aiqbench analysis
MicroPush offers a lightweight simulation environment specifically designed to accelerate the training of reinforcement learning policies for microrobotic manipulation tasks like pushing and assembly. By enabling zero-shot transfer from simulation to physical hardware, this tool addresses the critical bottleneck of slow and variable physical trials in developing autonomous micro-robotic systems. This advancement provides an efficient substrate for researchers aiming to scale up learning-based autonomy in constrained robotic domains.
MicroPush: A Lightweight Simulator for Sim-to-Real Microrobotic Pushing and Assembly
arXiv:2602.23607v2 Announce Type: replace Abstract: Developing learning-based autonomy for microrobotic manipulation is difficult because physical trials are slow, variable, and poorly suited to large-scale training. We present MicroPush, a lightweight simulation and benchmarking environment for pushing and assembly with magnetic rolling microrobots. The simulator combines overdamped planar dynamics with calibrated rolling actuation, reduced-order contact handling, short-range interaction corrections, and optional background flow, while providing common interfaces for planning, model-based control, reinforcement learning (RL), interactive visualization, and headless evaluation. In no-render tests, MicroPush remains substantially faster than real time, reaching 37.0x real time with 100 passive particles. We further evaluate the simulator through zero-shot sim-to-real policy transfer. A Residual RL-MPC policy is trained entirely in MicroPush, frozen after simulation training, and deployed on the physical system without real-world policy fine-tuning. A 20 {\mu}m magnetic microrobot pushes a similarly sized silica particle along a figure-eight trajectory with 1.91 {\mu}m median and 4.65 {\mu}m 95th-percentile path error, and the same autonomy stack performs sequential line, triangle, and six-particle hexagonal assembly. These results show that a deliberately reduced-order simulator can provide an efficient training and development substrate for sim-to-real microrobotic autonomy.
Abstract:Developing learning-based autonomy for microrobotic manipulation is difficult because physical trials are slow, variable, and poorly suited to large-scale training. We present MicroPush, a lightweight simulation and benchmarking environment for pushing and assembly with magnetic rolling microrobots. The simulator combines overdamped planar dynamics with calibrated rolling actuation, reduced-order contact handling, short-range interaction corrections, and optional background flow, while providing common interfaces for planning, model-based control, reinforcement learning (RL), interactive visualization, and headless evaluation. In no-render tests, MicroPush remains substantially faster than real time, reaching 37.0x real time with 100 passive particles. We further evaluate the simulator through zero-shot sim-to-real policy transfer. A Residual RL-MPC policy is trained entirely in MicroPush, frozen after simulation training, and deployed on the physical system without real-world policy fine-tuning. A 20 {\mu}m magnetic microrobot pushes a similarly sized silica particle along a figure-eight trajectory with 1.91 {\mu}m median and 4.65 {\mu}m 95th-percentile path error, and the same autonomy stack performs sequential line, triangle, and six-particle hexagonal assembly. These results show that a deliberately reduced-order simulator can provide an efficient training and development substrate for sim-to-real microrobotic autonomy.
Robotics (cs.RO); Systems and Control (eess.SY)
arXiv:2602.23607 [cs.RO]
(or arXiv:2602.23607v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2602.23607 Focus to learn more arXiv-issued DOI via DataCite
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 · 277 words
Related coverage

The Machines that Make the Machines
NVIDIA · Oct 7, 2026

Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AWS · Sep 4, 2026
News Archive 2026
Quantum Computing Report · Jul 25, 2026

NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA · May 21, 2026