摘要:
动态目标的多移动机器人主动协作观测方法是指以获取较优的观测结果为目的, 对携带同构/异构观测传感器的多个机器人系统的观测数据进行有效融合并同时对其行为进行协调优化的方法. 本文主要研究了三维环境中的多机器人动态目标主动协作观测的问题. 首先, 以扩展集员估计方法(Extended set-membership filter, ESMF)为基础, 将信息融合过程与算法本身存在的集合运算环节相结合, 提出了一种高精度的多机器人观测信息融合方法. 该方法在保证较高观测精度的同时, 并没有显著增加单机器人扩展集员估计算法的计算量, 因此具有较高的实时性. 此外, 利用最优观测角度的概念, 通过引入相对速度空间(Relative velocity coordinates, RVCs), 设计了多移动机器人协调行为优化方法, 该方法可以将多机器人协调行为优化问题转化为线性规划问题, 以实现具有较高实时性的多机器人三维动态目标主动协作观测. 最后, 为了验证所研究方法的可行性与有效性, 进行了三维空间动态目标协作观测仿真实验.
Abstract:
The solution to multiple mobile robot systems actively and cooperatively observing a moving target (MACO) means the algorithm that tries to pursuit optimal (sub-optimal) observations of the moving target by simultaneously fusing the observational data from multiple robot systems and regulating their behaviors cooperatively. In this paper, the 3D MACO method with two robots is studied. First, at the basis of extended set-membership filter (ESMF), a high precise observation fuse method is presented through combining the information fuse process and the set operations in ESMF algorithm. The new algorithm is as fast as the single ESMF algorithm since it has almost the same computational burden. Second, a coordinate behavior optimization method is given by combining the concept of optimal observational angle and the relative velocity coordinates (RVC) planning method. By using the RVC method, the coordinate behavior optimization can be transferred into a linear planning (LP) problem, which makes its real time application possible. Finally, 3D moving target observational simulations are conducted to verify the feasibility and validity of the proposed algorithm.
点击查看大图
下载:
