Reproduce “Task aware Dynamic Movement Primitives for failure detection and recovery in contact rich manipulation” 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
Bhavnashri A, Sobia Shafi, Krishnapuram Himavarshini, Anuj Tiwari
The authors present a framework that learns peg-in-hole insertion from one teleoperated demonstration using Dynamic Movement Primitives, while monitoring execution with Quadratic Discriminant Analysis stage classification and force-based anomaly detection. When contact deviates, it pauses, searches in a spiral, and resumes. Stage accuracy was 91.8% ± 5.4% across six trials; recovery was reported for lateral offsets up to 3 mm in one setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.