Learning Extrinsic Dexterity with Parameterized Manipulation Primitives

TL;DR

This paper introduces a hierarchical reinforcement learning method that learns parameterized manipulation primitives to enable robots to manipulate objects into graspable configurations in occluded scenarios, bypassing the need for explicit object detection or modeling.

Contribution

It presents a novel approach combining learned manipulation primitives with hierarchical RL to control object pose using depth data, facilitating complex interactions without detailed physical models.

Findings

Achieved 98% success rate in real robot experiments.

Successfully transferred from simulation to real-world tasks.

Operates directly on depth data without object detection or pose estimation.

Abstract

Many practically relevant robot grasping problems feature a target object for which all grasps are occluded, e.g., by the environment. Single-shot grasp planning invariably fails in such scenarios. Instead, it is necessary to first manipulate the object into a configuration that affords a grasp. We solve this problem by learning a sequence of actions that utilize the environment to change the object's pose. Concretely, we employ hierarchical reinforcement learning to combine a sequence of learned parameterized manipulation primitives. By learning the low-level manipulation policies, our approach can control the object's state through exploiting interactions between the object, the gripper, and the environment. Designing such a complex behavior analytically would be infeasible under uncontrolled conditions, as an analytic approach requires accurate physical modeling of the interaction and contact dynamics. In contrast, we learn a hierarchical policy model that operates directly on depth perception data, without the need for object detection, pose estimation, or manual design of controllers. We evaluate our approach on picking box-shaped objects of various weight, shape, and friction properties from a constrained table-top workspace. Our method transfers to a real robot and is able to successfully complete the object picking task in 98\% of experimental trials. Supplementary information and videos can be found at https://shihminyang.github.io/ED-PMP/.