Multi-resolution Modeling of Human with OctoMap for Human-Robot Interaction Studies

Özlem Örnek; Ahmet Yazici; Metin Özkan

Article Details

Multi-resolution Modeling of Human with OctoMap for Human-Robot Interaction Studies

Authors:

Özlem Örnek | Ahmet Yazici | Metin Özkan

Publisher:

Dokuz Eylul University

Main Topics:

Mühendislik

Journal:

Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 74, pp.303-316

Document Type:

Article

Publication Date:

5/2023

Document Language:

Turkish

Document ID (Org No.):

2001289552

pISSN:

1302-9304

Keywords:

Abstract:

 While industrial robots have been working in safety cages for many years, today, with Industry 4.0, the cages have been discarded and started to work in interaction with humans. Thus, studies in the field of human-robot interaction HRI, which aims to ensure that humans and robots work together safely and effectively, are increasingly continuing. There are many different types of HRIs for humans and robots in industrial environments. One of them is human-robots working together. In working together, human and robot work for the same purpose. It is important to ensure the physical and mental safety of people during work. The aim of this study is to create human models that will be used by the robot in the action plans to protect the mental health of the human during working together. While creating these models, it is recommended that some body parts of the human be modeled more precisely high resolution than other parts. It is aimed to improve mental health and productivity by providing confidence and comfort to people with high-resolution modeling of humans. To model the human, Octomap library, which is commonly used in robot applications to create a three-dimensional 3D environment map, with its octree structure was used. In determining the regions to be modeled with high resolution, human joint points, which can be provided by skeleton tracking algorithms, are used. Circular areas were created by selecting the points belonging to the human region of interested among the joint points. And multi-resolution modeling of the human was carried out by modeling the regions of the human body within these areas with high resolution and the areas outside of these areas with lower resolution. Experimental results with a data set are presented.

 While industrial robots have been working in safety cages for many years, today, with Industry 4.0, the cages have been discarded and started to work in interaction with humans. Thus, studies in the field of human-robot interaction HRI, which aims to ensure that humans and robots work together safely and effectively, are increasingly continuing. There are many different types of HRIs for humans and robots in industrial environments. One of them is human-robots working together. In working together, human and robot work for the same purpose. It is important to ensure the physical and mental safety of people during work. The aim of this study is to create human models that will be used by the robot in the action plans to protect the mental health of the human during working together. While creating these models, it is recommended that some body parts of the human be modeled more precisely high resolution than other parts. It is aimed to improve mental health and productivity by providing confidence and comfort to people with high-resolution modeling of humans. To model the human, Octomap library, which is commonly used in robot applications to create a three-dimensional 3D environment map, with its octree structure was used. In determining the regions to be modeled with high resolution, human joint points, which can be provided by skeleton tracking algorithms, are used. Circular areas were created by selecting the points belonging to the human region of interested among the joint points. And multi-resolution modeling of the human was carried out by modeling the regions of the human body within these areas with high resolution and the areas outside of these areas with lower resolution. Experimental results with a data set are presented.

 While industrial robots have been working in safety cages for many years, today, with Industry 4.0, the cages have been discarded and started to work in interaction with humans. Thus, studies in the field of human-robot interaction HRI, which aims to ensure that humans and robots work together safely and effectively, are increasingly continuing. There are many different types of HRIs for humans and robots in industrial environments. One of them is human-robots working together. In working together, human and robot work for the same purpose. It is important to ensure the physical and mental safety of people during work. The aim of this study is to create human models that will be used by the robot in the action plans to protect the mental health of the human during working together. While creating these models, it is recommended that some body parts of the human be modeled more precisely high resolution than other parts. It is aimed to improve mental health and productivity by providing confidence and comfort to people with high-resolution modeling of humans. To model the human, Octomap library, which is commonly used in robot applications to create a three-dimensional 3D environment map, with its octree structure was used. In determining the regions to be modeled with high resolution, human joint points, which can be provided by skeleton tracking algorithms, are used. Circular areas were created by selecting the points belonging to the human region of interested among the joint points. And multi-resolution modeling of the human was carried out by modeling the regions of the human body within these areas with high resolution and the areas outside of these areas with lower resolution. Experimental results with a data set are presented.

