智能履带车视觉导航技术研究
|
摘要
履带车辆依其良好的通过性能在工程建筑、军事、农业等领域发挥着十分重要的作用。随着传感技术、计算机科学、人工智能及其它相关学科的迅速发展，履带车正向着智能化方向发展。而履带车智能化水平的一个重要体现就是其导航能力的高低。
研究履带车的导航技术具有重要的意义。履带车工作环境一般比较恶劣，而且还经常出没于一些危险的环境，另外操作者在驾驶履带车长期工作时也会很疲劳，因此履带车的自主导航，可以将人从繁重无聊的工作中解放出来。
履带车常见的导航方式有电磁导航、超声波导航和视觉导航等。视觉传感器由于获取的信息量大，敏感度高，成本低，并且视觉导航可根据需要灵活地改变或扩充路径，因而视觉导航成为履带车导航研究得较多的方向。
视觉导航主要有两个研究方向：（1）完全意义上的视觉导航：它是通过模拟人的视觉来识别道路。（2）另一研究方向是标识线图像识别法，即“有线式”视觉导航。这种导航方法简单灵活，与完全意义上的视觉导航相比，具有更快的图像处理速度和更好的控制实时性。本文所研究的即为智能履带车“有线式”视觉导航。
本文以课题组自主开发研制的实验用智能履带车为实验平台，对履带车的视觉导航技术进行了研究，研究内容主要为：
实验用智能履带车结构及驱动系统设计
为了研究履带车的自主导航技术，课题组制作了实验用智能履带车。本文设计了其组织和控制结构，包括感知模块、规划模块和执行模块。
描述了履带车的转向过程，并介绍了实验用智能履带车通过工控机控制两台变频器，两台变频器分别控制两侧履带驱动电机，从而实现履带车转向的控制原理。
智能履带车的视觉系统
介绍了实验用智能履带车视觉系统构成以及各个模块的功能。分别为图像
采集模块用来获取路面数字图像；图像预处理模块是为路标线提取做准备的；路标线识别的任务是识别出环境中预先设定的路标线，并由此得到用于决策的信息，即位置偏差和方向偏差。
采用一种快速摄像机标定方法对实验用智能履带车的摄像机进行了标定，获得了摄像机内外参数，从而确定了视觉系统各个坐标系之间的映射关系。
对所采集的路面图像进行了预处理，包括图像灰度化，采用最大方差阈值法选择阈值进行的图像二值化，采用改进的领域平均法对图像进行噪声消除。通过对图像划分网格大大压缩了图像的数据量，提高了图像处理速度。采用逐行扫描求出路标线的位置，并用最小二乘法拟合路标线求得履带车模糊控制器的两个输入量：位置偏差和方向偏差。
智能履带车的决策系统
由于履带车视觉导航系统难以用精确的数学模型来描述，故在介绍了模糊控制的基本原理的基础上，为智能履带车视觉导航系统设计了模糊控制器。
图1 路标示意图
履带车视觉导航采用的是一个双输入单输出模糊控制器。两个输入为位置偏差B和方向偏差α（如图1），输出量为左右侧变频器控制电压之差系数ΔV*
采用加权平均法对控制量进行反模糊化：
假设反模糊化得到时刻的控制量，则求得两侧变频器的控制电压分别为：
，
这里，为转向前两侧变频器的控制电压。
智能履带车视觉导航实验研究
介绍了实验用智能履带车视觉导航系统，主要是由视觉系统、决策系统、驱动系统三部分组成。
选用Windows操作系统和Visual C++语言开发了智能履带车视觉导航系统软件。该软件可以使操作者方便地操作履带车，并能随时掌握履带车的运行状况。介绍了软件对图像采集的操作和通过PCL-728控制卡对两个变频器的操作。
利用自己设计研制的实验用智能履带车在实验室进行了大量的实验验证了导航系统的有效性。在介绍了履带车在做直线导航和曲线导航结果的基础上，分析了导航系统影响因素，并提出了几点改进措施。
结论
本文所介绍的智能履带车视觉导航系统较好的实现了视觉导航，但需要进一步的研究，以提高导航精度和可靠性。
Tracklayer plays an important role in project architecture, military affairs, agriculture etc，which can be attributed to its powerful overpassing ability. With the rapid development of sensor technology, computer science, artificial intelligence and other related subject, tracklayer is becoming more and more intelligentized. one of the vital embodiments of tracklayers’ intelligentized level is the ability of navigation.
