多机器人最大熵博弈协同定位算法

研究了多机器人观测到同一目标时的协同定位问题。建立了各个机器人相对观测一致程度的数学描述模型,进而提出用基于极大熵准则的最大熵博弈获取使相对观测一致程度最优的协同定位方式。针对博弈结果的多样性,相应地改变观测方程的雅克比矩阵,推导了可适应多机器人各种博弈结果的扩展Kalman滤波协同定位算法。仿真实验表明,方法可实现机器人团队在协同定位时有选择、更高效地共享相互间的观测信息;在保证协同定位精度提高的同时有效地消除了多机器人相对观测信息间的冲突。

A new cooperative localization algorithm based on maximum entropy gaming

This paper focuses on the problem of cooperative localization when an object is detected by robots simultaneously. As each robot has its own relative observation about the object, a mathematical model for comparing the consistency of these relative observations is presented. With that method, we propose a new cooperative localization algorithm based on maximum entropy gaming and Extended Kalman filter. As the gaming results are different, the Extended Kalman filter equations that can match any gaming result have been derived. Several simulation results showing that the proposed algorithm improves localization performance while avoids the relative observations conflict problem in cooperative localization.