Publication:
around a humanoid robot.
Pre-schoolers’ interest and caring behavior around NAO
by Emily Andreou
by Emily Andreou
BA student, Department of Multimedia and Graphic Arts,
Cyprus Interaction Lab,
Cyprus University of Technology,
Limassol, Cyprus
(emandre26@gmail.com)
Supervising Professor Dr. Andri Ioannou
Ioannou, A., Andreou, E., Christofi, M. (2014). Preschoolers’ interest and caring behaviour around a humanoid robot. TechTrends, in press.
First of all, I would like to thank my supervisor Professor Dr. Andri Ioannou for her help and guidance during the whole course of my dissertation, as well as Maria Christofi for her valuable assistance. I would also like to thank all the parents who entrusted their children to voluntarily participate in my research. Furthermore I would like to thank everyone who participated in any way in the completion of this study.
Abstract:
The study involved NAO and four children in pre-primary school aged 3-5 years. NAO was placed in a playground where the children played, together with a variety of toys. NAO appeared as a peer who played,
talked, danced and said stories. All interactions in the playground were video recorded. The analysis focused on children's behavior around NAO . The results showed that children can easily interact with this robot. They showed particular interest to NAO when he danced and when he was in need of help (e.g., fell down) demonstrating caring behavior such as kisses, hugs and cuddling.
talked, danced and said stories. All interactions in the playground were video recorded. The analysis focused on children's behavior around NAO . The results showed that children can easily interact with this robot. They showed particular interest to NAO when he danced and when he was in need of help (e.g., fell down) demonstrating caring behavior such as kisses, hugs and cuddling.
1. Introduction:
As time goes by, our life gets more and more depended on the use of technology. Pre elementary and elementary teachers incorporate the use of technology in their classes. For example [1],[2] stated that socialization between toddlers and robots is possible for extended periods of time and that humanoid robots can preserve long term relationships with humans. Also, [3] support that with the use of CRR, the robots can promote learning by encouraging the kids to teach the robots.
As time goes by, our life gets more and more depended on the use of technology. Pre elementary and elementary teachers incorporate the use of technology in their classes. For example [1],[2] stated that socialization between toddlers and robots is possible for extended periods of time and that humanoid robots can preserve long term relationships with humans. Also, [3] support that with the use of CRR, the robots can promote learning by encouraging the kids to teach the robots.
In this paper I decided to study the humanoid robot NAO and its possible use in the educational system. This study focuses on the behavior of children around the robot when it is placed in a room with toys and is presented as a friend of theirs, who talks, dances, tells stories etc.
Along the way two questions were raised:
-What kinds of behaviour arise around the humanoid robot NAO?
-How can NAO keep the interest of the children?
2. Method:
As a first step for this study I had to get familiarized with the NAO robot, its software and its capabilities. After my familiarization with NAO I studied the relevant bibliography and decided the role NAO would play in my study. I decided to give a childlike quality and gestures to NAO, so children of 3-5 years would interact with it comfortably. I programmed the robot using the Choregraphe program. I made use of some already programmed commands provided by the software; like sitting, getting up, walking etc. I also recorded some sentences NAO would need to say to the children because the Greek language wasn’t recognized by the software. Furthermore I programmed some dancing movements from choreographies of the songs Gangam-style and Macarena with the use of key-frame animation.
As a first step for this study I had to get familiarized with the NAO robot, its software and its capabilities. After my familiarization with NAO I studied the relevant bibliography and decided the role NAO would play in my study. I decided to give a childlike quality and gestures to NAO, so children of 3-5 years would interact with it comfortably. I programmed the robot using the Choregraphe program. I made use of some already programmed commands provided by the software; like sitting, getting up, walking etc. I also recorded some sentences NAO would need to say to the children because the Greek language wasn’t recognized by the software. Furthermore I programmed some dancing movements from choreographies of the songs Gangam-style and Macarena with the use of key-frame animation.
Image1: Programming gestures for NAO on Choregraphe program.
Afterwards I created a playground in a room of the Cyprus Interaction Lab of the Cyprus University of Technology. Besides the robot, I placed different toys like table games, dolls, cars, bears etc.
My scenario for the robot-children interaction was then conceived. It would last 35-40 minutes. The kids would enter the playroom and would have a few minutes to explore. Three minutes later NAO would get up, introduce himself and say good morning. In response, the kids would say good morning and introduce themselves either by their own will either by the encouragement of the researcher. Following that, NAO would sit down and the kids would be left free in the room to interact as they please. If the kids ignore the robot, it will get up and do some dance movements and I will note whether the children respond to that or not. According to the children’s behaviour I will keep NAO active and see if it engages their interest. At one point NAO will start telling a fairytale and again I will note the response to it. In the end NAO will get up and finish the story. If the kids aren’t tired NAO will do more gestures, dance moves etc.
Before the official experiment with the kids I did a pilot experiment with a 7 year old child. Everything ran smoothly and according to the scenario. I videotaped the whole process and took notes on his interaction with the robot. In the end I asked the participant what he wanted to see from the robot that he didn’t get to see.
Image2: pilot experiment with 1 child
In the main experiment, four children participated of 3-5 years old. The meeting took place after the parents’ agreement and they all met me at the Interaction lab. All the kids entered the playroom at the same time. In it, were the researcher that would control the activities and 2 more researches that would help with NAO and observe. The kids were then left free to interact with the robot and play with the games.
Image3: Main researcher and kids interacting with the robot
Data were collected from two sources.
All the interactions were taped by the 4 cameras we used. We covered all the sides so no action would be left unnoticed. Because of the nature of the experiment and the children it was decided that videotaping was necessary to capture moments that the researcher might have missed during the execution of the experiment.
