多机器人围捕与分布式目标跟踪的研究
文献类型：学位论文
| 作者 | 张文文 |
| 学位类别 | 工学硕士 |
| 答辩日期 | 2010-05-18 |
| 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 | 曹志强 |
| 关键词 | 多机器人系统 局部感知 协调围捕 无线传感器网络 多目标跟踪 multi-robot system local sensing cooperative hunting wireless sensor network multi-target tracking |
| 其他题名 | Research on Multi-robot Hunting and Distributed Target Tracking |
| 学位专业 | 控制理论与控制工程 |
| 中文摘要 | 多机器人系统是机器人研究领域一个富有挑战性的研究课题，通过群体协调可完成更为复杂的任务；无线传感器网络通过节点间的信息互补，可实现对环境无处不在的感知，与机器人系统相结合可进一步提高系统的整体性能，提高完成任务的效率。本文针对多机器人分布式协作围捕和混合网络中分布式目标跟踪开展研究，主要内容如下： 首先，对多机器人系统和无线传感器网络的研究现状进行了综述，重点阐述了多机器人协调围捕方法与传感器网络中多目标跟踪方面的研究现状，并对本文的选题背景和结构安排做了介绍。 其次，介绍了多异构机器人系统实验平台，包括多地面移动机器人和微小型无人机。描述了地面移动机器人的系统结构，以及驱动控制和传感器子系统，给出了多机器人通讯系统的实现。 第三，针对多机器人协调围捕问题，在多传感器信息处理与有效扇区的基础上，给出基于局部感知、辅以少量通讯信息的分布式地面机器人围捕方法。提出两种围捕模型：由搜索、绕障以及围捕三种状态构成的围捕模型I；以及包括初始固定跟随搜索、可变跟随搜索和围捕三种模式的围捕模型II，采用基于入侵者观察状态改变的通讯策略以降低通讯量。机器人通过所设计的局部协调规则实现交互协调。将多地面移动机器人与无人机结合起来，利用无人机机载视觉传感器为地面监控台提供相关环境信息，地面监控台进而根据目标的情况启动相应的地面机器人小组执行任务。实验验证了围捕方法的有效性。 第四，针对多目标跟踪这一任务，基于静态节点和动态节点组成的混合网络，给出了覆盖空洞的提取以及移动节点巡视路径的选取方法，静态节点网络采用具有通信能耗和K 级感知约束、基于动态跟踪簇的分布式多目标跟踪方法，考虑环境中存在覆盖空洞的情况，设计了移动节点和静态节点相互协作的联合目标跟踪策略，通过仿真进行了验证。 最后，对所取得的研究成果进行了总结，并阐述了下一步的工作。 |
| 英文摘要 | Multi-robot system is one challenging research subject in robotics, and it may accomplish more complicated tasks through inter-robot cooperation. Wireless sensor networks can implement ubiquitous sensing of the environment based on complementary information of sensor nodes deployed in the monitoring region. The multi-robot system provided with sensing information of WSN can further improve the group performance and task completion efficiency. This paper focuses on the multi-robot cooperative hunting and distributed multi-target tracking in hybrid WSN with both static and mobile sensor nodes. The main contents are as follows: Firstly, the current research work of multi-robot system and wireless sensor networks are introduced. The cooperative hunting approaches in MRS and multi-target tracking strategies in WSN are presented. The background and structure of this thesis are also given. Secondly, the heterogeneous multi-robot experimental system is introduced, which includes multiple ground mobile robots and an UAV. We describe the system architecture, drive control and sensors system of the mobile robot. The multi-robot communication system is presented. Thirdly, based on multi-sensor information processing and the concept of effective sectors, a distributed multi-robot hunting method with local sensing and modest communication is given. Two hunting models are proposed: the first one includes search, round-obstacle and hunting states; the second one has initial leader-fixed following&search, leader-changeable following&search and hunting modes, which adopts an event-trigger communication scheme based on the evader’s observation state to reduce the dependence on communication. The local coordination rules are given and the cooperation may emerge. When the UAV is involved in hunting, it provides the ground console with environment information through onboard camera and thus the ground console informs the corresponding robotic group to execute the task based on the evader information. The experiment results show the effectiveness of the hunting. Fourthly, for target tracking, based on the hybrid WSN with both static and mobile sensor nodes, we give the methods to extract sensing coverage holes and select the patrol path for mobile sensor nodes. The static network adopts a cluster-based distributed multi-target tracking approach with the constraints of communication cost and K-level coverage, and considering the sensing coverage hol... |
| 语种 | 中文 |
| 公开日期 | 2015-09-08 |
| 其他标识符 | 200728014628005 |
| 源URL | [http://ir.ia.ac.cn/handle/173211/7532]  |
| 专题 | 毕业生_硕士学位论文 |
| 推荐引用方式 GB/T 7714 | 张文文. 多机器人围捕与分布式目标跟踪的研究[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2010. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
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。