Active vision for dexterous grasping of novel objects

TL;DR

This paper presents an active vision approach for dexterous grasping of unfamiliar objects, improving success rates by guiding sensing based on anticipated actions to enhance reconstruction and safety.

Contribution

It introduces a three-stage active vision method that refines grasp planning through targeted sensing, outperforming baseline algorithms in dexterous object grasping.

Findings

Achieved 80.4% success rate in grasping tasks.

Outperformed a randomized algorithm with 64.3% success.

Enhanced safety and reliability of grasping through active sensing.

Abstract

How should a robot direct active vision so as to ensure reliable grasping? We answer this question for the case of dexterous grasping of unfamiliar objects. By dexterous grasping we simply mean grasping by any hand with more than two fingers, such that the robot has some choice about where to place each finger. Such grasps typically fail in one of two ways, either unmodeled objects in the scene cause collisions or object reconstruction is insufficient to ensure that the grasp points provide a stable force closure. These problems can be solved more easily if active sensing is guided by the anticipated actions. Our approach has three stages. First, we take a single view and generate candidate grasps from the resulting partial object reconstruction. Second, we drive the active vision approach to maximise surface reconstruction quality around the planned contact points. During this phase, the anticipated grasp is continually refined. Third, we direct gaze to improve the safety of the planned reach to grasp trajectory. We show, on a dexterous manipulator with a camera on the wrist, that our approach (80.4% success rate) outperforms a randomised algorithm (64.3% success rate).