Doing research on tracklayer’s navigation is of great significance. Tracklayer is mostly operated under unfavourable or even dangerous circumstances. Further more the operator is likely to be exhausted after working continuously for a long time. With the tracklayer’s navigation, the operators can be liberated from the tiring and boring work.
There are many methods frequently used in tracklayer’s navigation, such as electromagnetism navigation, ultrasonic navigation, vision navigation, etc. Vision sensor is capable of gaining more information with high sensitivity and low cost. In addition, its flexibility in transforming or extending paths also helps it attract more attention in the research.
Research on vision navigation can be made in two direction: (1) Vision navigation which imitate human vision to identify the path. (2) Vision navigation by following the initially fixed signs. The latter method is simpler and more flexible, compared to the former one, since it has faster image processing speed and better real time property. The research on my dissertation is focusing on the latter one, that is, intelligent tracklayer’s navigation by following sign.
In my dissertation, I did some research on the technology of intelligent tracklayer’s vision navigation based on the experimental intelligent tracklayer developed independently by our research group. The main contents of my dissertation are as follow:
1. Design the frame and drive system of the experimental intelligent tracklayer
Our research group developed an experimental intelligent tracklayer for the purpose of doing research on the intelligent tracklayer’s vision navigation.
The organizing and control framework of the experimental intelligent tracklayer is devised in my dissertation, that is the apperceiving module, the programming module and the enforcing module.
The process of tracklayer’s turning is described. An introduction concerning of the control principle of how to realize this progress is also made, that is the industry computer controls the two transducers and then the two transducers separately control the two electromotors.
2. Research on intelligent tracklayer’s vision system
My dissertation introduces that the intelligent tracklayer’s vision system are composed of three modules: Image collection module which is responsible for getting the digital image of the road, preprocessing module which is used to make preparation for road sign line identification, and identification module which is able to identify the road sign line laid in the circumstance as well as obtain the position warp and direction warp which are used for decision-making.
A fast and accurate calibration method is adopted to calibrate the CCD camera of the intelligent tracklayer so as to acquire the intrinsic and extrinsic parameters from which the mapping connection of all the coordinates can be achieved thoroughly.
The collected road image is preprocessed. This includes image gray scale transformation, image binarization whose threshold selection is based on the maximum variance, noise elimination which adoptes ameliorated adjacent region average method. Through the partition gridding of the image I make some accomplishment on compressing the data quantity of the image and increase the image processing velocity. I adopt progressive to locate the road sign line, and obtain the two input variables of the fuzzy controller, position warp and direction warp.
