Simple Torque-Observation Alignment for Zero-Shot Sim-to-Real Grasping with a Direct-Drive Gripper

2026-09-24

AI Summary

This study addresses the domain discrepancies in scale, offset, and noise between simulated and real-world torque observations in reinforcement learning. We propose a sensor-free torque alignment method that exploits the inherent characteristics of direct-drive motors. Specifically, linear mapping calibration is employed to correct scale deviations, a differential observation strategy eliminates systematic offsets, and Gaussian noise injection enhances robustness. Furthermore, a teacher-student distillation framework is introduced to facilitate policy transfer. Experimental results demonstrate that the proposed approach achieves zero-shot sim-to-real transfer with a 100% grasping success rate across nine in-distribution objects, significantly improving the deployment reliability of direct-drive grippers in real-world physical environments.

Abstract

Torque observations in reinforcement learning remain challenging because simulated and measured torque differ in scale, offset, and noise. In this paper, we propose a simple torque observation alignment method for robots with direct-drive (DD) actuators, in which motor current maps linearly to joint torque through a motor-type-specific torque constant K_tau. First, dynamometer calibration identifies K_tau* and corrects the scale mismatch between simulated and real torque. Second, the method uses delta_tau(t) = tau(t) - tau(t-1) as the observation in both domains to eliminate the constant offset instead of using the direct torque tau(t), which carries a domain-dependent bias. Third, Gaussian noise obtained from the dynamometer measurement data is injected during the learning process. To validate the proposed method, we train a teacher-student grasping policy entirely in simulation and deploy the distilled student on a multifingered DD gripper. The deployed policy performs proprioceptive grasping using only joint positions and torque differences. We conduct an ablation study comparing the proposed method with alternative alignment variants on nine in-distribution (ID) objects. The proposed method achieves 100% grasp success. These results demonstrate that the proposed alignment method improves the robustness of zero-shot policy transfer on the DD gripper against real-world torque-observation mismatches.

Problem

• Torque observation
• Sim-to-real transfer
• Reinforcement learning
• Direct-drive gripper
• Zero-shot

Innovation

• Torque-Observation Alignment
• Zero-Shot Sim-to-Real
• Direct-Drive Gripper
• Reinforcement Learning
• Teacher-Student Policy