aiqbench analysis
This research introduces a reinforcement learning framework that enables robots to perform precise peg-in-hole insertions directly from simulation, achieving perfect benchmark scores without real-world fine-tuning. By leveraging 3D force feedback and specialized reward structures, the system successfully generalizes across diverse geometries and tight clearances, significantly reducing collision risks. This advancement marks a critical step toward scalable, autonomous robotic assembly where complex manipulation skills can be trained virtually and deployed immediately in physical environments.
InsertAnything: Generalizable Contact-Rich Precision Insertion from Simulation to Reality
arXiv:2609.24511v1 Announce Type: new Abstract: Contact-rich precision insertion is a key manipulation skill in robotic assembly. Tight clearances make insertion more sensitive to alignment errors and prone to collisions and jamming, while variations in geometry and clearance across parts further complicate policy reuse. We present a reinforcement learning framework that trains insertion policies entirely in simulation for direct deployment without real-world demonstrations or policy fine-tuning. By combining target poses with compact three-dimensional fingertip force feedback, the policy learns to search for alignment and correct its motion despite errors in the estimated hole position. A decoupled gated reward coordinates alignment and insertion. Force-signal smoothing and state-independent standard deviations stabilize the learning process. The resulting policies perform real-world insertion across multiple hole geometries with a minimum nominal clearance of 0.02 mm and improve success while reducing peak contact forces under hole-position errors. Cross-clearance and cross-geometry evaluations further confirm policy generalization. The system achieved the first perfect score of 20/20 on ManipulationNet's peg-in-hole benchmark under its Human-in-the-Loop protocol, with fully autonomous insertion motions. A single policy trained only on a simulated hexagonal insertion task achieved an overall success rate of 95.0% across eight unseen real-world insertion tasks. These results show that learning entirely in simulation can yield precision insertion skills that can be deployed directly and reused across real-world tasks. The project website (https://mzhsoul.github.io/InsertAnything/) provides open-source simulation and real-robot experiment scripts, assets, and trained checkpoints.
Abstract:Contact-rich precision insertion is a key manipulation skill in robotic assembly. Tight clearances make insertion more sensitive to alignment errors and prone to collisions and jamming, while variations in geometry and clearance across parts further complicate policy reuse. We present a reinforcement learning framework that trains insertion policies entirely in simulation for direct deployment without real-world demonstrations or policy fine-tuning. By combining target poses with compact three-dimensional fingertip force feedback, the policy learns to search for alignment and correct its motion despite errors in the estimated hole position. A decoupled gated reward coordinates alignment and insertion. Force-signal smoothing and state-independent standard deviations stabilize the learning process. The resulting policies perform real-world insertion across multiple hole geometries with a minimum nominal clearance of 0.02 mm and improve success while reducing peak contact forces under hole-position errors. Cross-clearance and cross-geometry evaluations further confirm policy generalization. The system achieved the first perfect score of 20/20 on ManipulationNet's peg-in-hole benchmark under its Human-in-the-Loop protocol, with fully autonomous insertion motions. A single policy trained only on a simulated hexagonal insertion task achieved an overall success rate of 95.0% across eight unseen real-world insertion tasks. These results show that learning entirely in simulation can yield precision insertion skills that can be deployed directly and reused across real-world tasks. The project website (this https URL) provides open-source simulation and real-robot experiment scripts, assets, and trained checkpoints.
Robotics (cs.RO)
arXiv:2609.24511 [cs.RO]
(or arXiv:2609.24511v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.24511 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 · 322 words
Related coverage
Force-Aware Residual DAgger via Trajectory Editing for Precision Insertion with Impedance Control
arXiv · Jul 24, 2026
Amazon and University of Michigan give robots a sense of touch
Amazon Science · Jul 10, 2026
Impedance Control of Ship-Borne Manipulators via Optimization-based Task-Space Inverse Dynamics
arXiv · Jul 27, 2026
RouteRLT: Learning When and Which RL Specialist Should Control a Vision-Language-Action Policy
arXiv · Sep 24, 2026