 軟體系統架構...24
4.1 AI 識別物體能力...24
4.2 獲取物體深度資訊...25
4.3 自然語言處理...27
4.4 SLAM即時定位與地圖構建...28
4.5 Navigation導航與避障...29
4.6 移動機器人室內定位...30
第五章 實驗方法設計與實現...31
5.1 AI影像辨識須安裝套件與電腦規格...31
5.1.1 顯卡驅動與顯卡型號...31
5.1.2 Opencv與CUDA以及CUDNN版本型號...32
5.1.3 YOLO訓練之電腦規格...32
5.2 AI影像辨識安裝與測試...33
5.3 YOLO的訓練流程...35
5.3.1 準備數據集樣本...35
5.3.2 標記樣本...36
5.3.3 將XML轉換成TXT...38
5.3.4 數據集樣本分割...39
5.3.5 轉換成VOC文件...39
5.3.6 參數檔說明...40
5.4 YOLO的訓練...41
5.4.1 第一次訓練-無法辨識物體...41
5.4.2 第二次訓練-辨識距離過短...43
5.4.3 第三次訓練-辨識距離增加...45
5.4.4 第四次訓練-誤判尚未解決...47
5.4.5 第五次訓練-誤判解決...49
5.5 深度與手臂的轉換...51
5.5.1 深度與YOLO結合...52
5.5.2 三角演算法...53
5.6 PocketSphinx語音...54
5.7 SLAM...55
5.8 移動機器人室內定位...56
5.9 座標點標記...57
5.10 取物策略...58
第六章 實驗結果...62
6.1 SLAM建構地圖測試...62
6.2 Navigation導航與避障測試...65
6.3 物體檢測測試...70
6.4 機器人完整測試...72
第七章 結論與未來展望...75
7.1 結論...75
7.2 未來展望...76
參考文獻...77
Extended Abstract...79

1.Global Industrial Robot Data. Available from: https://www.digitimes.com.tw/iot/article.asp?cat=158&id=0000535766_EH78HXR253XOTK7H9FW66.
2.Global Service Robot Data. Available from: http://www.libnet.sh.cn:82/gate/big5/www.istis.sh.cn/list/list.aspx?id=11779.
3.Grisetti, G., C. Stachniss, and W. Burgard, Improved techniques for grid mapping with rao-blackwellized particle filters. IEEE transactions on Robotics, 2007. 23(1): p. 34.
4.Hart, P.E., N.J. Nilsson, and B. Raphael, A formal basis for the heuristic determination of minimum cost paths. IEEE transactions on Systems Science and Cybernetics, 1968. 4(2): p. 100-107.
5.Padgett, C. and K. Kreutz-Delgado, A grid algorithm for autonomous star identification. IEEE Transactions on Aerospace and Electronic systems, 1997. 33(1): p. 202-213.
6.Koenig, S. and M. Likhachev, Fast replanning for navigation in unknown terrain. IEEE Transactions on Robotics, 2005. 21(3): p. 354-363.
7.Khatib, O., Real-time obstacle avoidance for manipulators and mobile robots, in Autonomous robot vehicles. 1986, Springer. p. 396-404.
8.Fox, D., W. Burgard, and S. Thrun, The dynamic window approach to collision avoidance. IEEE Robotics & Automation Magazine, 1997. 4(1): p. 23-33.
9.ROS Melodic Package. Available from: http://wiki.ros.org/melodic.
10.Redmon, J., et al. You only look once: Unified, real-time object detection. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
11.蘇立珩, 基於機器學習與影像之機械手臂抓取, in 動力機械工程學系. 2018, 國立清華大學: 新竹市. p. 58.
12.林胤君, 機器視覺應用於六軸機器手臂抓取移動物體之實現, in 電機工程研究所. 2018, 高苑科技大學: 高雄市. p. 94.
13.Sauvola, J. and M. Pietikäinen, Adaptive document image binarization. Pattern recognition, 2000. 33(2): p. 225-236.
14.Madsen, A.L. and F.V. Jensen, Lazy propagation: a junction tree inference algorithm based on lazy evaluation. Artificial Intelligence, 1999. 113(1-2): p. 203-245.
15.Redmon, J. and A. Farhadi, Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767, 2018.
16.Huggins-Daines, D., et al. Pocketsphinx: A free, real-time continuous speech recognition system for hand-held devices. in 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings. 2006. IEEE.
17.Gao, Y., B.M.L. Srivastava, and J. Salsman. Spoken English Intelligibility Remediation with PocketSphinx Alignment and Feature Extraction Improves Substantially over the State of the Art. in 2018 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC). 2018. IEEE.
18.Taymans, W., et al., Gstreamer application development manual (1.2. 3). Publicado en la Web, 2013.
19.Varga, A.P. and R. Moore. Hidden Markov model decomposition of speech and noise. in International Conference on Acoustics, Speech, and Signal Processing. 1990. IEEE.
20.Eddy, S.R., Profile hidden Markov models. Bioinformatics (Oxford, England), 1998. 14(9): p. 755-763.
21.Miller, D.R., T. Leek, and R.M. Schwartz. A hidden Markov model information retrieval system. in SIGIR. 1999.
22.PocketSphin連續語音辨識框架. Available from: https://blog.csdn.net/zouxy09/article/details/7941585.
23.gmapping Package. Available from: http://wiki.ros.org/gmapping.
24.Lee, H.-C., et al., Probabilistic map merging for multi-robot RBPF-SLAM with unknown initial poses. Robotica, 2012. 30(2): p. 205-220.
25.Navigation Package. Available from: http://wiki.ros.org/navigation.
26.Lu, D.V. Navigation Illumination - Shedding Light on the ROS Navstack. Available from: https://roscon.ros.org/2014/wp-content/uploads/2014/07/ROSCON2014_DLu.pdf.
27.Monte Carlo localization. Available from. Available from: https://en.wikipedia.org/wiki/Monte_Carlo_localization.
28.Thrun, S., W. Burgard, and D. Fox, Probabilistic robotics. 2005: MIT press.
29.Darknet, CUDA and OpenCV Package install. Available from: https://blog.csdn.net/u014797226/article/details/78429936.
30.Darknet Package with Joseph Chet Redmon Available from: https://pjreddie.com/darknet/yolo/.
31.Darknet Package with AlexeyAB. Available from: https://github.com/AlexeyAB/darknet.