基于粒子滤波的壁面爬行机器人位姿跟踪研究

Pose Tracking of Wall-climbing Robot Based on Particle Filtering

为提高壁面爬行机器人的智能化水平,提出一种利用粒子滤波算法进行机器人位姿跟踪的方法。所述方法在机器人身上安装了发声器,将机器人位姿跟踪问题转换为声源跟踪问题处理。设计了一个声阵列传感器,采用可控波束形成定位技术观测机器人位姿,利用里程计描述机器人的位姿变换动态模型,并最终建立了基于贝叶斯估计的粒子滤波位姿跟踪算法。搭建了实验平台进行现场实验,结果表明该方法可行并有实际应用价值。

In order to improve the intelligence of wall-climbing robot, a pose tracking method of the robot based on particle filtering is presented. A speaker is mounted on the robot, and a acoustic array sensor is designed, thus the position of wall-climbing robot is detected by acoustic source tracking method with steered beamformer algorithm. The dynamic state of robot is described with odometer, and a particle filtering pose tracking algorithm is proposed. A experimental platform has been set up, and the method is proved efficient and practical.