technique

sim-to-real reinforcement learning

techniqueactiveprovisionalsim-to-real-reinforcement-learning-1e358c55·1 events·first seen 20d ago

Aliases: sim-to-real reinforcement learning

Co-occurring entities

multi-fingered dexterous hand Center-of-Pressure (CoP) tactile representation ball balancing multi-fingered robotic hand ball balancing task differentiable dynamics sensor calibration peg-in-hole insertion task

More like this (12)

sim-to-real transfer sim-to-real gap decoupled reinforcement learning shielded reinforcement learning self-play reinforcement learning reinforcement learning with belief-state rewards Entropy-Regularized Reinforcement Learning UniIntervene: Agentic Intervention for Efficient Real-World Reinforcement Learning curiosity-driven reinforcement learning rule-based reinforcement learning rewards reinforcement learning traffic smoothing Double Q-learning

Recent events (1)

5arXiv · cs.AI·20d ago·source ↗

Beyond Binary: Sim-to-Real Dexterous Manipulation with Physics-Grounded Contact Representation (CoP)

Researchers introduce Center-of-Pressure (CoP), a tactile representation grounded in physical principles designed to bridge the sim-to-real gap in contact-rich dexterous manipulation. CoP preserves dense contact information while remaining robust for sim-to-real transfer, supported by a differentiable-dynamics-based sensor calibration scheme that estimates taxel orientations without ground-truth force measurements. Evaluated on peg-in-hole insertion and ball balancing tasks, CoP-conditioned policies achieve zero-shot sim-to-real transfer on a multi-fingered robotic hand, outperforming binary-contact and raw-taxel baselines. An emergent finding is that CoP-conditioned policies implicitly encode task-relevant physical properties such as object mass.

Evaluation and Benchmarking Agent and Tool Ecosystem multi-fingered dexterous hand Center-of-Pressure (CoP) tactile representation ball balancing +5 more