【作者】 汪浩杰；

【导师】 付勇；

【作者基本信息】 华中科技大学 ， 控制理论与控制工程， 2007， 硕士

【摘要】 随着分布式人工智能技术的发展,多机器人系统的研究得到越来越多的关注。多机器人系统在大型船舶制造,现代物流系统中的仓库管理,空间探测,无人作战系统,还有社会和家庭服务等行业中都有广阔的应用前景。多机器人的智能围捕是一个常用的检验机器人学习策略优劣的平台,对智能移动机器人的产业化具有重要的意义。论文推导出了多机器人围捕的临界条件。在深入分析多机器人围捕数学模型的基础上,论文用求极值的方法,推导出了围捕机器人与逃跑者速度的最小比值。在智能水平和视野范围都相同的情况下,围捕机器人和逃跑者的速度之比只有大于这个临界值,围捕任务才可能完成。这个理论为实际工程设计中的资源分配提供了参考依据。论文提出了一种新的多机器人围捕方案。当围捕机器人的速度小于逃跑者时,采用常规的围捕方法很难完成任务。利用这种新的智能伏击策略,让各个围捕机器人相互协作,将逃跑者驱入最佳围捕区,然后按照文中提出的智能围捕策略实施围捕任务,提高了围捕的成功率。论文设计出了一种智能的逃跑策略。为了增大围捕任务的难度,缩小仿真实验与实际情况的差距,论文改进了“最近避让策略”,将围捕机器人的速度和方向都考虑在内,然后采用加权矢量和的方法确定逃跑的方向。这样提高了智能围捕平台测试结果的可信度。最后的仿真实验结果证明了围捕临界条件的存在性和新的围捕方案的可行性。更多还原

【Abstract】 As the development of Distributed Artificial Intelligence technology, the research of Multi-Robot System (MRS) is receiving more and more attention. MRS can be widely applied in industries such as large scale manufacturing like large ship manufacturing, warehouse management in modern logistics, space exploring, and unmanned combat system, as well as private and public service industry. The intelligent capture among the MARS is a testing platform for the effectiveness of robot learning strategy,it is of great significance to put intelligent mobile robot into industrial utility.Firstly, this thesis deduces the critical condition of collective capture. Analyzing thoroughly the mathematical model of MRS, this thesis deduces the minimum velocity ratio between the capture robot and evader with the method of working out the extreme value. Provided their intelligence and field of view are equal, the task of capture can be accomplished only if the velocity ratio between capture robot and evader are larger than this critical value. This theory provides guidance for the resources-allocating task in real engineering design.Secondly, it proposes a novel Multi-Robot scheme. It is rather difficult to complete the task with the general capture approach when the velocity of capture robot is less than that of the evader. Chasing the evader to the optimum capture zone with cooperation between capture robots, and implementing the intelligent capture strategy to finish the task, this new intelligent ambush strategy increases the success rate of capture. Thirdly, it designs an intelligent evasion strategy. In order to increase the difficulty of this task and fill the gap between the simulation experiment and real game, this paper improves Evading the Most Nearest Object Approach by taking into account of the vectors of all evasion direction and adding them all together. Thus it makes the testing result of the intelligent capture platform more convincible.Finally the statistics of the simulation experiment data justifies the existence of the critical condition under which the task is implemented and the feasibility of the scheme devised.更多还原