Memory Efficient Grasping Point Detection of Nontrivial Objects

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216275%3A25530%2F21%3A39918537" target="_blank" >RIV/00216275:25530/21:39918537 - isvavai.cz</a>

Výsledek na webu

<a href="https://ieeexplore.ieee.org/document/9446869" target="_blank" >https://ieeexplore.ieee.org/document/9446869</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ACCESS.2021.3086417" target="_blank" >10.1109/ACCESS.2021.3086417</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Memory Efficient Grasping Point Detection of Nontrivial Objects

Popis výsledku v původním jazyce

Robotic manipulation with a nontrivial object providing various types of grasping points is of an industrial interest. Here, an efficient method of simultaneous detection of the grasping points is proposed. Specifically, two different 3 degree-of-freedom end effectors are considered for simultaneous grasping. The method utilizes an RGB data-driven perception system based on a specifically designed fully convolutional neural network called attention squeeze parallel U-Net (ASP U-Net). ASP U-Net detects grasping points based on a single RGB image. This image is transformed into a schematic grayscale frame, where the positions and poses of the grasping points are coded into gradient geometric shapes. In order to approve the ASP U-Net architecture, its performance was compared with nine competitive architectures using metrics based on generalized intersection over union and mean absolute error. The results indicate its outstanding accuracy and response time. ASP U-Net is also computationally efficient enough. With a more than acceptable memory size (77 MB), the architecture can be implemented using custom single-board computers. Here, its capabilities were tested and evaluated on the NVIDIA Jetson NANO platform.

Název v anglickém jazyce

Memory Efficient Grasping Point Detection of Nontrivial Objects

Popis výsledku anglicky

Robotic manipulation with a nontrivial object providing various types of grasping points is of an industrial interest. Here, an efficient method of simultaneous detection of the grasping points is proposed. Specifically, two different 3 degree-of-freedom end effectors are considered for simultaneous grasping. The method utilizes an RGB data-driven perception system based on a specifically designed fully convolutional neural network called attention squeeze parallel U-Net (ASP U-Net). ASP U-Net detects grasping points based on a single RGB image. This image is transformed into a schematic grayscale frame, where the positions and poses of the grasping points are coded into gradient geometric shapes. In order to approve the ASP U-Net architecture, its performance was compared with nine competitive architectures using metrics based on generalized intersection over union and mean absolute error. The results indicate its outstanding accuracy and response time. ASP U-Net is also computationally efficient enough. With a more than acceptable memory size (77 MB), the architecture can be implemented using custom single-board computers. Here, its capabilities were tested and evaluated on the NVIDIA Jetson NANO platform.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

20204 - Robotics and automatic control

Projekt

<a href="/cs/project/EF17_049%2F0008394" target="_blank" >EF17_049/0008394: Spolupráce Univerzity Pardubice a aplikační sféry v aplikačně orientovaném výzkumu lokačních, detekčních a simulačních systémů pro dopravní a přepravní procesy (PosiTrans)</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

IEEE ACCESS

ISSN

2169-3536

e-ISSN

—

Svazek periodika

2021

Číslo periodika v rámci svazku

9

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

16

Strana od-do

82130-82145

Kód UT WoS článku

000673965700001

EID výsledku v databázi Scopus

2-s2.0-85107372074