臺灣博碩士論文加值系統

機器人自我定位是個具有挑戰性的話題，尤其需要加入人工智慧讓機器人更像人類，讓機器人能夠自己思考，了解自己的位置是不可或缺的，而定位法有非常多，在本篇論文應用了強健性高能夠擁有抗雜訊的蒙地卡羅定位法，應用在本實驗室的仿人形機器人，主要參加RoboCup與FIRA 的機器人競賽中，讓機器人能夠自我定位了解位於場地上的位置。
蒙地卡羅定位法中分為感測模型及運動模型，感測模型主要對四周的環境了解資訊，透過的是簡易型的網路攝影機，攝影機畫面角度的受限及影像空間與世界空間的轉換，應用逆透視來解決此問題，並且模擬出傳統感測模型需要昂貴的雷射掃瞄儀的掃描線，降低成本上的需求，來達到感測模型所需要的資訊需求。感測模型的更新及運動模型的預測，兩者互相修正，最後迭代出機器人最有可能的位置，而達到本論文的目標機器人自我定位的需求，最後分析其定位的位置座標資料，系統模擬與實際行走上的誤差，來探討分析數據的標準差是否為能合理接受範圍。

Localization is one of the most important challenges for a mobile robot. It is the fundamental of Artificial Intelligence (A.I), so the robot more like humans can think by itself, based on understanding of their position is necessary. There are a lot of researchers developing new methods for localization. In this paper the application of high robustness and against noise, method is Monte Carlo Localization is proposed. We design a self-localization robot for FIRA and RoboCup competition.
Monte Carlo localization method has sensor model and motion model. Sensor model searches the information on the environment, through a simple type camera. Especially the Inverse Perspective Mapping solves the problem of 3-D information with single camera, through the motion model and the sensor model of the MCL algorithm, the robot system can localize in the environment by updating its position weight recursively. The experiment includes the localization position influenced by error are also influenced.

誌謝
中文摘要
ABSTRACT
目錄
圖目錄
表目錄
1. 緒論
1.1 相關背景
1.2 研究動機
1.3 章節摘要
2. 問題描述
3. 研究方法
3.1 蒙地卡羅定位法
3.1.1 簡介
3.1.2 蒙地卡羅演算法流程
3.1.3 重新取樣( Resampling )
4. 定位演算法建立與設計
4.1 運動模型MOTION MODEL
4.1.1 運動方程式
4.2 感測模型SENSOR MODEL
4.2.1 逆透視(Inverse Perspective Mapping ;IPM)
4.2.2 IPM演算法
4.2.3 環境感測機制
4.2.4 混和機率密度
5. 軟硬體系統架構
5.1 仿人形機器人系統
5.2 影像視覺系統
5.3 影像校正
6. 實驗模擬與結果
6.1 模擬
6.2 實驗與分析
7. 結論與未來展望
8. 參考文獻

[1]AIBO Robot, 2011, http://support.sony-europe.com/aibo/
[2]ASIMO Robot, 2011, http://asimo.honda.com/
[3]RoboCup Official Site, 2011, http://www.robocup.org
[4]FIRA Site, 2011, http://www.fira.net/
[5]Dieter Fox, Wolfram Burgard, Sebastian Thrun, “Markov Localization for Mobile Robot in Dynamic Environments”, Artificial Intelligence,1999, pp. 1391-1427.
[6]Frank Dellaert, Dieter Fox, Wolfram Burgard, and Sebastian Thrun, “Monte Carlo localization for mobile robots”, IEEE International Conference, Robotics and Automation Vol. 2, May 1999, pp. 1322–1328.
[7]Sebastian Thrun, Wolfram Burgard , Dieter Fox, Probabilistic Robotics, MIT Press ,Cambridge, 2005.
[8]M. Bertozzi and A. Broggi, “GOLD: a parallel real-time stereovision system for generic obstacle and lane detection”, IEEE Transactions on Image Processing, vol. 7, no. 1, 1998, pp. 62–81.
[9]M. Bertozzi, A. Broggi, and A. Fascioli, “Stereo inverse perspective mapping: theory and applications”, Image and Vision Computing, vol. 16, no. 8, 1998, pp. 585-590.
[10]Chin-Teng Lin, Tzu-Kuei Shen, and Yu-Wen Shou, “Construction of Fisheye Lens Inverse PerspectiveMappingModel and Its Applications of Obstacle Detection”, EURASIP Journal on Advances in Signal Processing Volume 2010, Article ID 296598, pp. 23
[11]李浩銘，基於全方向視覺系統Monte Carlo定位的研究，國立高雄第一科技大學 系統資訊與控制研究所 碩士論文
[12]何丞堯，全方位視覺足球機器人之自我定位系統的設計與實現，淡江大學電機工程學系機器人工程碩士班 碩士論文
[13]Huimin Lu , Shaowu Yang, Hui Zhang, Zhiqiang Zheng, “A robust omnidirectional vision sensor for soccer robots”, Mechatronics, 2010 doi:10.1016/j.mechatronics.
[14]Sven Olufs and Markus Vincze, “An Efficient Area-based Observation Modelfor Monte-Carlo Robot Localization”, IEEE/RSJ International Conference ,Intelligent Robots and Systems, 2009
[15]Mohan Sridharan and Gregory Kuhlmann and Peter Stone, “Practical Vision-Based Monte Carlo Localization on a Legged Robot, Department of Computer Sciences”, University of Texas , Austin.