Position Information Sharing and Mutual Localization of Multiple Mobile Robots using Infrared Wireless Communication

本课题研究了基于红外无线通信系统的多移动机器人协同作业。在这项研究中，研究了一个系统，该系统使用通信伙伴而不是地标来创建他们的工作空间地图。首先，通过仿真计算了该红外无线通信系统相互位置信息的位置测量误差。采用三角测距法对三个机器人的位置测量进行了仿真。除算法不可计算的情况外，算法的准确率均在90%以上。然而，移动机器人不能移动到仅使用它们的相互位置信息确定的工作空间中的位置。这是因为它们的工作空间的坐标系往往与机器人之间共享的相互位置信息的坐标系不同。我们研究了如何检测建筑内部工作空间的坐标系统，使用由墙壁和柱子形成的角落和边缘作为地标。为了探测障碍物，移动机器人使用超声波传感器和电视摄像机。我们研究了如何使用传感器检测角落或边缘的方向和距离。利用超声波传感器进行距离测量，其精度取决于其检测电路的不同。通过对相位检测与包络检测的比较，证实了超声传感器测距误差较小。此外，我们还研究了一种利用电视摄像机对角和边缘进行方向检测的方法。电视摄像机的角度检测分辨率取决于其镜头和CCD器件。我们利用电视摄像机的图像验证了拐角和边缘方向检测的准确性。我们将在这些研究成果的基础上，研究一种创建适合远程控制移动机器人的工作空间地图的方法。

This is research of cooperative operation of multiple mobile robots that use an infrared wireless communication system. In this research, a system that creates maps of their working space using communicating partners instead of landmarks was studied.At first, the position measurement error of mutual position information that uses this infrared wireless communication system was calculated in a simulation. The simulation of position measurement of three robots was performed using triangulation ranging. The accuracy of the algorithm was more than 90%, except for the case where the algorithm is incomputable.However, mobile robots cannot move to a position in their working space that is decided using only their mutual position information. This is because the coordinate system of their working space often differs from the coordinate system of mutual position information shared by robots. We examined how to detect the coordinate system of the working space that is inside a building, using corners and edges that are formed by walls and pillars as landmarks.In order to detect obstacles, the mobile robots use ultrasonic sensors and TV cameras. We examined how to detect the direction of and distance from a corner or edge using sensors. The distance measurement using ultrasonic sensors has a difference in accuracy depending on its detection circuit. We confirmed small errors of distance measurement in experiments of ultrasonic sensors using phase detection that is compared with envelope detection.In addition, we examined a direction detection method of corners and edges using TV cameras. The resolution of angle detection of TV cameras depends on their lens and CCD devices. We confirmed the accuracy of direction detection of corners and edges using images from TV cameras.We will study a method of creating a working space map suitable for remote controlled mobile robots based on these research results.