Robocup类人机器人系统研究与实现
|
摘要
随着第一台电子计算机ENIAC在美国宾夕法尼亚大学诞生开始,人类便步入了信息时代；同时,代表着信息时代最高水准的机器人技术也开始孕育。机器人技术囊括了包括材料科学、机械结构、微电子技术、软件工程、传感器技术在内的众多学科,引起了世界各国的关注。在这个领域,日本、德国等工业强国走在了世界前列,无论在基础科学还是实物制造都取得了重大进展；我国由于起步较晚,加之工业化水平还无法同日本、德国等工业强国相比,进展较为缓慢。为推动机器人相关技术发展及学术交流,国际人工智能协会创立了Robocup (The World Robot Cup),即机器人足球世界杯,经过十几年的发展,现已经成同ACM/ICPC并列的世界两大赛事之一
本文Robocup类人组比赛为平台,对视觉系统及运动控制系统进行了讨论,并在实物Fantasia进行了验证。
视觉系统主要分析了Agent视觉及视觉调试原理及应用实现。分析了颜色空间、图像分割等基本问题,提出了基于RLE的改进聚类算法；在PC/104平台上使用Visual Studio 2008进行了实现,主要分为图像采集、图像分割、物体识别、特征颜色提取及存储等功能。
运动控制系统主要分析了机器人运动学模型的基础数学问题、稳定性判据,并提出了基于关键帧相似性的步态规划方法；为加快步态调试速度,开发了离线步态调试软件,并讨论了步态存储结构问题。
最后分析了Agent设计原理及相关模块实现,主要有网络通信、RS232通信、有限状态机等模块,最后阐述了Agent相关寻球及进攻策略。
经过在Robocup 2009中国公开赛上的验证表明,本文提出的基于RLE的改进聚类算法及基于关键帧相似性的步态规划方法,运行高效、稳定,并在比赛中取得了优异成绩。
As the first electronic computer ENIAC was born at the University of Pennsylvania, humans have entered the information age; Meanwhile, robot technology have begun to nurture, it represents the highest level of the information age. Robot technology including a number of disciplines, including materials science, mechanical structures, micro-electronics technology, software engineering, sensor technology, it has attracted world's attention. In this area, Japan, Germany and other industrial powers to go in the forefront of the world, Whether real robot manufactured or basic sciences have made significant progress; In China, with the late start, combined with the level of industrialization cannot compete with Japan, Germany and other industrial powers, made progress more slowly. To promote the development of robot-related technology and academic exchanges, International Association for Artificial Intelligence founded the Robocup (The World Robot Cup), after ten years of development, now has become with the ACM/ICPC alongside one of the world's two races.
This article makes Robocup humanoid competitions as a platform, the machine vision system and motion control system are discussed and verified in physical humanoid robot Fantasia.
The machine vision system analyzes agent vision system and vision debug system's theory and application implementation. Discuss color space, image segmentation and other basic issues, Improvements proposed clustering algorithm based on RLE; implement application with Visual Studio 2008 under PC/104 platform, Consists image acquisition, image segmentation, object recognition, feature color extraction and storage and other functions.
The motion control system kinematic model analyzes the mathematical basis and stability criterion of the problem, proposed stable gait plan based on key frame similarity algorithm. To speed up the debugging gait speed, off-line gait debugging software was developed, and discussed the storage structure of gait.
At the last, analyzes the design principles and related modules of the agent application; include network communication, RS232 communication, finite state machine module and the find ball and attack strategy.
In Robocup 2009 China Open, the verification shows that, the clustering algorithm based on RLE and stable gait plan based on key frame similarity algorithm running efficient, stable, and in the race has made outstanding achievements.
