Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data
In modern robot applications, there is often a need to manipulate previously unknown objects in an unstructured environment. The field of grasp-planning deals with the task of finding grasps for a given object that can be successfully executed with a robot. The predicted grasps can be evaluated acco...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-12-01
|
Series: | Applied Sciences |
Subjects: | |
Online Access: | https://www.mdpi.com/2076-3417/13/1/525 |
_version_ | 1797626284091637760 |
---|---|
author | Artúr István Károly Péter Galambos |
author_facet | Artúr István Károly Péter Galambos |
author_sort | Artúr István Károly |
collection | DOAJ |
description | In modern robot applications, there is often a need to manipulate previously unknown objects in an unstructured environment. The field of grasp-planning deals with the task of finding grasps for a given object that can be successfully executed with a robot. The predicted grasps can be evaluated according to certain criteria, such as analytical metrics, similarity to human-provided grasps, or the success rate of physical trials. The quality of a grasp also depends on the task which will be carried out after the grasping is completed. Current task-specific grasp planning approaches mostly use probabilistic methods, which utilize categorical task encoding. We argue that categorical task encoding may not be suitable for complex assembly tasks. This paper proposes a transfer-learning-based approach for task-specific grasp planning for robotic assembly. The proposed method is based on an automated pipeline that quickly and automatically generates a small-scale task-specific synthetic grasp dataset using Graspit! and Blender. This dataset is utilized to fine-tune pre-trained grasp quality convolutional neural networks (GQCNNs). The aim is to train GQCNNs that can predict grasps which do not result in a collision when placing the objects. Consequently, this paper focuses on the geometric feasibility of the predicted grasps and does not consider the dynamic effects. The fine-tuned GQCNNs are evaluated using the Moveit! Task Constructor motion planning framework, which enables the automated inspection of whether the motion planning for a task is feasible given a predicted grasp and, if not, which part of the task is responsible for the failure. Our results suggest that fine-tuning GQCNN models can result in superior grasp-planning performance (0.9 success rate compared to 0.65) in the context of an assembly task. Our method can be used to rapidly attain new task-specific grasp policies for flexible robotic assembly applications. |
first_indexed | 2024-03-11T10:08:09Z |
format | Article |
id | doaj.art-bd6ef850378d4a3c96bc8f72519d17dc |
institution | Directory Open Access Journal |
issn | 2076-3417 |
language | English |
last_indexed | 2024-03-11T10:08:09Z |
publishDate | 2022-12-01 |
publisher | MDPI AG |
record_format | Article |
series | Applied Sciences |
spelling | doaj.art-bd6ef850378d4a3c96bc8f72519d17dc2023-11-16T14:58:14ZengMDPI AGApplied Sciences2076-34172022-12-0113152510.3390/app13010525Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic DataArtúr István Károly0Péter Galambos1Antal Bejczy Center for Intelligent Robotics, Óbuda University, Bécsi út 96/B, 1034 Budapest, HungaryAntal Bejczy Center for Intelligent Robotics, Óbuda University, Bécsi út 96/B, 1034 Budapest, HungaryIn modern robot applications, there is often a need to manipulate previously unknown objects in an unstructured environment. The field of grasp-planning deals with the task of finding grasps for a given object that can be successfully executed with a robot. The predicted grasps can be evaluated according to certain criteria, such as analytical metrics, similarity to human-provided grasps, or the success rate of physical trials. The quality of a grasp also depends on the task which will be carried out after the grasping is completed. Current task-specific grasp planning approaches mostly use probabilistic methods, which utilize categorical task encoding. We argue that categorical task encoding may not be suitable for complex assembly tasks. This paper proposes a transfer-learning-based approach for task-specific grasp planning for robotic assembly. The proposed method is based on an automated pipeline that quickly and automatically generates a small-scale task-specific synthetic grasp dataset using Graspit! and Blender. This dataset is utilized to fine-tune pre-trained grasp quality convolutional neural networks (GQCNNs). The aim is to train GQCNNs that can predict grasps which do not result in a collision when placing the objects. Consequently, this paper focuses on the geometric feasibility of the predicted grasps and does not consider the dynamic effects. The fine-tuned GQCNNs are evaluated using the Moveit! Task Constructor motion planning framework, which enables the automated inspection of whether the motion planning for a task is feasible given a predicted grasp and, if not, which part of the task is responsible for the failure. Our results suggest that fine-tuning GQCNN models can result in superior grasp-planning performance (0.9 success rate compared to 0.65) in the context of an assembly task. Our method can be used to rapidly attain new task-specific grasp policies for flexible robotic assembly applications.https://www.mdpi.com/2076-3417/13/1/525grasp planningrobotic graspingrobot manipulationdeep learningGQCNNsynthetic data |
spellingShingle | Artúr István Károly Péter Galambos Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data Applied Sciences grasp planning robotic grasping robot manipulation deep learning GQCNN synthetic data |
title | Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data |
title_full | Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data |
title_fullStr | Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data |
title_full_unstemmed | Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data |
title_short | Task-Specific Grasp Planning for Robotic Assembly by Fine-Tuning GQCNNs on Automatically Generated Synthetic Data |
title_sort | task specific grasp planning for robotic assembly by fine tuning gqcnns on automatically generated synthetic data |
topic | grasp planning robotic grasping robot manipulation deep learning GQCNN synthetic data |
url | https://www.mdpi.com/2076-3417/13/1/525 |
work_keys_str_mv | AT arturistvankaroly taskspecificgraspplanningforroboticassemblybyfinetuninggqcnnsonautomaticallygeneratedsyntheticdata AT petergalambos taskspecificgraspplanningforroboticassemblybyfinetuninggqcnnsonautomaticallygeneratedsyntheticdata |