未知动态复杂环境下群机器人协同多层围捕  

Cooperative Multilayer Hunting by Swarm Robots in Unknown Dynamic Complex Environments

作　　者：张红强[1,2] 章兢[1] 周少武[2] 欧阳普仁[3] 吴亮红[2] 

机构地区：[1]湖南大学电气与信息工程学院,长沙410082 [2]湖南科技大学信息与电气工程学院,湘潭411201 [3]湖南科技大学机电工程学院,湘潭411201

出　　处：《电工技术学报》2015年第17期140-153,共14页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金(61174140;51374107);国家科技支撑计划项目课题(2012BAH09B02);国家自然科学青年基金(61203309);湖南自然科学基金(13JJ8014);长沙市科技计划(重大专项)项目(K1306004-11-1);湖南省教育厅优秀青年项目(12B043);博士点基金(20110161110035)资助

摘　　要：针对未知动态复杂环境下非完整移动群机器人多层围捕,提出了一种基于简化虚拟受力模型的多层协同自组织围捕方法。首先设计了未知杂乱环境下目标的运动模型,然后通过对杂乱环境下围捕行为的研究,构建了简化虚拟受力模型。基于此受力模型,提出了个体循障和多层协同围捕运动控制方法,该方法只需目标和两最近邻位置信息,计算简单,易于实现。最后证明了系统的稳定性,并获得了参数设置范围。不同情况下的仿真验证及对比结果表明,所提的多层协同围捕方法可使群机器人在未知复杂环境下保持较好的多层协同围捕队形,并具有良好的避障性、高扩展性和灵活性。A multilayer cooperative self-organizing hunting method,which is based on a simplified virtualforce model,is proposed for nonholonomic multilayer mobile swarm robots hunting in unknown dynamic complex environments. First, the motion models for the hunted target in unknown cluttered environments are designed. Then through researching hunting behavior under cluttered environments, a simplified virtual-force model is formed. Based on the virtual-force model, the control method is designed for swarm robots following motions of barriers and multilayer cooperative hunting the target. The method only needs the location information of the target and two nearest neighbors. So it is easy to be calculated and realized. After that,the stability of the hunting system is analyzed, and the control parameter ranges are obtained. Simulation results for different situations and comparative analyses demonstrate that the proposed method,which can make the group of robots maintain a good multilayer cooperative hunting formation in unknown complex environments, has good performance of obstacles avoidance,high scalability,and flexibility.

关 键 词：非完整移动机器人 群机器人 简化虚拟受力模型 队形保持 多层围捕 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]