 Endüstriyel robotlar, uzun yıllar güvenlik kafesleri içinde çalışırken, günümüzde Endüstri 4.0 ile birlikte kafesler atılmış ve insanlar ile etkileşim içerisinde çalışmaya başlamıştır. Böylece, insan ve robotların birlikte güvenli ve etkin olarak çalışmasını sağlamayı amaçlayan insan-robot etkileşimi Human-Robot Interaciton, HRI alanında çalışmalar artarak devam etmektedir. Endüstriyel ortamlarda insanlar ve robotlar için birçok farklı HRI türü bulunmaktadır. Bunlardan biri de insan ve robotların birlikte çalışmasıdır. Birlikte çalışmada insan ve robot aynı amaç için çalışmaktadır. Çalışma sırasında insanın fiziksel ve mental güvenliğinin sağlanması önemlidir. Bu çalışmada, amaç insan ve robotların birlikte çalışmaları süresince insanın mental sağlığını korumak üzere robotun yapacağı hareket planlarında kullanacağı insan modellerinin oluşturulmasıdır. Bu modeller oluşturulurken, insanın bazı vücut bölgelerinin diğer bölgelerden daha hassas olarak modellenmesi yüksek çözünürlüklü önerilmektedir. İnsanın yüksek çözünürlüklü modellenmesi ile insana güven ve konfor sağlanarak, mental sağlığın ve verimliliğin geliştirilmesi hedeflenmiştir. İnsanı modellemek için, sekizli ağaç octree yapısıyla robot uygulamalarında yaygın olarak üç boyutlu 3B ortam haritasının oluşturulmasında kullanılan, Octomap kütüphanesi kullanılmıştır. Yüksek çözünürlüklü modellenecek bölgelerin belirlenmesinde, iskelet takip algoritmaları ile sağlanabilen insan eklem noktaları kullanılmaktadır. Eklem noktaları arasıdan, ilgilenilen insan bölgesine ait olanlar noktalar seçilmek suretiyle dairesel alanlar oluşturulmuş ve bu alanlar içerisinde kalan insan vücuduna ait bölgeler yüksek, dışında kalan alanlar ise daha düşük çözünürlükle modellenerek insanın çoklu çözünürlüklü modellenmesi gerçekleştirilmiştir. Bir veri seti ile yapılan deneysel çalışma sonuçları sunulmaktadır.

Article Details

Multi-resolution Modeling of Human with OctoMap for Human-Robot Interaction Studies

Authors:

Özlem Örnek | Ahmet Yazici | Metin Özkan

Publisher:

Dokuz Eylul University

Main Topics:

Mühendislik

Journal:

Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 74, pp.303-316

Document Type:

Article

Publication Date:

5/2023

Document Language:

Turkish

Document ID (Org No.):

2001289552

pISSN:

1302-9304

Human-robot interaction

Octomap

Human modelingİnsan-robot etkileşimi

Octomap

İnsan modelleme

Issues

Article Details

Multi-resolution Modeling of Human with OctoMap for Human-Robot Interaction Studies

Authors: Özlem Örnek | Ahmet Yazici | Metin Özkan

Publisher: Dokuz Eylul University

Main Topics: Mühendislik

Journal: Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 74, pp.303-316

Document Type: Article

Publication Date: 5/2023

Document Language: Turkish

Document ID (Org No.): 2001289552

pISSN: 1302-9304

Human-robot interaction

Octomap

Human modelingİnsan-robot etkileşimi

Octomap

İnsan modelleme

Issues

Authors:

Özlem Örnek | Ahmet Yazici | Metin Özkan

Publisher:

Dokuz Eylul University

Main Topics:

Mühendislik

Journal:

Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 74, pp.303-316

Document Type:

Article

Publication Date:

5/2023

Document Language:

Turkish

Document ID (Org No.):

2001289552

pISSN:

1302-9304