本文Robocup类人组比赛为平台,对视觉系统及运动控制系统进行了讨论,并在实物Fantasia进行了验证。
视觉系统主要分析了Agent视觉及视觉调试原理及应用实现。分析了颜色空间、图像分割等基本问题,提出了基于RLE的改进聚类算法；在PC/104平台上使用Visual Studio 2008进行了实现,主要分为图像采集、图像分割、物体识别、特征颜色提取及存储等功能。
运动控制系统主要分析了机器人运动学模型的基础数学问题、稳定性判据,并提出了基于关键帧相似性的步态规划方法；为加快步态调试速度,开发了离线步态调试软件,并讨论了步态存储结构问题。
最后分析了Agent设计原理及相关模块实现,主要有网络通信、RS232通信、有限状态机等模块,最后阐述了Agent相关寻球及进攻策略。
经过在Robocup 2009中国公开赛上的验证表明,本文提出的基于RLE的改进聚类算法及基于关键帧相似性的步态规划方法,运行高效、稳定,并在比赛中取得了优异成绩。
As the first electronic computer ENIAC was born at the University of Pennsylvania, humans have entered the information age; Meanwhile, robot technology have begun to nurture, it represents the highest level of the information age. Robot technology including a number of disciplines, including materials science, mechanical structures, micro-electronics technology, software engineering, sensor technology, it has attracted world's attention. In this area, Japan, Germany and other industrial powers to go in the forefront of the world, Whether real robot manufactured or basic sciences have made significant progress; In China, with the late start, combined with the level of industrialization cannot compete with Japan, Germany and other industrial powers, made progress more slowly. To promote the development of robot-related technology and academic exchanges, International Association for Artificial Intelligence founded the Robocup (The World Robot Cup), after ten years of development, now has become with the ACM/ICPC alongside one of the world's two races.
This article makes Robocup humanoid competitions as a platform, the machine vision system and motion control system are discussed and verified in physical humanoid robot Fantasia.
The machine vision system analyzes agent vision system and vision debug system's theory and application implementation. Discuss color space, image segmentation and other basic issues, Improvements proposed clustering algorithm based on RLE; implement application with Visual Studio 2008 under PC/104 platform, Consists image acquisition, image segmentation, object recognition, feature color extraction and storage and other functions.
The motion control system kinematic model analyzes the mathematical basis and stability criterion of the problem, proposed stable gait plan based on key frame similarity algorithm. To speed up the debugging gait speed, off-line gait debugging software was developed, and discussed the storage structure of gait.
At the last, analyzes the design principles and related modules of the agent application; include network communication, RS232 communication, finite state machine module and the find ball and attack strategy.
In Robocup 2009 China Open, the verification shows that, the clustering algorithm based on RLE and stable gait plan based on key frame similarity algorithm running efficient, stable, and in the race has made outstanding achievements.
引文
[1]孙付春.步行机器人的行走控制[D].成都：成都理工大学测试计量学院,2006.
[2]蔡自兴.机器人学[M].北京：清华大学出版社,2000
[3]秦爱中.基于ZMP的双足步行机器人步态规划研究[D].西安：西北工业大学机械电子工程,2005
[4]J. Yamaguchi, A. Takanishi, I. Kato. Development a Biped Robot Walking Robot Compensation for Three-axis Moment by Trunk Motion[C]. IEEE International Conference on Robot and System,1993, Nagoya:561-566
[5]A. Takanishi, M. Ishida, YYamazaki and I. Kato. The Realization of Dynamic Walking by Biped Robot WL-10RD. IEEE International Conference on Autonomous Robot,1985, Tokyo:459-467
[6]Li Q, A. Takanishi and I. Kato. Learning Control of Compensation Trunk Motion for Biped Walking Robot Based on ZMP Stability Criterion. IEEE Int. Conf. on Robotics System,1992, Yokohama:597-603
[7]Kazuo Hirai, R&D Co. Ltd Wako Research Center, Current and Future Perspective of Honda of Dynamic Biped Locomotion Journal of Dynamic Systems [J], Measurement and Control, June 1987,Vol 109-155
[8]K. Hirai, M. Hirose and T. Takenaka. The Development of Honda Humanoid Robot [C]. IEEE International Conference on Robotics and Automation,1998, Leuven:1321-1326
[9]http://www.peopledaily.co.jp/GB/it/1070/2258457.html
[10]Y. F. Zheng, A Neural Gait Synthesizer for Autonomous Robot[C]. IEEE International Conference on Robotics and Automation,1990, Seattle:601-608
[11]Y.F.Zheng, Acceleration Compensation for Biped Robots to Reject External Disturbance[J], IEEE Trans on System, man and Cybernetics,1989,19(1),74-84
[12]Y. F. Zheng, Design and Motion Control of Practical Biped Robots [J], International J. of Robotics and Automation,1984,3(2):70-78
[13]王淑艳.仿人型跑步机器人的稳定性条件及步态规划[D].西安：西安电子科技大学机械制造及自动化,2006
[14]张博.小型双足人形机器人机构设计与步态规划[D].哈尔滨：哈尔滨工业大学机械电子工程,2006
[15]崔广斌.机器人足球仿真比赛中的步态规划研究[D].大连：大连理工大学计算机系,2008
[16]高亮.类人机器人视觉系统研究实现[D].大连：大连理工大学计算机系,2009
[17]http://baike.baidu.com/view/3427413.htm
[18]冯永超,罗敏,贺贵明.一种快速YUV-RGB彩色空间变换方法[J].微型机与应用.2002(7)：58-59.
