通讯范围受限条件下的多机器人主动SLAM

针对机器人通讯范围受限的问题，提出一种新的多机器人协作同时定位与建图(simultaneous localization and mapping, SLAM)方法。多个机器人采用基于最优控制的主动探索策略，创建自身周围区域的子地图，在每个建图周期内使用扩展的卡尔曼滤波器(extended Kalman filter, EKF)估计和维护子地图状态，并在一个周期结束后联络其通讯范围内的其他机器人，进行子地图的传递与融合。同时，为避免由通讯范围受限带来的地图过度融合问题，每个机器人保存每个建图周期内的局部子地图，待与其他机器人相遇时只传递并融合子地图的增量部分。仿真实验验证了该方法的有效性。

Multi-robot active SLAM under limited communication range

For the problem of communication constraint between robots caused by limited communication range,a new approach for cooperative multi-robot simultaneous localization and mapping(SLAM) is proposed.The robots adopt an active exploration strategy based on optimal control to build submaps around themselves, and use the extended Kalman filter(EKF) to estimate and maintain the states of submaps during every map-building cycle, and connect with other robots at the end of every cycle, sending and fusing submaps.At the same time, in order to avoid the over-fusion problem, which is introduced by limited communication range, it is suggested to save the submaps of every map-buildig cycle, sending and fusing the incremental parts when the robots meet each other. Simulation results validate the effectiveness of the presented approach.