Skip Navigation
Center Innovation Fund: LaRC CIF

Autonomous Assembly of Solar Array Modules by a Team of Robots, Year 2

Completed Technology Project
28 views

Project Description

Autonomous Assembly of Solar Array Modules by a Team of Robots, Year 2

Intelligent Precision Jigging Robots (IPJRs) enable a paradigm for ISA that can drastically reduce mission risk and cost while increasing performance. The IPJR paradigm uses: 1) gross positioning using long reach manipulation, and 2) localized, high precision positioning (``jigging'') to support joining. IPJRs provide the latter capabilities, enabling the structure to have simpler components, and reducing the number of tasks that a long reach manipulator needs to accomplish. By reducing the required complexity of the structural elements, IPJRs are more versatile, enabling reuse and lowering costs over time. Each of these features is unique in the field of robotic assembly, and has been previously explored by NASA and the University of Colorado Boulder in the context of a NASA Space Technology Research Fellowship (NSTRF) by the PI. This work also supports a recently-awarded joint NASA-Orbital/ATK TP project. This effort is investigating whether IPJRs and manipulators can assemble, disassemble, and reassemble solar arrays to high precision. ISA of solar arrays has several benefits, including: solar electric propulsion, and reuse/refurbishment missions with components that were not designed for ISA. This effort builds on the results of the 2016 fiscal year - design and construction of IPJR, long reach manipulator, and solar array module prototypes - and will utilize these prototypes for machine learning and state estimation research. The ultimate goal is to advance the state-of-the-art in robust assembly algorithms, utilizing advances in state estimation, object recognition, and solutions to partially-observable Markov decision processes (POMDPs). Assembly is at the intersection of several robotics topics, including simultaneous localization and mapping, path planning, error correction, reinforcement learning, and state classification using supervised and unsupervised learning. Progress and methods developed here are expected to have widespread applicability due to the generality of the robotics problems being addressed, and with respect to assembly, the ubiquity of welding and other general joining techniques in construction.

More »

Anticipated Benefits

Primary U.S. Work Locations and Key Partners

Technology Transitions

Light bulb

Suggest an Edit

Recommend changes and additions to this project record.
^