臺灣博碩士論文加值系統

本論文主要是完成人形足球機器人雙人傳球挑戰賽所使用的視覺與
策略系統。要完成整個比賽流程包含了很多方面研究題材的整合，包含
了運動控制、影像處理、定位系統以及機器人之間合作溝通的策略。本
論文中主要針對視覺以及策略系統作深入的探討。影像串流的資料是透
過一個CMOS 感測器經由USB 傳輸至嵌入式主機板Pico-820，使用主機
板上的CPU 來做運算。針對足球場地上的各項特徵物提出了辨識的方
法，並且使用了一個快速的影像分割法來使得物體辨識更加有效率，最
後再提出一個簡化的均值移動(Mean Shift)演算法，把整個視覺系統的計
算量做進一步的減少以提升效能。所有的策略也是由Pico-820 來做運算，
並且針對兩台機器人所扮演的不同角色設計了不同的控制策略。雙人傳
球的溝通係使用無線網路傳輸訊號，整體的通訊架構以及訊號的內容可
以讓兩台機器人的溝通正確無誤。最後實際實驗結果展現所提視覺與策
略系統，可成功地完成人形足球機器人雙人傳球功能。

The goal of this thesis is to achieve a technical challenge of double passing soccer game for humanoid soccer robots in RoboCup competition. There are many topics involved in order to finish the entire procedure, such as image processing algorithms, motion control, localization, and a proper scheme of strategy system. The development of vision and strategy system is mainly concerned in this thesis. The image information is captured by a CMOS sensor and then the image data are processed by an embedded CPU board Pico-820. Vision system works on the tasks of object recognition, which includes the goal, landmark poles and the interval of two black poles. A fast object segmentation method in the image with the need of only one execution loop is used to make the recognition more efficient. After recognitions, a simple mean shift algorithm is adopted to further improve the efficiency and performance of vision system. The computational time is reduced greatly by the algorithm and that time can be utilized to do the localization or control strategies. The strategies of double passing of each robot player are illustrated in flowcharts and the communication between them is via wireless network. The strategies of some more detailed situation are also discussed. All the strategies are also processed in Pico-820. Overall functions of vision and strategy system are programmed in C++. Finally, the experiment results demonstrate that the proposed vision and control strategy system can successfully perform the double passing soccer game.

Abstract Ⅰ
Acknowledgment Ⅲ
Contents Ⅳ
List of Figures Ⅶ
List of Tables XI
Chapter 1. Introduction 1
1.1 Motivation 1
1.2 Thesis Organization 3
Chapter 2. Hardware of the Humanoid Robot 5
2.1 Introduction 5
2.2 Mechanism Design 7
2.3 Vision and Control Strategy System 8
2.3.1 Image Sensor 8
2.3.2 Central Processing Unit 10
2.4 Actuators and Motion Controller 13
2.4.1 Actuators 13
2.4.2 Motion Controller 16
2.5 Power System 17

2.6 Summary 19
Chapter 3. Vision System 20
3.1 Introduction 20
3.2 A Brief Review of Basic Functions 21
3.3 Arrangement of Motors of Pan and Tilt 25
3.4 Object Recognition 27
3.4.1 Goal Recognition 27
3.4.2 Pole Recognition 33
3.4.3 Interval Detection 36
3.5 Simple Mean Shift Algorithm 38
3.5.1 Principal Concept of Mean Shift 38
3.5.2 Simplified Mean Shift Algorithm 40
3.5.3 The Result of Simplified Mean Shift Algorithm 42
3.6 Summary 44
Chapter 4. Control Strategy 45
4.1 Introduction 45
4.2 Rules of Double Passing Challenge 47
4.3 The Control Strategy for Common Behaviors 48
4.3.1 Control Strategies of Player A and B 48
4.3.2 The Processing Method of Falling down 51
4.3.3 The Communication Structure of Double Passing 52
4.3.4 Special problem of Double Passing 55
4.4 Summary 56
Chapter 5. Experimental Results 57
5.1 Introduction 57
5.2 Experimental comparisons of simple mean shift algorithm 58
5.3 Experimental Results of Double Passing Strategy 61
Chapter 6. Conclusions and Future Works 66
6.1 Conclusions 66
6.2 Future Works 68