3. Research on intelligent tracklayer’s decision-making system
Since it’s difficult to establish an accurate mathematics model for the intelligent tracklayer’s vision navigat
研究履带车的导航技术具有重要的意义。履带车工作环境一般比较恶劣，而且还经常出没于一些危险的环境，另外操作者在驾驶履带车长期工作时也会很疲劳，因此履带车的自主导航，可以将人从繁重无聊的工作中解放出来。
履带车常见的导航方式有电磁导航、超声波导航和视觉导航等。视觉传感器由于获取的信息量大，敏感度高，成本低，并且视觉导航可根据需要灵活地改变或扩充路径，因而视觉导航成为履带车导航研究得较多的方向。
视觉导航主要有两个研究方向：（1）完全意义上的视觉导航：它是通过模拟人的视觉来识别道路。（2）另一研究方向是标识线图像识别法，即“有线式”视觉导航。这种导航方法简单灵活，与完全意义上的视觉导航相比，具有更快的图像处理速度和更好的控制实时性。本文所研究的即为智能履带车“有线式”视觉导航。
本文以课题组自主开发研制的实验用智能履带车为实验平台，对履带车的视觉导航技术进行了研究，研究内容主要为：
实验用智能履带车结构及驱动系统设计
为了研究履带车的自主导航技术，课题组制作了实验用智能履带车。本文设计了其组织和控制结构，包括感知模块、规划模块和执行模块。
描述了履带车的转向过程，并介绍了实验用智能履带车通过工控机控制两台变频器，两台变频器分别控制两侧履带驱动电机，从而实现履带车转向的控制原理。
智能履带车的视觉系统
介绍了实验用智能履带车视觉系统构成以及各个模块的功能。分别为图像
采集模块用来获取路面数字图像；图像预处理模块是为路标线提取做准备的；路标线识别的任务是识别出环境中预先设定的路标线，并由此得到用于决策的信息，即位置偏差和方向偏差。
采用一种快速摄像机标定方法对实验用智能履带车的摄像机进行了标定，获得了摄像机内外参数，从而确定了视觉系统各个坐标系之间的映射关系。
对所采集的路面图像进行了预处理，包括图像灰度化，采用最大方差阈值法选择阈值进行的图像二值化，采用改进的领域平均法对图像进行噪声消除。通过对图像划分网格大大压缩了图像的数据量，提高了图像处理速度。采用逐行扫描求出路标线的位置，并用最小二乘法拟合路标线求得履带车模糊控制器的两个输入量：位置偏差和方向偏差。
智能履带车的决策系统
由于履带车视觉导航系统难以用精确的数学模型来描述，故在介绍了模糊控制的基本原理的基础上，为智能履带车视觉导航系统设计了模糊控制器。
图1 路标示意图
履带车视觉导航采用的是一个双输入单输出模糊控制器。两个输入为位置偏差B和方向偏差α（如图1），输出量为左右侧变频器控制电压之差系数ΔV*
采用加权平均法对控制量进行反模糊化：
假设反模糊化得到时刻的控制量，则求得两侧变频器的控制电压分别为：
，
这里，为转向前两侧变频器的控制电压。
智能履带车视觉导航实验研究
介绍了实验用智能履带车视觉导航系统，主要是由视觉系统、决策系统、驱动系统三部分组成。
选用Windows操作系统和Visual C++语言开发了智能履带车视觉导航系统软件。该软件可以使操作者方便地操作履带车，并能随时掌握履带车的运行状况。介绍了软件对图像采集的操作和通过PCL-728控制卡对两个变频器的操作。
利用自己设计研制的实验用智能履带车在实验室进行了大量的实验验证了导航系统的有效性。在介绍了履带车在做直线导航和曲线导航结果的基础上，分析了导航系统影响因素，并提出了几点改进措施。
结论
本文所介绍的智能履带车视觉导航系统较好的实现了视觉导航，但需要进一步的研究，以提高导航精度和可靠性。
Tracklayer plays an important role in project architecture, military affairs, agriculture etc，which can be attributed to its powerful overpassing ability. With the rapid development of sensor technology, computer science, artificial intelligence and other related subject, tracklayer is becoming more and more intelligentized. one of the vital embodiments of tracklayers’ intelligentized level is the ability of navigation.
Doing research on tracklayer’s navigation is of great significance. Tracklayer is mostly operated under unfavourable or even dangerous circumstances. Further more the operator is likely to be exhausted after working continuously for a long time. With the tracklayer’s navigation, the operators can be liberated from the tiring and boring work.
There are many methods frequently used in tracklayer’s navigation, such as electromagnetism navigation, ultrasonic navigation, vision navigation, etc. Vision sensor is capable of gaining more information with high sensitivity and low cost. In addition, its flexibility in transforming or extending paths also helps it attract more attention in the research.
