基于最大一致容积Kalman滤波器的多机器人协同定位CSCD

机器人位置信息是多机器人系统执行任务的前提,单个机器人因传感器载荷和作用范围的限制,难以完成复杂环境中的定位任务,多个机器人通过协作可实现大范围下的位置确定。将配置多个传感器的同构机器人群替换为配置单个或少量传感器的异构机器人可降低硬件成本,并且通过设计协同算法,不会降低定位精度。提出了一种将容积Kalman滤波器与最大一致思想融合后的新型滤波算法,并将该算法应用于麦克纳姆轮机器人系统。通过仿真和实物验证了最大一致容积Kalman滤波器的协同定位效果。

Multi-Robot Cooperative Localization Based on Maximum Consensus Cubature Kalman Filter

Robot position information is a precondition for multi-robot systems to perform their tasks. Due to the limitation of sensor load and measure range, individual robot is difficult to perform localization task in complex environments. But multiple robots can achieve deterministic positions at large ranges through collaboration. Replacing the homogeneous robot population with multiple sensors to heterogeneous robots with single or small sensors can reduce the hardware cost, and the positioning accuracy is not reduced by designing collaborative algorithms. A novel filtering algorithm is proposed by combining the cubature Kalman filter with the idea of maximum consistency, and the algorithm is used to a Mecanum wheels robot system. The collaborative location effect of the maximum consistent cubature Kalman filter is verified by simulation and physical object.