EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

DexAnyTwist: Learning general dexterous twisting with hybrid manipulation system identification

Xing Liu, Yunlong Dong, Jun Wan, Linan Deng, Feng Hua, Yi Shen, Min Yu, Guijun Ma, Cheng Cheng, Haitao Song, Han Ding, Ye Yuan

National Science Review · 09.06.2026

DOI: 10.1093/nsr/nwag351 ↗

Kurz gefasst

Fachartikel in der Zeitschrift National Science Review, online erschienen am 09.06.2026. Verfasst von Xing Liu und 11 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Developing general-purpose dexterous manipulation, particularly for intricate twisting tasks, remains a long-standing challenge. Large-scale training on massive datasets encompassing a vast spectrum of geometric and physical properties is a promising path. Yet, the hybrid dynamical space inherent in heterogeneous datasets creates optimization conflicts, where standard learning paradigms fail to synthesize a cohesive control strategy across disparate physical laws. This paper presents DexAnyTwist, a framework that reformulates the general twisting task as a problem of identification and control within a hybrid dynamical system. DexAnyTwist employs an iterative strategy to progressively partition the hybrid manipulation task into dynamically consistent subsystems. This approach utilizes forward partitioning to isolate data modes based on expert performance and backward refinement to optimize specialized policies. A learned gating mechanism then dynamically composes these experts, effectively mitigating interference between distinct dynamic modes. Beyond achieving state-of-the-art performance on our large-scale dataset, our framework spontaneously evolves manipulation primitives consistent with human twisting, and demonstrates superior zero-shot generalization to novel real-world geometries. This study not only solves a complex class of twisting tasks but also establishes a scalable pathway for robots to acquire general-purpose contact-rich manipulation skills.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Liu, X., Dong, Y., Wan, J., Deng, L., Hua, F., Shen, Y., Yu, M., Ma, G., Cheng, C., Song, H., Ding, H., & Yuan, Y. (2026). DexAnyTwist: Learning general dexterous twisting with hybrid manipulation system identification. National Science Review. https://doi.org/10.1093/nsr/nwag351

Vancouver

Liu X, Dong Y, Wan J, Deng L, Hua F, Shen Y, et al. DexAnyTwist: Learning general dexterous twisting with hybrid manipulation system identification. National Science Review. 2026. doi:10.1093/nsr/nwag351

BibTeX
@article{Liu2026,
  author = {Liu, Xing and Dong, Yunlong and Wan, Jun and Deng, Linan and Hua, Feng and Shen, Yi and Yu, Min and Ma, Guijun and Cheng, Cheng and Song, Haitao and Ding, Han and Yuan, Ye},
  title = {DexAnyTwist: Learning general dexterous twisting with hybrid manipulation system identification},
  journal = {National Science Review},
  year = {2026},
  doi = {10.1093/nsr/nwag351}
}
Kostenlos starten →