Infrastructure to Support Research on Long-Term Visual Navigation of Multiple Autonomous Robots

所要求的基础设施将用于开发一个先进的测试环境，用于户外基于视觉的机器人导航。两个地面漫游者和一个空中机器人（一架自动直升机）将配备最先进的立体摄像机。这些机器人将部署在多伦多大学航空航天研究所的场地上（面积约40，000平方米），并将通过高速无线网络连接在一起。实验将比较视觉导航的精度与地面实况测量提供的全站仪测量仪器。为了容纳立体摄像机产生的大量信息，并促进机器学习算法所需的数据挖掘，视觉数据将被传输到大型存储阵列。可以通过无线网络访问此存储阵列。使用所要求的基础设施，将在机器人的长期自主视觉导航方面取得突破。

The requested infrastructure will be used to develop an advanced test environment for outdoor, vision-based robotic navigation. Two ground rovers and one aerial robot (an autonomous helicopter) will be outfitted with state-of-the-art stereo cameras. These robots will be deployed on the grounds of the University of Toronto Institute for Aerospace Studies (an area of approximately 40,000 square metres), and will be linked together through a high-speed wireless network. Experiments will compare the accuracy of visual navigation with ground truth measurements provided by a total station survey instrument. To accommodate the large quantity of information generated by the stereo cameras, and to facilitate the data mining required by machine learning algorithms, visual data will be transferred to a large storage array. This storage array will be accessible over the wireless network. Using the requested infrastructure, breakthroughs will be made in long-term autonomous visual navigation for robots.