Description: NASA's Moon-to-Mars roadmap calls for robotic systems which can complement and eventually replace humans for both routine and unplanned assembly, maintenance and inspection tasks for orbital and lunar surface infrastructure. The proposed innovation is an integrated perception and planning system to enable safe and productive interaction with cables and fluid lines, a ubiquitous element in any spacecraft or ground habitat, without human oversight. Handling cables and fluid lines involves a wide array of motion primitives and configurations. For example, they may need to be pulled aside to allow others to be installed, for panels to be inspected, or to unblock a passageway. Perhaps a gentle tug is necessary to confirm the connector it is linked to or to validate its stiffness before moving it more forcefully. As flexible materials, they settle in unpredictable configurations, and robots need to be able to adapt to what they see in the moment, rather than memorizing precomputed response patterns. Funding will be put towards developing software for four key capabilities required to safely interact with cables and fluid lines: view planning, detection, mapping and manipulation planning. Development will occur mainly in simulation and then a representative end-to-end scenario will be demonstrated in a lab environment. This technology's space customers are NASA and other space agencies, as well as commercial space operators. It is intended to be deployed as a tech demo on Astrobee onboard the International Space Station. L5 would offer "Maintenance as a Service" at a rate which would drastically undercut the equivalent human labor rate. For Earth applications, it would support L5's ongoing strawberry harvesting activity, specifically for the perception and manipulation of tangled strawberry stems, whose occlusion can prevent ripe specimens from being picked.

Benefits: In the short term, this technology is intended to be deployed on Astrobee on the International Space Station (ISS) as a technology demonstration. By using this existing asset, we can gain operational experience in a typical space environment, which can in turn help inform the design of Gateway assembly, maintenance and inspection elements. There is a window of 2-4 years where it could be matured beyond the demo stage for operational use onboard the ISS, offloading routine tasks from astronauts and allowing more productive use of their time. Once the ISS is decommissioned, as the technology is platform-agnostic, it can be ported to whatever hardware is used for future orbital or lunar infrastructure, for both initial assembly and ongoing maintenance. In both the short and long term, it will support and enable tasks including moving safely through cluttered environments, inspecting occluded panels and installing and reconfiguring wired equipment. While the initial effort is for internal habitat inspection and maintenance, similar tasks on the exterior of habitats can also be accomplished. The same benefits as for NASA space assets can equally apply to other space agencies and commercial space station operators. Commercial stations under development by Axiom Space, Vast and Blue Origin, which are due to launch in the late 2020s, as well as assets from ESA and JAXA, will likely need a similar level and type of upkeep to the ISS and NASA's Gateway assets. Beyond space, the techniques used for cables and fluid lines can transfer to other thin flexible elements. L5's ongoing development efforts are in automated strawberry harvesting, for which a semantic understanding of strawberry stems and the ability to move them out of the way would enable more complex field conditions to be handled, increasing the yield per harvester. Similar motion primitives for detecting and manipulating stems would also be applicable in orange, avocado, blueberry or artichoke harvesting or orchard management. Beyond agriculture, there is overlap with the techniques used to inspect and maintain undersea cables and above ground power lines, which are among the most dangerous occupations.