Research on vision navigation can be made in two direction: (1) Vision navigation which imitate human vision to identify the path. (2) Vision navigation by following the initially fixed signs. The latter method is simpler and more flexible, compared to the former one, since it has faster image processing speed and better real time property. The research on my dissertation is focusing on the latter one, that is, intelligent tracklayer’s navigation by following sign.
In my dissertation, I did some research on the technology of intelligent tracklayer’s vision navigation based on the experimental intelligent tracklayer developed independently by our research group. The main contents of my dissertation are as follow:
1. Design the frame and drive system of the experimental intelligent tracklayer
Our research group developed an experimental intelligent tracklayer for the purpose of doing research on the intelligent tracklayer’s vision navigation.
The organizing and control framework of the experimental intelligent tracklayer is devised in my dissertation, that is the apperceiving module, the programming module and the enforcing module.
The process of tracklayer’s turning is described. An introduction concerning of the control principle of how to realize this progress is also made, that is the industry computer controls the two transducers and then the two transducers separately control the two electromotors.
2. Research on intelligent tracklayer’s vision system
My dissertation introduces that the intelligent tracklayer’s vision system are composed of three modules: Image collection module which is responsible for getting the digital image of the road, preprocessing module which is used to make preparation for road sign line identification, and identification module which is able to identify the road sign line laid in the circumstance as well as obtain the position warp and direction warp which are used for decision-making.
A fast and accurate calibration method is adopted to calibrate the CCD camera of the intelligent tracklayer so as to acquire the intrinsic and extrinsic parameters from which the mapping connection of all the coordinates can be achieved thoroughly.
The collected road image is preprocessed. This includes image gray scale transformation, image binarization whose threshold selection is based on the maximum variance, noise elimination which adoptes ameliorated adjacent region average method. Through the partition gridding of the image I make some accomplishment on compressing the data quantity of the image and increase the image processing velocity. I adopt progressive to locate the road sign line, and obtain the two input variables of the fuzzy controller, position warp and direction warp.
3. Research on intelligent tracklayer’s decision-making system
Since it’s difficult to establish an accurate mathematics model for the intelligent tracklayer’s vision navigat