In the end of the experiment during an informal focus group interview, the children were asked some questions like:
- What did you like the most from the playground?
- What did you like most from the robot?
- Would you like your own robot?
- What would you do with it?
During the interview one of the kids prepared a drawing of the robot which she happily gave to the researcher.
3. Analysis and Results:
Two researchers examined the experiment. After the transcription of the conversations recorded on the videos, the researchers watched the videos and created a coding system that categorizes the kinds of interactions (verbal and non-verbal) of the children around the robot. Table 1 shows the coding system in which three categories of interactions are included. Coding unit was the turn taking of the participants. Based on table 1 the frequency of each category was measured as presented on table 2.
Two researchers examined the experiment. After the transcription of the conversations recorded on the videos, the researchers watched the videos and created a coding system that categorizes the kinds of interactions (verbal and non-verbal) of the children around the robot. Table 1 shows the coding system in which three categories of interactions are included. Coding unit was the turn taking of the participants. Based on table 1 the frequency of each category was measured as presented on table 2.
Table 1: Coding system of interactions with NAO
The analysis of the video showed that the participants were most exited during the dancing stage. They were most interested in NAO when he was dancing and they were trying to touch and explore him. Most kids didn’t want NAO to stop dancing so they were asking the researcher if NAO would dance more and to not turn him off. They were also talking to NAO as if he were a kid, “Why don’t you dance?”, “Sit” (table 1). It was also observed that the children were comfortable with NAO and communicated with him as if he were friend.
During dancing NAO faced a technical problem and fell down which drew the kids interest. They tried to care for him and touched him sympathetically. The kids were worried because they thought he was crying (music from Tai-Chi dance). One of the kids kissed NAO on the head twice and the other kids were asking why he is crying and if he is hurt. The kids didn’t see NAO as a toy but as a kid just like them. Only one girl tried to move the robots arms as if it were a toy.
Table 2: Analytic count of the categories per child
Following are the conclusions from the informal interviews the researcher took from the kids.
Child 1: “K” mentioned he liked the robot and that he considered him his friend. What he liked most was when NAO was ”crying” and if he had a robot himself he would push it down often so he would hear it cry. The kid also mentioned that he wanted to baptize NAO and change his name probably because it was hard for him to remember his name as he was asking the robots name often.
Child 2: “N” said he liked the robot, he considered him his friend and that he would like to go to school with him.
Child 3: “P” said he liked NAO’s eyes because they changed colour.
Child 4: “X” also liked NAO’s changing eye colour and the fact that he was saying ouch when he fell down. If she had a robot would push it down so she could hear it say ouch. During the interview she drew a picture of NAO and showed it to the researcher. She drew the head, two ears, two legs and arms. She also added special buttons that when you press them the robot falls and says ouch and starts crying.
Image 4: “X’s” drawing
Generally, results showed that kids of 3-5 years old can easily interact with humanoid robots (the kids explored NAO and touched him like he was a kid as well). Moreover that the robot can hold their interest when it dances and gives them the chance to take care of it.
In the process of this research two questions were raised that can be answered as following.
What kinds of behaviour arise around the humanoid robot NAO?
The video analysis showed that the participants interacted with NAO. The interactions were mostly touching and questions towards him or towards the researcher concerning NAO’s behaviour (table1). Results from table 2 showed that girls reacted more emotionally and caressed NAO. In contrary, boy’s touch was more explorative and to see how he would react. Finally research showed that kids addressed NAO as one of their own.
How can NAO keep the interest of the children?
Results showed that the kids were most interested when NAO was dancing or fell down. During dancing they stopped all other activities and were just watching NAO. Some kids were saying hello to NAO and one kid was clapping and dancing along. Their interest was also peaked when NAO fell down and they showed signs of compassion.
Results showed that the kids were most interested when NAO was dancing or fell down. During dancing they stopped all other activities and were just watching NAO. Some kids were saying hello to NAO and one kid was clapping and dancing along. Their interest was also peaked when NAO fell down and they showed signs of compassion.
In conclusion, results of this research are consistent with findings of previous researches. For example Tanaka, Cicourel, Movellan (2007) found that toddlers presented a variety of social behaviours, like taking care of the robot and gradually interacting with it as one of their own rather than a toy. Similarly this research showed that children were interested when NAO fell down and were compassionate and touched the robot to comfort it.
In addition, the conclusions of Kanda, Sato, Saiwaki και Ishiguro (2007), also showed that the kids developed friendly attitude towards the robot. On the contrary their research showed that the robot kept the kids attention the whole time something that didn’t happen in this research. Here the kid’s attention was engaged to the robot only when it gave them the chance to take care of it whilst it lost it when it was storytelling and had delayed response.
Applications in education
According to my findings I conclude that humanoid robots can aid in the education of young children. The kids interacted with NAO and addressed him like one of their own. This shows that the humanoid robot NAO can be smoothly introduced to educational activities.
The kids showed interest when NAO fell which shows that humanoid robots can evoke feelings and behaviors to the kids and so they can be used to raise awareness and educate them on how to behave in certain situations.
References
1] F. Tanaka , A. Cicourel and J. R. Movellan, "Socialization between toddlers and robots at an early childhood education center", Proc. Nat. Acad. Sci., vol. 104, pp.17954 -17958, 2007
2] T. Kanda , R. Sato , N. Saiwaki and H. Ishiguro "A two-month field trial in an elementary school for long-term human–robot interaction", IEEE Trans. Robot., vol. 23, no. 5, pp.962 -971, 2007
3] F. Tanaka, S. Matsuzoe, “Learning verbs by teaching a care-receiving robot by children: An experimental report.” In Human-Robot Interaction (HRI), 2012 7th ACM/IEEE International Conference on (pp. 253-254).
Thanks for appreciate!!