References 69
Biography 71

[1]Y. T. Su, C. Y. Hu, and T. H. S. Li, “FPGA-based fuzzy PK controller and image processing system for small-sized humanoid robot, in Proc. IEEE International Conference on Systems, Man and Cybernetics, SMC, 2009, pp. 1039-1044.
[2]Y. T. Su, Development and implementation of visual and control systems for humanoid robot, Master Thesis, Dept. of E.E., N.C.K.U., Taiwan, June 2006.
[3]C. M. Chang, Design and implementation of vision and strategy system for humanoid robot soccer competition, Master Thesis, Dept. of E.E., N.C.K.U., Taiwan, July 2009.
[4]Y. Sumi, Y. Ishiyama and F. Tomita, “Robot-vision architecture for real-time 6-DOF object localization, Computer Vision and Image Understanding, vol. 105, pp. 218-230, 2007.
[5]C. C. Cheah, D. Q. Wang, and Y. C. Sun, “Region-Reaching Control of Robots, IEEE Trans. Robotics, vol. 23, pp. 1260-1264, 2007.
[6]Robocup 2010, http://www.robocup2010.org/
[7]FIRA, http://www.fira.net/
[8]H. P. Ng, S.H. Ong, K. W. C. Foong, P. S. Goh and W. L. Nowinski, “Medical image segmentation using k-means clustering and improved watershed algorithm, in Proc. IEEE Southwest Symposium on Image Analysis and Interpretation, 2006, pp. 61-65.
[9]P. Yongsheng, J. D. Birdwell and S. M. Djouadi, “Preferential image segmentation using trees of shapes, IEEE Trans. Image Processing, vol. 18, pp. 854-866, 2009.
[10]A. Bugeau and P. Perez, “Detection and segmentation of moving objects in complex scenes, Computer Vision and Image Understanding, vol. 113, pp. 458-476, 2009.
[11]Z. Shiping, X. Xi, Z. Qingrong and K. Belloulata, “An image segmentation Algorithm in image processing based on threshold segmentation, in Proc. Third International IEEE Conference on Signal-Image Technologies and Internet-Based System, SITIS, 2007, pp. 673-678.
[12]O. M. Alia, R. Mandava, D. Ramachandram and M. E. Aziz, “A novel image segmentation algorithm based on harmony fuzzy search algorithm, in Proc. International Conference of Soft Computing and Pattern Recognition, SOCPAR, 2009, pp. 335-340.
[13]N. Alajlan, M. Kamel and E. Jernigan, “Detail preserving impulsive noise removal, Signal Processing: Image Communication, vol. 19, pp. 993-1003, 2004.
[14]X. Cai, “Reduction of speckle noise in the reconstructed image of digital holography, Optik-International Journal for Light and Electron Optics, vol. 121, pp. 394-399, 2010.
[15]S. Yuan and Y. Tan, “Impulse noise removal by a global-local noise detector and adaptive median filter, Signal Processing, vol. 86, pp. 2123-2128, 2006.
[16]G. Sen Gupta and D. Bailey, “Discrete YUV look-up tables for fast colour segmentation for robotic applications, in Canadian Conference on Electrical and Computer Engineering, CCECE, 2008, pp. 000963-000968.
[17]A. N. Bishop, A. V. Savkin and P. N. Pathirana, “Vision-based target tracking and surveillance with robust set-valued state estimation, IEEE Signal Processing Letters, vol. 17, pp. 289-292, 2010.
[18]Q. Wei and D. Schonfeld, “Robust control-based object tracking, IEEE Trans. Image Processing, vol. 17, pp. 1721-1726, 2008.
[19]A. Marin-Hernandez, V. Ayala-Ramirez and M. Devy, “Collaborative tracking for robotic tasks, in Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS, 2003, pp. 272-277.
[20]N. D. Binh, “A robust framework for visual object tracking, in Proc. International Conference on Computing and Communication Technologies, RIVF, 2009, pp. 1-8.
[21]A. Pooransingh, C. A. Radix and A. Kokaram, “The path assigned mean shift algorithm: A new fast mean shift implementation for colour image segmentation, in Proc. 15th IEEE International Conference on Image Processing, ICIP, 2008, pp. 597-600.
[22]H. Wang, B. Yang, G. Tian and A. Men, “Object tracking by applying mean-shift algorithm into particle filtering, in Proc. 2nd IEEE International Conference on Broadband Network ＆ Multimedia Technology, IC-BNMT, 2009, pp. 550-554.
[23]Y. Yafeng and M. Hong, “Adaptive mean shift for target-tracking in FLIR imagery, in Proc. Wireless and Optical Communications Conference, WOCC, 2009, pp. 1-3.
[24]Z. Q. Wen and Z. X. Cai, “Mean shift algorithm and its application in tracking of objects, in Proc. International Conference on Machine Learning and Cybernetics, 2006, pp. 4024-4028.
[25]Y. Liu and S. Peng, “A new motion detection algorithm based on snake and mean shift, in Proc. Congress on Image and Signal Processing, CISP, 2008, pp. 140-144.
[26]The rules of RoboCup 2010 Humanoid League Soccer,
http://www.tzi.de/humanoid/bin/view/Website/Downloads/