Scaling data volume and diversity is critical for generalizing embodied intelligence. While synthetic data generation offers a scalable alternative to expensive physical data acquisition, transferring robotic manipulation policies from simulation to the real world (sim-to-real) remains a formidable challenge due to the domain gap. This paper presents HyperSim, a holistic framework spanning from synthetic data generation to policy training and seamless real-world deployment. To systematically bridge the sim-to-real gap, HyperSim is realized through three core pillars: high-fidelity environment synthesis, adversarial trajectory generation, and sim-and-real co-training. Collectively, these modules address domain discrepancies by enhancing visual fidelity, expanding data coverage, and enforcing domain-invariant representations. We rigorously validate HyperSim through a large-scale empirical study involving 400 real-world task executions across two representative manipulation models. Assessed across three fine-grained metrics, our complete pipeline achieves remarkable sim-to-real success rates of 80% and 95% with ACT and \pi_{0}, respectively. Furthermore, policies trained on our adversarial trajectories exhibit significantly enhanced robustness against dynamic uncertainties, achieving a 35% higher completion rate under physical perturbations.
@article{arxiv.2605.26638,
title = {HyperSim: A Holistic Sim-To-Real Framework For Robust Robotic Manipulation},
author = {Junyi Dong and Haotian Luo and Ziwei Xu and Shengwei Bian and Heng Zhang and Sitong Mao and Jingyi Guo and Yang Xu and Wenhao Chen and Qiuyu Feng and Yao Mu and Ping Luo and Shunbo Zhou and Xiaodong Wu},
journal= {arXiv preprint arXiv:2605.26638},
year = {2026}
}