The Mini Wheelbot Dataset: High-Fidelity Data for Robot Learning

2026-01-16

AI Summary

This work addresses the limited availability of real-world data and high hardware barriers in learning-based control for unstable systems by introducing a high-fidelity, multimodal dataset centered on the open-source Mini Wheelbot—a self-balancing single-wheel robot. The dataset features 1 kHz synchronized recordings of sensor measurements, state estimates, motion-capture ground-truth poses, and third-person video, spanning multiple hardware units, floor surfaces, and control strategies—including nonlinear model predictive control and reinforcement learning. Data collection employs pseudo-random binary excitation signals and high-precision motion capture for both generation and validation. As the first open-source, high-frequency benchmark dataset that spans diverse hardware and environmental conditions for balancing robots, it has already enabled reproducible research in dynamics modeling, state estimation, and time-series classification.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

The development of robust learning-based control algorithms for unstable systems requires high-quality, real-world data, yet access to specialized robotic hardware remains a significant barrier for many researchers. This paper introduces a comprehensive dynamics dataset for the Mini Wheelbot, an open-source, quasi-symmetric balancing reaction wheel unicycle. The dataset provides 1 kHz synchronized data encompassing all onboard sensor readings, state estimates, ground-truth poses from a motion capture system, and third-person video logs. To ensure data diversity, we include experiments across multiple hardware instances and surfaces using various control paradigms, including pseudo-random binary excitation, nonlinear model predictive control, and reinforcement learning agents. We include several example applications in dynamics model learning, state estimation, and time-series classification to illustrate common robotics algorithms that can be benchmarked on our dataset.

Problem

• robot learning
• unstable systems
• high-fidelity data
• dataset accessibility
• balancing robot

Innovation

• high-fidelity dataset
• robot learning
• balancing reaction wheel unicycle
• synchronized multi-modal data
• open-source robotics