【作者】 王晓丽； 洪奕光；

【Author】 Xiaoli Wang,Yiguang Hong Key Laboratory of Systems and Control,Institute of Systems Science Chinese Academy of Sciences,Beijing 100190

【机构】 中国科学院数学与系统科学研究院系统控制重点实验室；

【摘要】 本文主要研究如何利用动态多智能体系统覆盖跟踪一个移动的目标.每个智能体都备有一个感应器和一个执行器,感应范围是无穷大的但是执行范围是有限的.如果目标能落在某个智能体的执行区域内,我们称该多智能体网络完成覆盖该目标的任务.由于每个智能体只能拿到该目标位置或速度的部分信息,因此首先需要考虑对该目标重构的问题.本文通过设计联合估计算法实现在线估计目标.已有很多结果考虑的目标以静态为主,这里我们考虑如何利用联合观测器和覆盖跟踪控制器跟踪一个动态的目标.虽然目标的动力系统和传感器对目标的量测输出都不可避免受到外界的干扰,但是利用给出的联合估计算法,我们能得到该估计误差的界,进而得知目标会落在以该目标的估计位置为中心,估计误差为半径的球内.这样覆盖跟踪目标的问题就可以转化为覆盖一个移动区域的问题.我们只需设计算法覆盖这一移动的区域.为了覆盖该移动区域,我们利用联合观测器的信息和队形控制的算法为智能体设计分布式的协议.利用我们给出的协议能够实现对目标的覆盖,大大提高了系统性能.另外我们能给出完成该覆盖跟踪任务需要的最少智能体的个数,该数目依赖于理想队形策略的选择、联合估计算法在线估计目标误差的界和智能体执行区域的半径.更多还原

【Abstract】 This paper discuss the problem of coverage tracking a moving target by mobile agents.Every agent is equipped with a sensor and an actuator,where the sensing range is infinite and the actuating range is finite.If the target is in an actuator of some agent,the target is covered by the agents.Under our consideration,the sensor measurement only gives some partial information about the position or velocity of the target in question.Therefore,the estimation of its state based on measurements from the sensors is considered first.The collaborative estimation algorithms are proposed to get the information of the moving target.Conventionally,the target was assumed to be static.Here with a combination of observer-based state estimator and coverage tracking controller we investigate the problem of tracking a moving target.Although the target dynamics and the measurement output contain uncertainties,we can limit the estimation error within some disk centered at the estimation of the position of the target.In this way,the target coverage tracking problem can be viewed as the problem of covering a moving region.Then the coverage protocols for mobile agents are built based on the constructed observers and formation strategies.With the given protocols,the coverage mission is achieved.The number of the desired agents to complete the mission is also estimated,which may vary if we take different formation schemes.It is based on the bound of the estimation error and the range of the actuation region.更多还原