Infrastructure for Scalable Robot Learning

Axis Robotics Platform

A Cross-Simulator, Cross-Machine Infrastructure for Scalable Teleoperation and Generalizable Robot Learning

Axis Robotics Team
Axis AI Platform Architecture

Axis Robotics Platform: A unified infrastructure bridging web-based teleoperation, GPU-accelerated realistic augmentation, and sim-to-real deployment.

Abstract

Robotic learning systems increasingly rely on large-scale demonstration data and realistic simulation for training and evaluation. However, existing workflows are often fragmented across simulators, operating systems, and compute environments: lightweight simulators enable broad access for teleoperation but lack high-fidelity rendering and physics, while GPU-based simulation stacks enable realism but are expensive and restrictive to access.

We present Axis Robotics Platform, a unified infrastructure that supports (i) multi-simulator development and execution, (ii) cross-machine orchestration spanning heterogeneous compute (web, Mac/Windows, and Linux GPU servers), and (iii) an end-to-end data pipeline from web-based teleoperation to photorealistic augmentation and downstream model training with sim-to-real deployment.

In Axis Robotics Platform, users teleoperate robots directly in a browser via a MuJoCo WebAssembly frontend, producing demonstrations without specialized hardware or compute. Demonstrations are then uploaded in a unified trajectory format and replayed on GPU-based Linux servers to generate realistic, domain-randomized rollouts using an IsaacSim-based augmentation backend. We further provide two production pipelines: (1) data cleaning and trajectory refinement, and (2) model training and sim-to-real evaluation, enabling rapid iteration from data to real-world deployment.

System Overview

System Overview

Platform Architecture

Explore the interconnected modules that make up the Axis Robotics Platform infrastructure, bridging lightweight data collection with heavy-duty simulation.

WASM UI
Task Generation
Resample Only
Smoothed & Resampled

Drag the slider to compare Resample Only vs. Smoothed & Resampled trajectory.

5x Speed

Trajectory Optimization Validation

Quantitative evaluation of trajectory optimization across LIBERO tasks. Our offline cleaning pipeline significantly reduces high-frequency artifacts (acceleration and jerk) from raw web-teleoperated data.

LIBERO Task Mean Acc. ↓ Mean Jerk ↓ Pos. Dev. (m) Removed Ratio
Before After (Ours) Before After (Ours)
Task 1: Place Black Bowl on Top of cabinet 0.2458 0.0738 2.6425 1.3363 0.0235 6.46%
Task 2: Place Rear Butter in Cabinet Top Drawer and Close It 0.3067 0.0943 3.7395 1.6208 0.0654 5.73%
Task 3: Place the black bowl on the plate 0.3559 0.1011 5.1358 2.3634 0.0201 7.17%
Task 4: Place the black bowl on top of the cabinet 0.3225 0.0859 4.6724 2.0785 0.0142 3.77%
Task 5: Place the frying pan on the stove 0.1832 0.1055 2.4771 1.1374 0.0106 1.24%
Task 6: Place the moka pot on the stove 0.1819 0.1090 2.5239 1.6784 0.0064 1.94%
Task 7: Turn on the stove 0.2689 0.1295 3.8291 2.3017 0.0142 3.81%
Task 8: Close Cabinet Bottom Drawer 0.1574 0.0741 1.9118 0.6659 0.0057 4.59%
Task 9: Place the black bowl into the cabinet’s bottom drawer 0.1551 0.0913 1.8665 1.3193 0.0161 2.03%
Task 10: Place Wine Bottle on Wine Rack 0.1841 0.1114 2.0298 1.0022 0.0625 4.86%
Task 11: Close Cabinet Top Drawer 0.1224 0.0683 1.5424 0.5822 0.0053 2.62%
Task 12: Place the black bowl into the cabinet’s top drawer 0.2109 0.1418 2.6997 2.1633 0.0961 11.4%
Task 13: Place Black Bowl on Plate 0.1682 0.0816 2.2262 1.1552 0.2511 2.14%
Task 14: Place the black bowl on top of the cabinet 0.1375 0.0736 1.8499 0.9924 0.0108 6.73%
Task 15: Place the right moka pot on the stove 0.1890 0.1130 2.5768 1.4929 0.0134 3.19%
Task 16: Turn off the stove 0.2149 0.1213 2.8268 1.5823 0.0219 8.93%
Average 0.2128 0.0985 2.7844 1.4670 0.0398 4.79%

BibTeX

@misc{AxisWebInfra,
  author = {{AxisAIOrg}},
  title = {AxisWebInfra},
  year = {2026},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/AxisAIOrg/AxisWebInfra}},
}