【摘要】 多目位姿实时监测系统是一种高精度的数字化测量设备，可用于获取目标物在空间中三维位置信息。本文将多个麦克纳姆移动机器人和多目视觉系统相结合构建智能体集群系统，该系统由定位子系统、通信子系统、控制子系统3部分组成。多目视觉系统作为定位系统通过反光靶球对智能体进行定位，通信子系统将定位信息实时输入控制子系统，控制子系统识别不同的智能体，并将相机坐标系下位姿关系转换到世界坐标系下，再根据不同智能体彼此间的位置关系，设计相应的协同控制算法解算单个智能体所需的姿态控制输入，控制智能体达到指定的位置。测试结果表明，用多目位姿实时监测系统辅助多智能体定位，定位误差小，可令整个智能体集群拥有良好的协同控制性能。更多还原

【Abstract】 Multi-camera pose monitoring system is a high precision digital measuring equipment which can be used to obtain the three-dimensional position information of the target in environment. We constructed a new agent system consisting of many Mecanum wheel mobile robots and multi-camera pose monitoring system, which can be divided into three parts: positioning subsystem, communication subsystem and control subsystem. First, the multi-camera system acts as a positioning system to locate the agent through the reflective target balls, then the communication subsystem sends large positioning information to the control subsystem in real time. The control subsystem identifies different agents, and converts the pose relationship under the camera coordinate system to the world coordinate system. Finally, according to the position relationship between different agents, the corresponding cooperative control algorithm is designed to solve the input parameters required by a single agent, and control the agent to reach the specified position precisely. The experiment results show that the agent positioning assisted by the multi-camera pose monitoring system has small positioning error, which can make the whole agent cluster have good cooperative control performance.更多还原