MoDex: A Diffusion Policy for Sequential Multi-Object Dexterous Grasping

Institution Name
Conference name and year

Abstract

This work addresses the problem of sequentially grasping multiple objects with a single dexterous gripper without releasing previously grasped objects, a task that has been limited to approaches requiring known object models or poses. The proposed solution, MoDex, overcomes these constraints by training an opposition-space and point-cloud-conditioned diffusion policy to output the next gripper pose. The opposition space condition specifies which robotic fingers participate in each grasp, allowing the gripper to use only a subset of available degrees of freedom while preserving the remaining degrees of freedom for subsequent grasps. To ensure MoDex is sim-to-real transfer-ready, we propose atwo-step training process: first, the policy is pre-trained via imitation learning onexpert demonstrations, followed by reinforcement learning fine-tuning to furtherenhance robustness. MoDex is evaluated against state-of-the-art baselines on a MuJoCo-simulated Franka Panda robot equipped with an Allegro Hand, and in the real world using the same hardware. The experimental results demonstratethat MoDex consistently outperforms existing learning-based methods, achieving4.75-36.25% and 6.67-17.78% higher success rates in simulation and real experiments, respectively.

Experimental Results

MoDex achieves the best simulation performance across all grasp stages and transfers to real hardware after training only in simulation.

Simulation Success Rates (%)

Method Stage 1 Stage 2 Stage 3 Average
BC-RNN52.5013.7526.2530.83
PPO0.000.000.000.00
SeqDiffuser1.670.000.000.56
MoDex-BC60.0046.2541.2549.17
MoDex (Ours)70.0050.0055.0058.33

Ablation Study Success Rates (%)

Variant Stage 1 Stage 2 Stage 3
DP3 (full)60.0046.2541.25
w/o Grasp History Context70.003.7538.75
MoDex (full)70.0050.0055.00
DPPO w/o grasp reward67.5045.0045.00
DPPO w/o maintain reward66.2546.2541.50

Real-World Success Rates (%)

Method Stage 1 Stage 2 Stage 3
MoDex-BC40.0020.006.67
MoDex57.7826.6720.00

Additional Figures

Real robot experiment setup
Real-world Franka Panda and Allegro Hand experiment setup.
Representative real-world success and failure rollouts
Representative real-world rollouts, including success cases and failure modes.

BibTeX

@article{modex,
  title={MoDex: A Diffusion Policy for Sequential Multi-Object Dexterous Grasping},
  author={Authors},
  journal={Conference/Journal Name},
  year={2026},
}