引文
[1] 赵登峰等. 履带车的视觉导航方法. 建筑机械 2002，3：21～24
[2] 王积分，张新荣编著. 计算机图象识别. 中国铁道出版社，1988，第1版 p28～30,p67～77
[3] 章卫国，杨向忠著. 模糊控制理论与应用. 西北工业大学出版社， 1999，第1版 p48～58
[4] 罗飞路，傅恩锡. CCD摄像机内外部参数快速准确的标定方法. 湖南大学学报 1997，4：71～74
[5] 关柏青. 基于彩色视觉和模糊控制的移动机器人路径跟踪. 河北工业大学硕士论文，2002
[6] Ernst D.Dickmanns, The Development of Machine Vision for Road Vehicle in the Last Decade. Processing of the IEEE Intelligent Vehicles Symposium 2002(6)
[7] E.D, Dickmanns, V.Granefe. Dynamic Monocualr Machine Vision and Application of Dynamic monocular Vision. International Journal of Machine Vision and Application, 1998(1)
[8] Masayoshi Aoki. Image Processing in ITS. IEEE International Conference on Intelligent Vehicles, 1998
[9] 闻新等编著. MATLAB模糊逻辑工具箱的分析与应用. 科学出版社，2001，第1版
[10] 韩安荣主编. 通用变频器及其应用. 机械工业出版社，2000，第2版 p66
[11] Hessburg T and Tomizuka M. Fuzzy Logic Control for Lateral Guidence. IEEE Control System Magazine, 1994(8)
[12] 冯炎. 新型视觉引导AGV图像处理技术的研究. 吉林大学硕士论文，2002 p13,p21,p55～57
[13] J.M. Sanchiz，F. Pla，J.A. Marchant. An apprcach to the vision tasks involved in an autonomous crop protection vehicle. Engineering Application of Artificial Intelligence Ⅱ(1998) p175～187
[14] 刘国民，黄海东 履带底盘转向轨迹的研究. 水利电力机械 1998，5：p19
[15] 卢韶芳，刘大维. 自主式移动机器人导航研究现状及其相关技术. 农业机械学报 2002，2 p112～115
[16] 周俊，姬长英. 自主车辆导航系统中的多传感器融合技术. 农业机械学报 2002，5 p113～116
[17] 顾柏园等. 实验型高速智能汽车导航技术. 机器人技术与应用 2002，5 p16～20
[18] 章毓晋. 图像工程（上册）——图像处理与分析. 清华大学出版社，2000，第1版
[19] 容观奥. 计算机图象处理. 清华大学出版社， 2000，第1版
[20] A. Broggi, M.Bertozzi et al. The ARGO Autonomous Vehicle’s Vision and Control System. Intelligent Control and System, 1999(3)
[21] A.O.Djekoune, K.Achour, Visual Guidence Control Based on the Hough Transform. Proceeding of 2000 IEEE International Conference on Intelligenct Vehicles, 2000(10)
[22] 史文革. 微机图像格式大全. 海洋出版社，1992，第1版
[23] 史力晨等. 高速履带车辆在坚实地面转向机理研究. 装甲兵工程学院学报 2003，1 p29～32
[24] 袁军，黄心汉，陈锦江. 基于多传感器的智能机器人信息融合、控制结构和应用. 机器人， 1994（5）p313～320
[25] C.M. Pearson, P.J. Probert. Navigation of Autonomous Vehicles Using Linear Array Cameras. The International Journal of Robotics Research, 1998(6)
[26] 鸿冬清等. 模糊智能控制. 化学工业出版社 1998
[27] Hong Che_wun. A fuzzy Throttle Control for Dynamic Driving Patten Simulation, International Journal of Vehicle Design, 1995(2)
[28] YiMa, Jana Kosecka, and Shankar S.Sastry. Vision Guided Navigation for a Nonholonomic Mobile Robot. IEEE Transactions on Robotics and Automation, 1999(6)
[29] M. Elarbi Boudihir, R.Nourine. Visual Guidence of Autonomous Vehicle Based on Fuzzy Perception. IEEE International Conference on Intelligent Vehicles,1998
[30] Ramehs Rajagopalan, R.M.H. Cheng. Binary Camera for AGV Navigation. Fourth World Conference On Robotics Research, 1991
[31] 晶辰工作室. 最新流行图像格式实用参考手册. 电子工业出版社，1998，第1版 p22～43
[32] M.A. Sotelo, F.J. Rodriguez, L.Magedalena. Vision-based Navigation System for Autonomous Urban Transport Vehicle in Outdoor Environments. Proceedings of the IEEE Intelligent Vehicles Symposium 2002, 2002(6)
