全区域覆盖自主移动机器人无信标边界生成和识别的研究

【作者】 岳峰；

【导师】 王华坤；

【作者基本信息】 南京理工大学 ， 机械设计及理论， 2004， 硕士

【摘要】 本文以全区域覆盖自主移动机器人为研究对象，着重研究其工作区域边界生成和识别技术。作者摒弃了目前国内外一些专利和机器人产品大多采用基于信标的边界生成和识别方法，提出无信标边界自主识别解决方案。文中采用组合定位系统建立边界“电子地图”，机器人在工作过程中通过定位系统实时获得当前的位置信息，并将其与存储的边界信息进行对比，判断是否到达边界。文中应用RBF人工神经网络的方法实现移动机器人航向角的精确测量，并提出了基于环境特征的定位误差修正方案，以上策略都使得基于定位系统的边界生成和识别方案具有较高的精度。最后，作者根据全区域覆盖自主移动机器人的特点，对整个方案的硬件系统和软件系统进行了设计。本课题研究为全区域覆盖移动机器人的边界生成和识别提供了一种低成本、高精度的设计方案，具有一定的理论意义和实用价值。更多还原

【Abstract】 In the paper, a new solution to the boundary set-up and identification of unmarked work area for mobile robot is proposed. Combined localization systems are used to establish the "digital map" of the boundary and also used to acquire the current position information of mobile robot to judge whether the boundary is reached during the course of working. According to the specialties of outdoor environment, the localization systems with combined sensors and the localization algorithm based on RBF neural network are designed. The strategy based on the environment features is also applied to eliminate the accumulated error of the localization systems. The results of simulation and experiments show that the real-time locating information is reliable to ensure the veracity of boundary identification. The research presented in the paper is valuable to the development of the region-coverage mobile robots.更多还原