# Leveraging Extrinsic Dexterity for Occluded Grasping on Grasp Constraining Walls

arXiv: https://arxiv.org/abs/2507.14721 · HTML: https://mlsanity.com/paper/2507.14721/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods.
- Builds on 24 of its 27 references that are in the map.
- First posted on arXiv 19 Jul 2025.

## Placed in

- Deep reinforcement learning methods / [RL for robot locomotion and manipulation](https://mlsanity.com/deeprl/rl-for-robot-locomotion-and-manipulation/)

## Builds on (most cited first, 24 of 24)

- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [Soft Actor-Critic Algorithms and Applications](https://mlsanity.com/paper/1812.05905/) (2018)
- [6-DOF GraspNet: Variational Grasp Generation for Object Manipulation](https://mlsanity.com/paper/1905.10520/) (2019)
- [Deep Reinforcement Learning for High Precision Assembly Tasks](https://mlsanity.com/paper/1708.04033/) (2017)
- [HACMan: Learning Hybrid Actor-Critic Maps for 6D Non-Prehensile Manipulation](https://mlsanity.com/paper/2305.03942/) (2023)
- [Learning to Grasp the Ungraspable with Emergent Extrinsic Dexterity](https://mlsanity.com/paper/2211.01500/) (2022)
- [Learning Diverse and Physically Feasible Dexterous Grasps with Generative Model and Bilevel Optimization](https://mlsanity.com/paper/2207.00195/) (2022)
- [Efficient Sim-to-real Transfer of Contact-Rich Manipulation Skills with Online Admittance Residual Learning](https://mlsanity.com/paper/2310.10509/) (2023)
- [Learning Insertion Primitives with Discrete-Continuous Hybrid Action Space for Robotic Assembly Tasks](https://mlsanity.com/paper/2110.12618/) (2021)
- [Manipulation Trajectory Optimization with Online Grasp Synthesis and Selection](https://mlsanity.com/paper/1911.10280/) (2019)
- [Contact Mode Guided Motion Planning for Quasidynamic Dexterous Manipulation in 3D](https://mlsanity.com/paper/2105.14431/) (2021)
- [ContactDexNet: Multi-fingered Robotic Hand Grasping in Cluttered Environments through Hand-object Contact Semantic Mapping](https://mlsanity.com/paper/2404.08844/) (2024)
- [Learning Pregrasp Manipulation of Objects from Ungraspable Poses](https://mlsanity.com/paper/2002.06344/) (2020)
- [One-Shot Transfer of Long-Horizon Extrinsic Manipulation Through Contact Retargeting](https://mlsanity.com/paper/2404.07468/) (2024)
- [One-Shot Manipulation Strategy Learning by Making Contact Analogies](https://mlsanity.com/paper/2411.09627/) (2024)
- [Learning Generalizable Pivoting Skills](https://mlsanity.com/paper/2305.02554/) (2023)
- [Multi-Stage Reinforcement Learning for Non-Prehensile Manipulation](https://mlsanity.com/paper/2307.12074/) (2023)
- [HACMan++: Spatially-Grounded Motion Primitives for Manipulation](https://mlsanity.com/paper/2407.08585/) (2024)
- [Learning Extrinsic Dexterity with Parameterized Manipulation Primitives](https://mlsanity.com/paper/2310.17785/) (2023)
- [Decomposed Vector-Quantized Variational Autoencoder for Human Grasp Generation](https://mlsanity.com/paper/2407.14062/) (2024)
- [Object manipulation through contact configuration regulation: multiple and intermittent contacts](https://mlsanity.com/paper/2310.00798/) (2023)
- [Robust In-Hand Manipulation with Extrinsic Contacts](https://mlsanity.com/paper/2403.18960/) (2024)
- [Robotic Object Insertion with a Soft Wrist through Sim-to-Real Privileged Training](https://mlsanity.com/paper/2408.17061/) (2024)
- [COMBO-Grasp: Learning Constraint-Based Manipulation for Bimanual Occluded Grasping](https://mlsanity.com/paper/2502.08054/) (2025)

## Built on by (most cited first, 1 of 1)

- [Emerging Extrinsic Dexterity in Cluttered Scenes via Dynamics-aware Policy Learning](https://mlsanity.com/paper/2603.09882/) (2026)

## Abstract (arXiv)

This study addresses the problem of occluded grasping, where primary grasp configurations of an object are not available due to occlusion with environment. Simple parallel grippers often struggle with such tasks due to limited dexterity and actuation constraints. Prior works have explored object pose reorientation such as pivoting by utilizing extrinsic contacts between an object and an environment feature like a wall, to make the object graspable. However, such works often assume the presence of a short wall, and this assumption may not always hold in real-world scenarios. If the wall available for interaction is too large or too tall, the robot may still fail to grasp the object even after pivoting, and the robot must combine different types of actions to grasp. To address this, we propose a hierarchical reinforcement learning (RL) framework. We use Q-learning to train a high-level policy that selects the type of action expected to yield the highest reward. The selected low-level skill then samples a specific robot action in continuous space. To guide the robot to an appropriate location for executing the selected action, we adopt a Conditional Variational Autoencoder (CVAE). We condition the CVAE on the object point cloud and the skill ID, enabling it to infer a suitable location based on the object geometry and the selected skill. To promote generalization, we apply domain randomization during the training of low-level skills. The RL policy is trained entirely in simulation with a box-like object and deployed to six objects in real world. We conduct experiments to evaluate our method and demonstrate both its generalizability and robust sim-to-real transfer performance with promising success rates.