[33] Y Goto and A Stents. The CMU System for Mobile Robot Navigation. Processing IEEE International Conference of Robotics and Automation. 1987 p99～105
[34] T Lozano-Perez. Spatial Planning: A Configuration Space Approach. IEEE Transactions on Computer. C-32,1983(2) p108～120
[35] J Canny and B Donald. Simplified Voronoi Diagram. Discrete and Computation Geometry,1988(3) p219～236
[36] V Boschian and A Pruski. Grind Modeling of Robot Cells:A Memory Efficient Approach. Journal of Intelligent and Robotics System,Oct,1993,Vol.8 p89～94
[37] 李兵. 基于视觉的智能车辆自主导航控制的研究. 吉林大学博士论文，2003
[38] S.Bentalba, A.El Hajjaji . Fuzzy Path Tracking Control of a Vehicle. IEEE International Conference on Intelligent Vehicles, 1998 p195～200
[39] R Chatila and J P Laumond. Position Referencing and Consistent World Modeling for Mobile Robot. Processing of the IEEE International Conference on Robotics and Automation. 1985 p138～145
[40] J Borenstein and Y Koren. Histogramic in Motion Mapping for Mobile Robot Obstacle Avoidance. IEEE Transaction on Robotics and Automation, 1991,Vol.7(No.4) p535～539
[41] D Gaw et al. Mininum-Time Navigation of an Unmanned Mobile Robot in a 2-1/2-D World with Obstacles. Processing of IEEE International Conference on Robotics and Automation. 1986 p1670～1677
[42] N.C. Rowe and R.S. Ross. Optional Grid-Free Path Planning Across Arbitrarily Contoured Terrain with Anisotropic Friction and Gravity Effect. IEEE Trans. On Robotics and Automation. 1990,Vol.6(No5):p540～553
[43] 金小平等. 移动机器人的动态路径规划及控制. 机器人，1990，Vol.12(No.6) p:10～16
[44] 沈明霞，姬长英. 农业机器人视觉导航技术发展与展望. 农业机械学报，2001（1）
[45] 赵登峰等. 基于模糊神经网络的智能履带车路径跟踪系统. 农业工程学报 2003，2
[46] C.H.Leung and A.M.S.Zalzala. Genetic Solution for the Motion of Wheeled Robotic System in Dynamic Environments. IEEE Conference Publication, 1994,Vol.1:p389～392
[47] T.S. Chang et al. An Obstacle Avoidance Algorithm for an Autonomous Land Vehicle. Int. Journal of Robotics and Automation,1987, Vol.2(No.1):p21～24
[48] V Lumelsky et al. A paradigm for Incorporating Vision in the Robot Navigation Function. IEEE Int. Conf. Robotics and Automation, 1998. p734～739
[49] Ramsh Rajagopalan, R.M.H. Cheng. Architecture of Operation System software of an Autonomous Vehicle. Computers in Engineering, Volume two, ASME 1991
[50] K Ohshima et al. Fuzzy Control in Tracking of Welding Line. International Conference on New advances in Welding and Allied Process ,Bejing,1991, p238～243
[51] 董再励，王光辉，田彦涛等. 自主移动机器人激光全局定位系统研究. 机器人，2000（3） p207～210
[52] Y.J. Zhang. Quantitative study of 3-D gradient operators. Image and vision computing. 1993(11) p611～622
[53] 何克忠. 智能移动机器人的一种体系结构. 机器人，1992（3）
[54] 戴广平编著. 电动机变频器与电力拖动. 中国石化出版社. 1999.3
[55] 林家明. 面阵CCD摄像机光学镜头参数及其相互关系. 光学技术， 2000（3）
[56] 邱茂林，马颂德，李毅. 计算机视觉中摄像机标定技术综述. 自动化学报，2000（1）
[57] 徐士良. C常用算法程序集. 清华大学出版社. 1996
[58] 刘向杰，周孝信，柴天佑 模糊控制研究的现状与新发展. 信息与控制 1999（4）