[19]Keith Jack. Video Demystified[M]. Butterworth-Heinemann,2005
[20]http://baike.baidu.com/view/541362.html
[21]Tai, X., E. Hodneland, et al. Level set methods for watershed image segmentation[J]. Lecture notes in computer science (2007),4485:178
[22]Luc Vincent, Pierre Soille. Watersheds in Digital Spaces:An Efficient Algorithm Based on Immersion Simulations[J]. IEEE Trans On Pattern Analysis and Machine Intelligence,1991, Vol.13, No.6,583-598
[23]Fahim, A., Salem, et al. An efficient enhanced k-means clustering algorithm [J]. Journal of Zhejiang University-Science.2006.7(10):1626-1633
[24]陆其明.DirectShow开发指南[M].北京：清华大学出版社,2003.
[25]Golomb, S. Run-length encodings(Run-length encodings, determining explicit form of Huffman coding when applied to geometric distribution)[J]. IEEE Transactions on Information Theory,1966.12:399-401
[26]王艳平.Windows网络与通讯程序设计[M].北京：人民邮电大学出版社,2009.
[27]Erich Gamma, Richard Helm, et al. Design Patterns:Elements of Reusable Object-Oriented Software[M].1994.
[28]RoboCup Soccer Humanoid League Rules and Setup for the 2009 competition in Xian, China
[29]徐利华,陈早生.二值图像中的游程编码区域标记[J].光电工程,2004,31(6)：63-65.
[30]ROSENFELD A, PFALTZ J L. Sequential Operations in Digital Picture Processing[J]. Journal of the ACM(S0004-541),1966,13(4):471-494.
[31]ROSENFELD A, KAK A C. Digital Picture Processing [M]. New York:Academic Press,1982:241-242.
[32]LUMIA R, SHAPIRO L, ZUNIGA O. A New Connected Components Algorithm for Virtual Memory Computers [J]. Computer Vision, Graphics, and Image Processing,1983,22(2):287-300.
