Within Arms' Reach: Infrastructure for Research in Support of the Canadarm3 Program

加拿大在空间机器人方面有着悠久的开创性贡献，从1981年美国宇航局的航天飞机（太空运输系统，或STS）上首次发射的原始加拿大手臂机器人机械手开始，这是STS计划成功的关键。加拿大2号是一种更有能力和更复杂的双手操纵器，于2001年安装在国际空间站上，并继续提供必要的服务和停泊支持。加拿大在太空的下一步将把我们从地球轨道带到月球——加拿大3号将服务于月球门户，月球轨道上的一个空间站。Canadarm3和月球门户代表了加拿大空间机器人技术的潜在复兴，并有机会推进加拿大在这一关键领域的领导地位。

Canada has a long history of seminal contributions to space robotics beginning with the original Canadarm robotic manipulator, which first launched on NASA's Space Shuttle (Space Transportation System, or STS) in 1981 and was key to the success of the STS program. The Canadarm2, a more capable and sophisticated bimanual (dual-hand) manipulator, was installed on board the International Space Station in 2001, and continues to provide essential servicing and berthing support. Canada's next step in space will take us beyond Earth orbit to the Moon - the Canadarm3 will service the Lunar Gateway, a space station in lunar orbit. Canadarm3, and the Lunar Gateway, represent a potential renaissance in Canadian space robotics and an opportunity to advance Canada's leadership in this key field. The original Canadarm, and the Canadarm2, were designed to be tele-operated by astronauts on-orbit or by ground controllers and hence had very limited autonomy capabilities. The arm destined for the Lunar Gateway, however, will require a new generation of autonomy algorithms to be developed. Gateway is expected to be unstaffed for long periods - the round-trip travel time for signals from Earth to lunar orbit will be too long to make safe tele-operation feasible. In turn, advanced autonomy solutions will be needed to enable the Canadarm3 to operate without direct human guidance. This proposal is for the acquisition of a bimanual mobile manipulator, or a robot that incorporates a mobile base and two manipulator arms with specialized grippers (hands). This system will replicate, on a smaller scale, the mobility and dexterity of the bimanual Canadarm3. Such a testbed will be critical to developing the advanced autonomy necessary for on-orbit operations where the emphasis will be on ensuring safety despite imperfect models, scene clutter, and challenging lighting conditions; these conditions can be mocked up with the requested platform to validate autonomy algorithms on the ground. The requested equipment will enable key Canadarm3-related research projects, across four robotics groups at the University of Toronto Institute for Aerospace Studies, including: hybrid motion planning for mobile manipulation; model learning, anomaly detection, and performance improvement for flexible manipulators; on-orbit localization and mapping, and deep learning for worksite change detection. The mobile manipulator acquired through this grant will be a unique hardware platform in Canada and comparable to the best hardware at just a handful of laboratories worldwide. Using the requested equipment, the co-applicants' research will move beyond the current state of the art, solving problems that are essential to the Canadarm3 program and that have yet to be fully addressed by the community. The findings will also benefit several existing and potential terrestrial partner companies nationally who are interested in mobile manipulation for warehousing, services, and healthcare, for example.