[59] 何斌等编著. Visual C++数字图像处理. 人民邮电出版社，2001，第1版
[2] 王积分，张新荣编著. 计算机图象识别. 中国铁道出版社，1988，第1版 p28～30,p67～77
[3] 章卫国，杨向忠著. 模糊控制理论与应用. 西北工业大学出版社， 1999，第1版 p48～58
[4] 罗飞路，傅恩锡. CCD摄像机内外部参数快速准确的标定方法. 湖南大学学报 1997，4：71～74
[5] 关柏青. 基于彩色视觉和模糊控制的移动机器人路径跟踪. 河北工业大学硕士论文，2002
[6] Ernst D.Dickmanns, The Development of Machine Vision for Road Vehicle in the Last Decade. Processing of the IEEE Intelligent Vehicles Symposium 2002(6)
[7] E.D, Dickmanns, V.Granefe. Dynamic Monocualr Machine Vision and Application of Dynamic monocular Vision. International Journal of Machine Vision and Application, 1998(1)
[8] Masayoshi Aoki. Image Processing in ITS. IEEE International Conference on Intelligent Vehicles, 1998
[9] 闻新等编著. MATLAB模糊逻辑工具箱的分析与应用. 科学出版社，2001，第1版
[10] 韩安荣主编. 通用变频器及其应用. 机械工业出版社，2000，第2版 p66
[11] Hessburg T and Tomizuka M. Fuzzy Logic Control for Lateral Guidence. IEEE Control System Magazine, 1994(8)
[12] 冯炎. 新型视觉引导AGV图像处理技术的研究. 吉林大学硕士论文，2002 p13,p21,p55～57
[13] J.M. Sanchiz，F. Pla，J.A. Marchant. An apprcach to the vision tasks involved in an autonomous crop protection vehicle. Engineering Application of Artificial Intelligence Ⅱ(1998) p175～187
[14] 刘国民，黄海东 履带底盘转向轨迹的研究. 水利电力机械 1998，5：p19
[15] 卢韶芳，刘大维. 自主式移动机器人导航研究现状及其相关技术. 农业机械学报 2002，2 p112～115
[16] 周俊，姬长英. 自主车辆导航系统中的多传感器融合技术. 农业机械学报 2002，5 p113～116
[17] 顾柏园等. 实验型高速智能汽车导航技术. 机器人技术与应用 2002，5 p16～20
[18] 章毓晋. 图像工程（上册）——图像处理与分析. 清华大学出版社，2000，第1版
[19] 容观奥. 计算机图象处理. 清华大学出版社， 2000，第1版
[20] A. Broggi, M.Bertozzi et al. The ARGO Autonomous Vehicle’s Vision and Control System. Intelligent Control and System, 1999(3)
[21] A.O.Djekoune, K.Achour, Visual Guidence Control Based on the Hough Transform. Proceeding of 2000 IEEE International Conference on Intelligenct Vehicles, 2000(10)
[22] 史文革. 微机图像格式大全. 海洋出版社，1992，第1版
[23] 史力晨等. 高速履带车辆在坚实地面转向机理研究. 装甲兵工程学院学报 2003，1 p29～32
[24] 袁军，黄心汉，陈锦江. 基于多传感器的智能机器人信息融合、控制结构和应用. 机器人， 1994（5）p313～320
[25] C.M. Pearson, P.J. Probert. Navigation of Autonomous Vehicles Using Linear Array Cameras. The International Journal of Robotics Research, 1998(6)
[26] 鸿冬清等. 模糊智能控制. 化学工业出版社 1998
[27] Hong Che_wun. A fuzzy Throttle Control for Dynamic Driving Patten Simulation, International Journal of Vehicle Design, 1995(2)
[28] YiMa, Jana Kosecka, and Shankar S.Sastry. Vision Guided Navigation for a Nonholonomic Mobile Robot. IEEE Transactions on Robotics and Automation, 1999(6)
[29] M. Elarbi Boudihir, R.Nourine. Visual Guidence of Autonomous Vehicle Based on Fuzzy Perception. IEEE International Conference on Intelligent Vehicles,1998
[30] Ramehs Rajagopalan, R.M.H. Cheng. Binary Camera for AGV Navigation. Fourth World Conference On Robotics Research, 1991
[31] 晶辰工作室. 最新流行图像格式实用参考手册. 电子工业出版社，1998，第1版 p22～43
[32] M.A. Sotelo, F.J. Rodriguez, L.Magedalena. Vision-based Navigation System for Autonomous Urban Transport Vehicle in Outdoor Environments. Proceedings of the IEEE Intelligent Vehicles Symposium 2002, 2002(6)
