---
title: PRISM
canonical_url: "https://www.modelscope.cn/datasets/Turboyuuuu/PRISM"
md_url: "https://www.modelscope.cn/datasets/Turboyuuuu/PRISM.md"
repository: Turboyuuuu/PRISM
last_updated: 2026-09-29
license: "Apache License 2.0"
storage_size: "49 GB"
downloads: 22212
stars: 0
---

# PRISM

> PRISM - Turboyuuuu 在 ModelScope 开源的数据集。PRISM: Precision and contact-rich Real-world Industrial Skill Dataset with Multimodal Sensing

Turboyuuuu/PRISM 是 ModelScope 魔搭社区上的数据集，存储大小 49 GB，采用 Apache License 2.0 许可。

- **Repository**: Turboyuuuu/PRISM
- **License**: Apache License 2.0
- **Storage size**: 49 GB
- **Downloads**: 22212
- **Stars**: 0
- **Last updated**: 2026-09-29

Source: https://www.modelscope.cn/datasets/Turboyuuuu/PRISM

---

<p align="center">
  <strong>PRISM: Precision and contact-rich Real-world Industrial Skill Dataset with Multimodal Sensing</strong>
</p>

## Overview

Most existing robot learning datasets focus on short-horizon and low-contact manipulation tasks, such as pick-and-place. These datasets are often insufficient for industrial assembly and other contact-rich operations, where robots must handle tight tolerances, sustained contact, friction, insertion, alignment, and force-sensitive interactions.

PRISM addresses this gap by collecting diverse industrial manipulation demonstrations with synchronized multimodal sensing. The dataset includes multi-view RGB-D observations, force/torque measurements, tactile signals, and robot-state information, providing a realistic benchmark for learning contact-rich manipulation policies.

## Dataset Highlights

- **25+** industrial manipulation tasks
- **5,000+** robot trajectories
- **5,000+** paired human demonstrations
- **45+** hours of demonstrations
- **~27M** images across visual and visuotactile streams
- Multi-view **RGB-D** sensing
- **6DoF force/torque** sensing
- **Tactile** and proprioceptive observations
- Multiple robot platforms and teleoperation interfaces

## Citation

```
@misc{yu2026prismprecisioncontactrichrealworld,
      title={PRISM: Precision and contact-rich Real-world Industrial Skill dataset with Multimodal sensing}, 
      author={Tengbo Yu and Jiahao Wu and Hanning Wang and Rui Chen and Chuanhou Liu and Chuang Sun and Hangxin Liu},
      year={2026},
      eprint={2608.17962},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2608.17962}, 
}
```