Roads? Where We're Going We Don't Need Roads! A Robot Platform for Offroad Self-Driving Research

道路？

• 批准号: RTI-2022-00494
• 负责人: Barfoot, Timothy
• 金额: $ 10.93万
• 依托单位: University of Toronto
• 依托单位国家: 加拿大
• 项目类别: Research Tools and Instruments
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=743075
• 关键词: Roads; Where; re; Going; Don

While there is currently world-wide effort focussed on the development of self-driving cars and trucks, there are also many other applications of autonomous mobile robots that are either already in operation or can be brought to market in the near term. In this research, we are interested in offroad self-driving, where (at least some of the time) a robot cannot rely on the presence of road infrastructure for navigation and often has to contend with rough terrain. Examples of these applications are last-mile delivery, campus shuttles, agriculture, mining, mowing, search & rescue, environment monitoring, safety/security, and planetary exploration. The University of Toronto has a long-term research effort focussed on offroad self-driving for such applications, with a proven track record of results in terms of publications and technology transfer to industry/government partners. For example, new algorithms allowing autonomous operations of mobile robots have been provided to MDA Robotics and the Canadian Space Agency (for planetary exploration), Clearpath Robotics (for warehouse and outdoor logistics such as airports), Drone Delivery Canada (for unmanned aerial vehicle delivery), and Electric Sheep Robotics (for large-scale mowing). The proposal requests a new state-of-the-art unmanned ground vehicle (UGV) equipped with sensors (a stereo camera, a laser-rangefinder, and a radar) to continue cutting-edge research into these important offroad robotics applications. There is still much research needed to enable offroad applications including being robust to all kinds of weather and terrain, being safe enough to work alongside humans, and being efficient enough to move quickly in challenging environments. Such a platform serves as a critical stepping stone before shipping software to end users, by allowing frequent local experiments to be conducted. These experiments are not only a necessary part of the scientific process but also provide students, postdocs, and faculty important hands-on learning opportunities to acquire the skills they will need to keep Canada at the forefront of this exciting field.

虽然目前全球范围内的努力都集中在自动驾驶汽车和卡车的开发上，但自主移动机器人也有许多其他应用，要么已经投入运营，要么可以在短期内推向市场。在这项研究中，我们对越野自动驾驶感兴趣，在这种情况下(至少在某些时候)，机器人不能依赖道路基础设施的存在进行导航，并且经常不得不与崎岖的地形作斗争。这些应用的例子包括最后一英里递送、校园班车、农业、采矿、割草、搜索和救援、环境监测、安全/安保和行星探测。多伦多大学有一项长期的研究工作，专注于此类应用的越野自动驾驶，在向行业/政府合作伙伴进行出版物和技术转让方面，有经过验证的成果记录。例如，已向MDA Robotics和加拿大航天局(用于行星探索)、ClearPath Robotics(用于仓库和机场等户外物流)、无人机递送加拿大(用于无人驾驶飞行器交付)和电动绵羊机器人(用于大规模割草)提供了允许移动机器人自主操作的新算法。该提案要求配备传感器(立体摄像机、激光测距仪和雷达)的最先进的无人地面车辆(UGV)继续对这些重要的越野机器人应用进行尖端研究。要实现越野应用，还需要进行大量研究，包括对各种天气和地形的健壮，足够安全地与人类一起工作，以及足够高效地在具有挑战性的环境中快速移动。这样的平台允许频繁地进行本地试验，是向最终用户交付软件之前的关键垫脚石。这些实验不仅是科学过程中必要的一部分，而且还为学生、博士后和教职员工提供了重要的实践学习机会，以获得他们所需的技能，使加拿大保持在这一令人兴奋的领域的前沿。