[33] Y Goto and A Stents. The CMU System for Mobile Robot Navigation. Processing IEEE International Conference of Robotics and Automation. 1987 p99～105
[34] T Lozano-Perez. Spatial Planning: A Configuration Space Approach. IEEE Transactions on Computer. C-32,1983(2) p108～120
[35] J Canny and B Donald. Simplified Voronoi Diagram. Discrete and Computation Geometry,1988(3) p219～236
[36] V Boschian and A Pruski. Grind Modeling of Robot Cells:A Memory Efficient Approach. Journal of Intelligent and Robotics System,Oct,1993,Vol.8 p89～94
[37] 李兵. 基于视觉的智能车辆自主导航控制的研究. 吉林大学博士论文，2003
[38] S.Bentalba, A.El Hajjaji . Fuzzy Path Tracking Control of a Vehicle. IEEE International Conference on Intelligent Vehicles, 1998 p195～200
[39] R Chatila and J P Laumond. Position Referencing and Consistent World Modeling for Mobile Robot. Processing of the IEEE International Conference on Robotics and Automation. 1985 p138～145
[40] J Borenstein and Y Koren. Histogramic in Motion Mapping for Mobile Robot Obstacle Avoidance. IEEE Transaction on Robotics and Automation, 1991,Vol.7(No.4) p535～539
[41] D Gaw et al. Mininum-Time Navigation of an Unmanned Mobile Robot in a 2-1/2-D World with Obstacles. Processing of IEEE International Conference on Robotics and Automation. 1986 p1670～1677
[42] N.C. Rowe and R.S. Ross. Optional Grid-Free Path Planning Across Arbitrarily Contoured Terrain with Anisotropic Friction and Gravity Effect. IEEE Trans. On Robotics and Automation. 1990,Vol.6(No5):p540～553
[43] 金小平等. 移动机器人的动态路径规划及控制. 机器人，1990，Vol.12(No.6) p:10～16
[44] 沈明霞，姬长英. 农业机器人视觉导航技术发展与展望. 农业机械学报，2001（1）
[45] 赵登峰等. 基于模糊神经网络的智能履带车路径跟踪系统. 农业工程学报 2003，2
[46] C.H.Leung and A.M.S.Zalzala. Genetic Solution for the Motion of Wheeled Robotic System in Dynamic Environments. IEEE Conference Publication, 1994,Vol.1:p389～392
[47] T.S. Chang et al. An Obstacle Avoidance Algorithm for an Autonomous Land Vehicle. Int. Journal of Robotics and Automation,1987, Vol.2(No.1):p21～24
[48] V Lumelsky et al. A paradigm for Incorporating Vision in the Robot Navigation Function. IEEE Int. Conf. Robotics and Automation, 1998. p734～739
[49] Ramsh Rajagopalan, R.M.H. Cheng. Architecture of Operation System software of an Autonomous Vehicle. Computers in Engineering, Volume two, ASME 1991
[50] K Ohshima et al. Fuzzy Control in Tracking of Welding Line. International Conference on New advances in Welding and Allied Process ,Bejing,1991, p238～243
[51] 董再励，王光辉，田彦涛等. 自主移动机器人激光全局定位系统研究. 机器人，2000（3） p207～210
[52] Y.J. Zhang. Quantitative study of 3-D gradient operators. Image and vision computing. 1993(11) p611～622
[53] 何克忠. 智能移动机器人的一种体系结构. 机器人，1992（3）
[54] 戴广平编著. 电动机变频器与电力拖动. 中国石化出版社. 1999.3
[55] 林家明. 面阵CCD摄像机光学镜头参数及其相互关系. 光学技术， 2000（3）
[56] 邱茂林，马颂德，李毅. 计算机视觉中摄像机标定技术综述. 自动化学报，2000（1）
[57] 徐士良. C常用算法程序集. 清华大学出版社. 1996
[58] 刘向杰，周孝信，柴天佑 模糊控制研究的现状与新发展. 信息与控制 1999（4）
[59] 何斌等编著. Visual C++数字图像处理. 人民邮电出版社，2001，第1版