Underwater robotic grasping is difficult due to degraded, highly variable imagery and the expense of collecting diverse underwater demonstrations. We introduce a system that (i) autonomously collects successful underwater grasp demonstrations via a self-supervised data collection pipeline and (ii) transfers grasp knowledge from on-land human demonstrations through a depth-based affordance representation that bridges the on-land-to-underwater domain gap and is robust to lighting and color shift.

We develop a compact long-reach manipulator with a deployable composite boom designed for autonomous cable outfitting and construction tasks on the lunar surface. We employ control strategies to mitigate deflection, vibration, and blossoming effects inherent to the deployable structure. Experiments demonstrate endpoint accuracy under 15 mm for boom lengths up to 1.8 m, enabling fine manipulation across large workspaces.

A small mobile robot equipped with a long-reach manipulator is studied for large-scale cleaning of lunar solar arrays. Endpoint force sensing and a velocity-based admittance control policy enables stable contact regulation (~2N), demonstrating autonomous maintenance capabilities.