QQ客服 官方微博 反馈留言

题名	多机器人协调探测未知环境的研究
作者	苏丽颖
学位类别	工学博士
答辩日期	2003-05-22
授予单位	中国科学院研究生院
授予地点	中国科学院自动化研究所
导师	谭民
关键词	多机器人系统 环境探测 导航 地图构建 Multi-robot System Environment Exploration Navigation Map Building
其他题名	RESEARCH OF UNKNOWN ENVIRONMENT EXPLORATION BY MULTIPLE COOPERATIVE ROBOTS
学位专业	控制理论与控制工程
中文摘要	随着科学技术的发展，机器人学的研究已从工业制造领域扩展到航空航天、 军事、核工业、服务业等许多不同领域。许多复杂任务如果仅由单个机器人来 完成，存在许多不足，所以需要研究多个机器人的协调作业。而且随着作业环 境的逐渐扩展，在很多情况下，机器人对其作业环境是没有预先了解的，比如 危险的矿井、核废墟、外星球等，很难通过其他的措施对这些环境进行探测， 所以就需要派机器人到这些环境中进行探测以及其他的作业。这些研究也是机 器人智能化的基础。因此本文就多机器人系统探测环境展开研究。 本文首先对多机器人系统的特点和优势，机器人在探测环境中的相关内容 以及研究现状等进行了综述性介绍，并对本文的选题背景和主要内容作了介绍。 其次，本文介绍了多机器人探测环境的基础技术，即信息融合技术。在多 机器人构建地图的研究中，提出了采用基于D—S证据推理的融合算法，设计了 一种机器人之间进行协调避碰的策略；研究了多机器人在环境探测中的定位问 题，提出采用卡尔曼滤波器融合不同机器人对同一环境特征的探测信息，从而 减小了各机器人对环境物体的定位误差，提高了机器人对环境物体的定位精度。 第三，提出了一种机器人完整探测未知环境的方法，即基于虚拟离心力的 导航算法。通过信息共享，以螺旋形折线的方式逐层扩展，该算法实现了多个 机器人协调高效构建环境地图。 第四，提出了一种基于障碍物影响系数的路径规划算法，设计了一个代价 函数，综合考虑了以障碍物为参照的局部实时感知信息对机器人运动选择的影 响和以目标点为参照的距离转换全局规划，兼具全局优化和实时局部优化的长 处，增强了规划的反应能力，在多机器人协调的任务中研究了其应用。 第五，在探测出环境信息的基础上，研究了环境拓扑图的提取。针对栅格 式地图的特征，提出将栅格地图的环境特征划分成8种基本的拐角类型，在此 拐角分类的基础上将环境分区，从而获得环境的拓扑图。 第六，为了满足研究的需要，本文开发了开放式的多机器人仿真实验平台， 用户可以将各种任务程序添加进去，以验证其控制算法。 最后，对本论文所进行的工作和和取得的成果加以总结，并指出了需要进 一步研究的工作。
英文摘要	With the development of science and technology, the Robotics research has extended from manufacturing industry to many different domains such as aerospace & aviation, military application, nucleus industry and service industry. There exists some insufficiency for single robot to complete complicated tasks, which makes necessary the cooperative operation among multiple robots. And with the expansion of robot working environments, there is no a priori knowledge of the environment in many robotic tasks. For example, the hazardous mine, nucleus relics and far celestial bodies. It is difficult to acquire the detailed information about those environments. But robots can be launched to explore those environments and execute more operations. Such relevant researches also lay the foundation for robot intelligence. So the exploration of unknown environments by multi-robot systems is studied in this dissertation. Firstly, the characteristics and the advantages of the multi-robot systems are summarized. The research actuality of the robotic exploration and the relative aspects are reviewed. The background and the main contents of this dissertation are described briefly. Secondly, information fusion technology, one of the basic technologies of environment exploration by multiple robots, is introduced. D-S evidential reasoning based multi-sensor fusion algorithm is proposed to be adopted in the research of map building by multiple robots. A strategy is designed for the collision avoidance between cooperative robots. Also, the localization problem during the environment exploration by multiple robots is studied. Kalman filters are proposed to be adopted for the position information fusion of the same environment feature detected by different robots. With Kalman filters, the environmental localization errors of separate robot can be decreased and so the localization precision of robots can be improved. Thirdly, a virtual centrifugal force based navigation algorithm is presented for the complete exploration of the unknown environments by robots. By sharing the detected environment information, multiple cooperative robots can efficiently map the whole environment in a shape of spiral expansion. Fourthly, an obstacle influence coefficient based path planning algorithm is presented, in which a cost function is designed to integrate the influence of locally sensed real-time information on the robot motion selection with a detected obstacle as reference and the global planning of distance transaction with the target position as reference. That planning method possesses both advantages of the global optimization and the real-time local optimization. With that method, the reaction ability of robots is increased in the planning tasks. And the application of the algorithm in the cooperative multi-robot tasks is studied. Fifthly, a method of extracting topological graph from the detected environment information is studied. Contraposing
语种	中文
其他标识符	725
内容类型	学位论文
源URL	[http://ir.ia.ac.cn/handle/173211/5766]
专题	毕业生_博士学位论文
推荐引用方式 GB/T 7714	苏丽颖. 多机器人协调探测未知环境的研究[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2003.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。