Reproduce “On the Calibration of Force/Torque Sensors in Robotics” 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
Fredrik Bagge Carlson
This work by Fredrik Bagge Carlson develops methods for calibrating wrist-mounted force/torque sensors using only robot movements in free air, eliminating the need for specialized external equipment. The proposed techniques accurately estimate sensor kinematic and kinetostatic parameters, including the gravity vector, demonstrating robust convergence even with noisy measurements and poor initial conditions.
Key insights
- Force/torque sensor calibration can be effectively performed in 'free air' using only robot movements and orientation data, removing the dependency on external force application or specialized jigs.
- The relaxation-based method for known gravity concurrently estimates payload mass, mitigating errors that arise from incorrect prior mass estimates.
- Multiple distinct algorithmic approaches for calibrating with an unknown gravity vector (eigendecomposition, nullspace, iterative least-squares) were found to be mathematically equivalent and yield identical solutions, revealing a fundamental structure to the problem.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.