aiqbench analysis
DynaForge introduces a planning-guided residual learning framework specifically designed to generate high-quality demonstrations for dynamic robot manipulation tasks, addressing limitations in current static or simplified replay methods. By combining global planning with object-centric inverse kinematics and focusing reinforcement learning on evolving competence frontiers, the approach significantly improves demonstration success rates and enables effective sim-to-real transfer across multiple physical tasks. This advancement directly enhances the efficiency and robustness of policy learning for robots operating in complex, real-world environments where dynamic interactions are critical.
DynaForge: Planning-Guided Residual Learning for Dynamic Manipulation Demonstration Generation
arXiv:2609.25631v1 Announce Type: new Abstract: Dynamic object manipulation is essential for robots operating in real-world environments, yet methods for generating high-quality demonstrations remain limited. Methods designed for static tasks do not readily transfer to dynamic settings. Among dynamic demonstration generators, planning-based methods can fail near contact, while DOMINO-style replay simplifies dynamic interactions and may limit the experience available for policy learning. We present DynaForge, a planning-guided framework that learns residual corrections for dynamic manipulation demonstration generation. DynaForge combines low-frequency global planning with high-frequency object-centric inverse kinematics across task phases, and applies a residual policy to correct actions during dynamic interaction. An implicit curriculum groups rollouts under matched conditions and selects mixed-success groups, focusing residual reinforcement learning on the evolving competence frontier. On Can and Bottle, it uses 0.73x as many optimizer steps as vanilla GRPO at the same nominal environment-step budget, with higher observed final success rates. Across nine simulation tasks, DynaForge increases mean demonstration-generation success from 41.30% of the planning prior to 78.37%. With 800 demonstrations per task, DP3 policies trained on DynaForge data achieve 49.11% mean success, compared with 7.07% for DOMINO data. On three real-world dynamic tasks, DynaForge-trained policies achieve 30-60% success, compared with 0-10% for DOMINO-trained policies, showing the ability of DynaForge for sim-to-real transfer.
Abstract:Dynamic object manipulation is essential for robots operating in real-world environments, yet methods for generating high-quality demonstrations remain limited. Methods designed for static tasks do not readily transfer to dynamic settings. Among dynamic demonstration generators, planning-based methods can fail near contact, while DOMINO-style replay simplifies dynamic interactions and may limit the experience available for policy learning. We present DynaForge, a planning-guided framework that learns residual corrections for dynamic manipulation demonstration generation. DynaForge combines low-frequency global planning with high-frequency object-centric inverse kinematics across task phases, and applies a residual policy to correct actions during dynamic interaction. An implicit curriculum groups rollouts under matched conditions and selects mixed-success groups, focusing residual reinforcement learning on the evolving competence frontier. On Can and Bottle, it uses 0.73x as many optimizer steps as vanilla GRPO at the same nominal environment-step budget, with higher observed final success rates. Across nine simulation tasks, DynaForge increases mean demonstration-generation success from 41.30% of the planning prior to 78.37%. With 800 demonstrations per task, DP3 policies trained on DynaForge data achieve 49.11% mean success, compared with 7.07% for DOMINO data. On three real-world dynamic tasks, DynaForge-trained policies achieve 30-60% success, compared with 0-10% for DOMINO-trained policies, showing the ability of DynaForge for sim-to-real transfer.
8 pages, 6 figures. Under review
Robotics (cs.RO)
arXiv:2609.25631 [cs.RO]
(or arXiv:2609.25631v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.25631 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 303 words
Related coverage
The Machines that Make the Machines
NVIDIA · Oct 7, 2026
DROM: A Language-Guided Diffusion Framework for Multi-Skill Robotic Manipulation
arXiv · Sep 30, 2026
NVIDIA Research Advances Robotics From Simulation to the Real World
NVIDIA · May 28, 2026
ACDC: Adaptive Curriculum Planning with Dynamic Contrastive Control for Goal-Conditioned Reinforcement Learning in Robotic Manipulation
arXiv · Aug 4, 2026