Domain-Adaptive Communication-Rate Optimization for Sim-to-Real Humanoid-Robot Wireless XR Teleoperation

2026-05-18

AI Summary

This work addresses the excessive communication overhead caused by high-frequency motion transmission in wireless XR-based humanoid teleoperation, which hinders practical deployment. The authors propose an end-to-end system framework that optimizes communication efficiency through dimension-wise sampling rate control, significantly reducing energy consumption while preserving motion trajectory reconstruction accuracy. To mitigate distribution shifts in sim-to-real transfer, they introduce a PAC-Bayes generalization bound analysis that elucidates the impacts of latent-space density ratio estimation, finite-sample bias, and encoder bias. Guided by this analysis, they design a PPO algorithm integrating density-ratio weighting and trust-region regularization. Experiments demonstrate that the method effectively balances reconstruction error and communication cost on public humanoid teleoperation datasets and exhibits robustness and generalization across diverse wireless channels and dynamic trajectories.

Abstract

Wireless extended reality (XR) teleoperation provides embodied interaction capability for collecting humanoid robot demonstrations, but the large-scale adoption is restricted by the overhead of high-frequency motion transmission. This paper develops a system framework that integrates sampling, transmission, interpolation, and reconstruction and formulates a communication-rate optimization that aims to minimize the communication energy while maintaining the reconstruction accuracy of robot motion trajectories through dimension-wise sampling-rate control. Since acquiring real-time feedback from physical robots is limited by hardware costs, it is necessary to solve the problem through simulator interaction with offline real-domain data correction. To guide sim-to-real adaptation, we provide a PAC-Bayes generalization characterization that reveals the effects of latent density-ratio estimation, finite-sample deviation, and encoder bias. Building on this analysis, we propose a proximal policy optimization (PPO) method with density-ratio weighting and trust-region regularization. Experiments on public humanoid teleoperation dataset show that the proposed method improves the tradeoff between reconstruction error and communication energy consumption under sim-to-real distribution shift. We further analyze the effectiveness of the proposed algorithm across various wireless channels and dynamic motion trajectories.

Problem

• sim-to-real
• communication-rate optimization
• wireless XR teleoperation
• humanoid robot
• distribution shift

Innovation

• domain adaptation
• communication-rate optimization
• sim-to-real transfer
• PAC-Bayes generalization
• proximal policy optimization