Towards Space-Based Environmentally-Adaptive Grasping

TL;DR

This paper presents a space-based robotic grasping system that learns control policies in a structured latent space, achieving high success rates efficiently and demonstrating robustness to environmental variations.

Contribution

It introduces a latent manifold approach for space manipulation tasks, enabling faster, more robust learning compared to traditional visual methods.

Findings

Achieved over 95% success in space grasping tasks with less than 1 million steps.

Faster convergence than state-of-the-art visual baselines.

Latent space reasoning improves robustness to environmental variations.

Abstract

Robotic manipulation in unstructured environments requires reliable execution under diverse conditions, yet many state-of-the-art systems still struggle with high-dimensional action spaces, sparse rewards, and slow generalization beyond carefully curated training scenarios. We study these limitations through the example of grasping in space environments. We learn control policies directly in a learned latent manifold that fuses (grammarizes) multiple modalities into a structured representation for policy decision-making. Building on GPU-accelerated physics simulation, we instantiate a set of single-shot manipulation tasks and achieve over 95% task success with Soft Actor-Critic (SAC)-based reinforcement learning in less than 1M environment steps, under continuously varying grasping conditions from step 1. This empirically shows faster convergence than representative state-of-the-art visual baselines under the same open-loop single-shot conditions. Our analysis indicates that explicitly reasoning in latent space yields more sample-efficient learning and improved robustness to novel object and gripper geometries, environmental clutter, and sensor configurations compared to standard baselines. We identify remaining limitations and outline directions toward fully adaptive and generalizable grasping in the extreme conditions of space.