Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.17513 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913045406547968 |
|---|---|
| author | Luo, Siyuan Zhou, Bingyang Zhang, Chong Liu, Xin Huang, Zhenhao Yang, Gang Han, Zhengtao Hu, Xiaotian Yang, Eric Yu, Rymon Zeng, Ziqiu Shi, Fan |
| author_facet | Luo, Siyuan Zhou, Bingyang Zhang, Chong Liu, Xin Huang, Zhenhao Yang, Gang Han, Zhengtao Hu, Xiaotian Yang, Eric Yu, Rymon Zeng, Ziqiu Shi, Fan |
| contents | Simulation frameworks such as Isaac Sim have enabled scalable robot learning for locomotion and rigid-body manipulation; however, contact-rich simulation remains a major bottleneck for deformable object manipulation. The continuously changing geometry of soft materials, together with large numbers of vertices and contact constraints, makes it difficult to achieve high accuracy, speed, and stability required for large-scale interactive learning. We present FLASH, a GPU-native simulation framework for contact-rich deformable manipulation, built on an accurate NCP-based solver that enforces strict contact and deformation constraints while being explicitly designed for fine-grained GPU parallelism. Rather than porting conventional single-instruction-multiple-data (SIMD) solvers to GPUs, FLASH redesigns the physics engine from the ground up to leverage modern GPU architectures, including optimized collision handling and memory layouts. As a result, FLASH scales to over 3 million degrees of freedom at 30 FPS on a single RTX 5090, while accurately simulating physical interactions. Policies trained solely on FLASH-generated synthetic data in minutes achieve robust zero-shot sim-to-real transfer, which we validate on physical robots performing challenging deformable manipulation tasks such as towel folding and garment folding, without any real-world demonstration, providing a practical alternative to labor-intensive real-world data collection. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17513 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | FLASH: Fast Learning via GPU-Accelerated Simulation for High-Fidelity Deformable Manipulation in Minutes Luo, Siyuan Zhou, Bingyang Zhang, Chong Liu, Xin Huang, Zhenhao Yang, Gang Han, Zhengtao Hu, Xiaotian Yang, Eric Yu, Rymon Zeng, Ziqiu Shi, Fan Robotics Simulation frameworks such as Isaac Sim have enabled scalable robot learning for locomotion and rigid-body manipulation; however, contact-rich simulation remains a major bottleneck for deformable object manipulation. The continuously changing geometry of soft materials, together with large numbers of vertices and contact constraints, makes it difficult to achieve high accuracy, speed, and stability required for large-scale interactive learning. We present FLASH, a GPU-native simulation framework for contact-rich deformable manipulation, built on an accurate NCP-based solver that enforces strict contact and deformation constraints while being explicitly designed for fine-grained GPU parallelism. Rather than porting conventional single-instruction-multiple-data (SIMD) solvers to GPUs, FLASH redesigns the physics engine from the ground up to leverage modern GPU architectures, including optimized collision handling and memory layouts. As a result, FLASH scales to over 3 million degrees of freedom at 30 FPS on a single RTX 5090, while accurately simulating physical interactions. Policies trained solely on FLASH-generated synthetic data in minutes achieve robust zero-shot sim-to-real transfer, which we validate on physical robots performing challenging deformable manipulation tasks such as towel folding and garment folding, without any real-world demonstration, providing a practical alternative to labor-intensive real-world data collection. |
| title | FLASH: Fast Learning via GPU-Accelerated Simulation for High-Fidelity Deformable Manipulation in Minutes |
| topic | Robotics |
| url | https://arxiv.org/abs/2604.17513 |