Closing the Reality Gap: Zero-Shot Sim-to-Real Deployment for Dexterous Force-Based Grasping and Manipulation

2026-07-06

AI Summary

This work addresses the challenge of transferring dexterous manipulation policies trained in simulation directly to real-world five-fingered hands, which is hindered by dense contact interactions and non-ideal actuation characteristics. The authors propose a reinforcement learning framework that integrates dense tactile sensing with joint torque perception to achieve zero-shot sim-to-real transfer. Key innovations include parallel forward-kinematics-based tactile simulation, precise calibration from motor currents to joint torques, randomized modeling of actuator dynamics, and an asymmetric Actor-Critic Proximal Policy Optimization (PPO) algorithm operating in a fused tactile-torque observation space. Without any fine-tuning, the method successfully enables controlled grip-force tracking and in-hand object reorientation on a physical dexterous hand, demonstrating human-like robust manipulation capabilities.

Abstract

Human-like dexterous hands with multiple fingers offer human-level manipulation capabilities but remain difficult to train the control policies that can deploy on real hardware due to contact-rich physics and imperfect actuation. We present a sim-to-real reinforcement learning method that leverages dense tactile feedback combined with joint torque sensing to explicitly regulate physical interactions. To enable effective sim-to-real transfer, we introduce (i) a computationally fast tactile simulation that computes distances between dense virtual tactile units and the object via parallel forward kinematics, providing high-rate, high-resolution touch signals needed by RL; (ii) a current-to-torque calibration that eliminates the need for torque sensors on dexterous hands by mapping motor current to joint torque; and (iii) actuator dynamics modeling with randomization to account for non-ideal torque-speed effects and bridge the actuation gaps. Using an asymmetric actor-critic PPO pipeline, we train policies entirely in simulation and deploy them directly to a five-finger hand. The resulting policies demonstrate two essential human-hand skills: (1) command-based controllable grasp force tracking and (2) reorientation of objects in the hand, both of which are robustly executed without fine-tuning on the robot. By combining tactile and torque in the observation space with scalable sensing and actuation modeling, our system provides a practical solution to achieve reliable dexterous manipulation. To our knowledge, this is the first demonstration of controllable grasping on a multi-finger dexterous hand trained entirely in simulation and transferred zero-shot on real hardware.

Problem

• dexterous manipulation
• sim-to-real transfer
• force-based grasping
• tactile feedback
• zero-shot deployment

Innovation

• sim-to-real
• dexterous manipulation
• tactile simulation
• torque estimation
• zero-shot transfer