【作者】 李强；

【导师】 刘国栋；

【作者基本信息】 江南大学 ， 控制理论与控制工程， 2011， 硕士

【摘要】 随着机器人科学和技术的发展,人们在机器人方面的研究不断深入,机器人的应用也日益广泛。传统的单一机器人系统已不再满足人们的需求,而相互协作的多机器人系统有着诸多优点被人们所重视,已成为机器人研究的热点,其中队形控制是多机器人协作系统的重要研究方向。文中首先介绍了多机器人协作系统的发展、内容、特性等方面,探讨了队形控制与协作系统的关系,并深入学习了机器人系统的体系结构,重点分析了面向协作的分层式控制架构,在此基础上设计了适用于队形控制的体系结构。其次,针对机器人队形控制在复杂环境下出现的队形控制失败及机器人走失的现象,将领航者法和行为法相结合,依据通信和协作手段,提出了以跟踪链方式穿越障碍的策略,其中融合了机器人处理事件的独立性和协调性,增强了机器人在复杂环境下的适应能力,同时在避障活动障碍时借鉴人为经验有预见的主动避开。再次,针对机器人队形形成方法中存在的问题进行研究。当前队形形成方法对其快速性、合理性的考虑较少,如何在尽量短的时间内并且尽可能避免碰撞的要求下形成队形,提出了一种距离目标位置最小距离最大值优先分配目标位置的控制方法,其中构建了机器人与目标位置的距离矩阵,并对矩阵进行映射分析,方便了机器人进行数据分析查询,较快的确定合理的目标位置。此外,在机器人队形形成的过程中,应用模糊控制理论,并且依据人为经验制定出模糊规则,使机器人灵活的避开障碍并顺利组成指定队形。最后,以实验室的类人足球机器人为研究对象,针对类人机器人建模困难,在不依据数学模型的情况下进行队形控制研究。其中制定了机器人的队形参考规则,构建出机器人队形信息知识库,并且依据有限的传感信息,提出了基于主从知识联想的平行四边形法,较容易的得到当前机器人下一步的目标位置信息,进而获得机器人的运动矢量,在队形控制精度要求不高的情况下有较好的控制效果。另外,在机器人的队形控制中采用模糊控制进行动静障碍物的避障。更多还原

【Abstract】 As robots science and technology development , the people in robot research unceasingly thorough, and the application of robots are also increasing widely. The traditional and single robot system no longer satisfy people’s needs, and mutual cooperation multi-robot system has many merits, has become a robot research hotspot,and formation control is an important research direction among multi-robot operation system.Firstly, this paper introduces the multi-robot operation system about development, content, character etc, discusses the relationship of formation control and cooperation system, and learns the robot system structure, and focuses on analyzing the objectoriented hierarchical control architecture of collaboration, and designs a structure that suitable for formation control system.Secondly, in view of the phenomenon that robots formation control often failure and lost under the complex environment , combines the method of pilot and behavior,and put forward to tracking chain mode through obstacles strategy, incorporates the independence and coordination, and enhance the robot in complicated environment adaptiveness, at the same time foresees active avoiding barriers that reference human’s experience in the avoidance activities obstacles.Again, aiming at the robot formation shape method of the problems in research. The current formation method consider less about faster and rationality , and how to form formation in short period of time and avoid collision as far as possible, then puts forward a method that maximum of minimum distance priority target location distribution, which constructs the robot and target position matrix, and analysis by mapping, and convenient robot analyzes the data, rapidly determinae the reasonable target location. In addition, during the formation, the application of the fuzzy control theory, and based on human experience formulate fuzzy rules, allowing robots flexible avoiding obstacles and success of designated formation.Finally, by the laboratory humanoid robot soccer as the research object, in view of the difficulty in humanoid modeling,which formulate the robot’s formation reference rules, and construct robot formation information repository, and depending on the limited sensor information, then propose a method that base on master-slave knowledge lenovo parallelogrammic,it is easy to get the next target location information for current robots, and then get the motion vector, the robots formation will get a good control effect on the condition that accuracy is not high. In addition, using fuzzy control to avoide obstacles in the formation control.更多还原