DGBench: An Open-Source, Reproducible Benchmark for Dynamic Grasping

TL;DR

This paper presents DGBench, an open-source benchmark for dynamic grasping that enables reproducible testing and comparison of perception systems in environments with unpredictable motion, demonstrating improved success rates with a multi-camera setup.

Contribution

The paper introduces DGBench, a reproducible benchmark for dynamic grasping, and evaluates a multi-camera perception system showing advantages over traditional configurations.

Findings

Multi-camera perception system outperforms single-camera setups.

Significant improvement in grasp success rate on a real robot.

Benchmark enables standardized evaluation of dynamic grasping methods.

Abstract

This paper introduces DGBench, a fully reproducible open-source testing system to enable benchmarking of dynamic grasping in environments with unpredictable relative motion between robot and object. We use the proposed benchmark to compare several visual perception arrangements. Traditional perception systems developed for static grasping are unable to provide feedback during the final phase of a grasp due to sensor minimum range, occlusion, and a limited field of view. A multi-camera eye-in-hand perception system is presented that has advantages over commonly used camera configurations. We quantitatively evaluate the performance on a real robot with an image-based visual servoing grasp controller and show a significantly improved success rate on a dynamic grasping task.