[33]胡涛,郭宝平,郭轩.基于改进游程连通分析的快速目标描述算法[J].光电工程.2009,36(2)：116-120
[2]蔡自兴.机器人学[M].北京：清华大学出版社,2000
[3]秦爱中.基于ZMP的双足步行机器人步态规划研究[D].西安：西北工业大学机械电子工程,2005
[4]J. Yamaguchi, A. Takanishi, I. Kato. Development a Biped Robot Walking Robot Compensation for Three-axis Moment by Trunk Motion[C]. IEEE International Conference on Robot and System,1993, Nagoya:561-566
[5]A. Takanishi, M. Ishida, YYamazaki and I. Kato. The Realization of Dynamic Walking by Biped Robot WL-10RD. IEEE International Conference on Autonomous Robot,1985, Tokyo:459-467
[6]Li Q, A. Takanishi and I. Kato. Learning Control of Compensation Trunk Motion for Biped Walking Robot Based on ZMP Stability Criterion. IEEE Int. Conf. on Robotics System,1992, Yokohama:597-603
[7]Kazuo Hirai, R&D Co. Ltd Wako Research Center, Current and Future Perspective of Honda of Dynamic Biped Locomotion Journal of Dynamic Systems [J], Measurement and Control, June 1987,Vol 109-155
[8]K. Hirai, M. Hirose and T. Takenaka. The Development of Honda Humanoid Robot [C]. IEEE International Conference on Robotics and Automation,1998, Leuven:1321-1326
[9]http://www.peopledaily.co.jp/GB/it/1070/2258457.html
[10]Y. F. Zheng, A Neural Gait Synthesizer for Autonomous Robot[C]. IEEE International Conference on Robotics and Automation,1990, Seattle:601-608
[11]Y.F.Zheng, Acceleration Compensation for Biped Robots to Reject External Disturbance[J], IEEE Trans on System, man and Cybernetics,1989,19(1),74-84
[12]Y. F. Zheng, Design and Motion Control of Practical Biped Robots [J], International J. of Robotics and Automation,1984,3(2):70-78
[13]王淑艳.仿人型跑步机器人的稳定性条件及步态规划[D].西安：西安电子科技大学机械制造及自动化,2006
[14]张博.小型双足人形机器人机构设计与步态规划[D].哈尔滨：哈尔滨工业大学机械电子工程,2006
[15]崔广斌.机器人足球仿真比赛中的步态规划研究[D].大连：大连理工大学计算机系,2008
[16]高亮.类人机器人视觉系统研究实现[D].大连：大连理工大学计算机系,2009
[17]http://baike.baidu.com/view/3427413.htm
[18]冯永超,罗敏,贺贵明.一种快速YUV-RGB彩色空间变换方法[J].微型机与应用.2002(7)：58-59.
[19]Keith Jack. Video Demystified[M]. Butterworth-Heinemann,2005
[20]http://baike.baidu.com/view/541362.html
[21]Tai, X., E. Hodneland, et al. Level set methods for watershed image segmentation[J]. Lecture notes in computer science (2007),4485:178
[22]Luc Vincent, Pierre Soille. Watersheds in Digital Spaces:An Efficient Algorithm Based on Immersion Simulations[J]. IEEE Trans On Pattern Analysis and Machine Intelligence,1991, Vol.13, No.6,583-598
[23]Fahim, A., Salem, et al. An efficient enhanced k-means clustering algorithm [J]. Journal of Zhejiang University-Science.2006.7(10):1626-1633
[24]陆其明.DirectShow开发指南[M].北京：清华大学出版社,2003.
[25]Golomb, S. Run-length encodings(Run-length encodings, determining explicit form of Huffman coding when applied to geometric distribution)[J]. IEEE Transactions on Information Theory,1966.12:399-401
[26]王艳平.Windows网络与通讯程序设计[M].北京：人民邮电大学出版社,2009.
[27]Erich Gamma, Richard Helm, et al. Design Patterns:Elements of Reusable Object-Oriented Software[M].1994.
[28]RoboCup Soccer Humanoid League Rules and Setup for the 2009 competition in Xian, China
[29]徐利华,陈早生.二值图像中的游程编码区域标记[J].光电工程,2004,31(6)：63-65.
[30]ROSENFELD A, PFALTZ J L. Sequential Operations in Digital Picture Processing[J]. Journal of the ACM(S0004-541),1966,13(4):471-494.
[31]ROSENFELD A, KAK A C. Digital Picture Processing [M]. New York:Academic Press,1982:241-242.
[32]LUMIA R, SHAPIRO L, ZUNIGA O. A New Connected Components Algorithm for Virtual Memory Computers [J]. Computer Vision, Graphics, and Image Processing,1983,22(2):287-300.
[33]胡涛,郭宝平,郭轩.基于改进游程连通分析的快速目标描述算法[J].光电工程.2009,36(2)：116-120