Reproduce “Which Physical Tactile Cues Matter for Contact-Rich Imitation Learning? A Controlled Real-Robot Study” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xuhao Qin, Xiyan Huang, Haoyu Dong, Xi Zhang, Chenxi Xiao
The authors compare tactile representations for robot manipulation, deriving tactile RGB, surface geometry and indentation, and marker flow from one vision-based tactile sensor stream. They evaluate cue configurations alongside vision across three tabletop tasks using imitation learning. Full tactile input succeeds in 71/90 trials (78.9%), versus 52/90 (57.8%) for vision alone; results are limited to this sensor, gripper, and tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.