1Peking University
2Galbot
3Beijing Academy of Artificial Intelligence
* equal contributions
† corresponding author
Left: Our pipeline siginificantly outperforms analytic-based dexterous grasp synthesis baselines on almost all metrics in the simulation benchmark. Right-top: Our synthesized large-scale, high-quality grasp dataset. Right-bottom: Real-world deployment of the learning-based model trained on our dataset.
Robotic dexterous grasping is important for interacting with the environment. To unleash the potential of data-driven models for dexterous grasping, a large-scale, high-quality dataset is essential. While gradient-based optimization offers a promising way for constructing such datasets, previous works suffer from limitations, such as inefficiency, strong assumptions in energy design, or limited object sets for experiments. Moreover, the lack of a standard benchmark for comparing different methods and datasets hinders progress in this field. To address these challenges, we develop a highly efficient synthesis system and a comprehensive benchmark with MuJoCo for dexterous grasping. We formulate grasp synthesis as a bilevel optimization problem, combining a novel lower-level quadratic programming (QP) with an upper-level gradient descent process. By leveraging recent advances in CUDA-accelerated robotic libraries and GPU-based QP solvers, our system can parallelize thousands of grasps and synthesize over 49 grasps per second on a single 3090 GPU. Our synthesized grasps for both Shadow and Allegro hand achieve a success rate above 75% in simulation, with a penetration depth and contact distance of under 1 mm, outperforming existing baselines on nearly all metrics. Compared to the previous large-scale dataset, DexGraspNet, our dataset significantly improves the performance of learning models, with a success rate from around 40% to 80% in simulation. Real-world testing of the trained model on the Shadow Hand achieves an 81% success rate across 20 diverse objects.
@article{chen2024bodex,
title={BODex: Scalable and Efficient Robotic Dexterous Grasp Synthesis Using Bilevel Optimization},
author={Chen, Jiayi and Ke, Yubin and Wang, He},
journal={arXiv preprint arXiv:2412.16490},
year={2024}
}
If you have any questions, please feel free to contact Jiayi Chen at jiayichen@pku.edu.cn, Yubin Ke at 2200013213@stu.pku.edu.cn, and He Wang at hewang@pku.edu.cn.