Abstract
Generative robot policies predict short action chunks but lack explicit long-horizon intent. Recent methods expose longer-horizon structure through language plans, subgoal images, or video forecasts, which are costly to generate and still need to be translated into robot motion. Predicting future robot motions avoids this translation, but a dense, time-indexed trajectory requires numerous parameters to cover the full remaining task, and over a short horizon it largely repeats the action chunk and adds little guidance for action generation. We propose Proprioceptive Action Models (PAM), which jointly generate a compact, timing-free sketch of the robot's remaining joint-space path and a dense executable action chunk within a single transformer denoiser. The sketch parameterizes the path by arc length rather than time, capturing geometric intent invariant to execution timing. Block-causal attention and a staggered denoising schedule maintain directed sketch-to-action dependence, ensuring the action tokens condition on a progressively cleaner sketch throughout sampling. In simulation, PAM improves over its action-only counterparts on Push-T and LIBERO-Long; on four real-world bimanual tasks, it raises success from 47.5% to 75.0%. Project page: https://nicehiro.github.io/pam_dp/
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Wang, F., Huang, S., Sun, H., Lyu, S., He, C., Duan, A., Zhou, P., & Navarro-Alarcon, D. (2026). Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies. https://omanscience.com/en/articles/proprioceptive-sketches-as-long-horizon-intent-for-generative-action-policies
MLA 9
Wang, Fangyuan, et al. "Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies." https://omanscience.com/en/articles/proprioceptive-sketches-as-long-horizon-intent-for-generative-action-policies.
Chicago (author–date)
Wang, Fangyuan, Songhao Huang, Haoxiang Sun, Shipeng Lyu, Chengyang He, Anqing Duan, Peng Zhou, and David Navarro-Alarcon. 2026. "Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies." https://omanscience.com/en/articles/proprioceptive-sketches-as-long-horizon-intent-for-generative-action-policies.
Harvard
Wang, F., Huang, S., Sun, H., Lyu, S., He, C., Duan, A., Zhou, P. and Navarro-Alarcon, D. (2026) 'Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies', Available at: https://omanscience.com/en/articles/proprioceptive-sketches-as-long-horizon-intent-for-generative-action-policies.
Vancouver
Wang F, Huang S, Sun H, Lyu S, He C, Duan A, et al. Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies. https://omanscience.com/en/articles/proprioceptive-sketches-as-long-horizon-intent-for-generative-action-policies
IEEE
F. Wang, S. Huang, H. Sun, S. Lyu, C. He, A. Duan, P. Zhou, and D. Navarro-Alarcon, "Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies," https://omanscience.com/en/articles/proprioceptive-sketches-as-long-horizon-intent-for-generative-action-policies.