多机器人编队控制研究进展

Research development of multi-robot formation control

  • 摘要: 首先介绍了传统的编队控制方法的定义、特点和常用方法及优缺点,并将传统编队控制时代定义为前编队控制时代.随着多智能体技术的发展,将多智能体技术引入到编队控制问题中,诞生了众多新的研究成果,称为后编队控制时代.后编队控制时代以多智能体技术为基础,随着通信技术、计算机技术、人工智能技术的发展而逐渐壮大起来,并受到了学者的广泛关注.前编队控制时代强调多机器人通过编队协作完成单个机器人无法实现的任务,提高任务完成效率且缩短任务完成时间.后编队控制时代则是在前编队控制时代的基础上,更强调低成本、同步性和协同性,但却不那么重视每个个体的任务分工,甚至是按照规则自由分配任务,不再有"不可替代"的个体存在.最后给出了研究编队控制问题的基本思路和目前尚待解决的关键问题.

     

    Abstract: Multi-robot formation control is one of the most important research directions in the field of robotics. In this paper, the definition and characteristics of multi-robot formation control were given and several traditional formation control methods were introduced, such as leader-follower method, behavior-based method, artificial potential method, and virtual structure method, and their advantages and disadvantages were stated. The era in which these traditional control methods are applied is termed as "Former Formation Control Era" (FFCE). Following the development of multi-agent theory, the multi-agent technology has become popular and is considered as an efficient solution to the multi-robot system formation control problem, and it has generated several favorable results. The era in which this technology is adopted is termed the "Post Formation Control Era" (PFCE). In this era, the multi-agent theory has gained much attention following the developments in the communication technology, computation technology, and artificial intelligence technology. The FFCE emphasizes that multiple robots are able to accomplish complex tasks that are unachievable by a single robot, while forming a desired formation, which improves the efficiency of a required task as well as shorten the completion time of the task. However, the PFCE emphasizes much more cost-effectiveness, synchronization, and coordination than the FFCE. Meanwhile, the PFCE does not pay much attention to the task assignment for each individual, but assigns each sub-task autonomously to each individual on the basis of simple rules, and there exists no "irreplaceable" individual. Finally, this paper presented the basic steps for solving the formation control problem using multi-agent technologies.

     

/

返回文章
返回