View
40
Download
1
Category
Preview:
Citation preview
A comparison of humanoid and non humanoid robots in supporting the
learning of pupils with severe intellectual disabilities
Sarmad Aslam School of Medicine
University of Nottingham
PJ Standen Rehabilitation and Ageing,
School of Medicine University of Nottingham
Nick Shopland Interactive Systems Research Group
Nottingham Trent University
Andy Burton Interactive Systems Research Group
Nottingham Trent University
David Brown Interactive Systems Research Group
Nottingham Trent University
• Technology has revolutionised learning in the classroom
• Challenging for Pupils with Intellectual disabilities(ID) to use technology
• Special schools can not afford expensive technology
• Pupils with ID disadvantaged from the lack of technology use
2
Introduction
• Robots have been used in special needs schools
• Humanoid Robots have been used to enhance communication in pupils with ID
• The cost of a humanoid robot is higher than most special schools can afford
• A cheaper robot which could be an alternative to the relatively expenses humanoid robot
3
Introduction
• To compare the humanoid and a non-humanoid robot in enhancing communication skills in children with Intellectual disabilities.
• Hypothesis: No significant difference between the two robots
4
Introduction
• Design: single case experimental ABAB design
• Robot demonstration at a special needs school to the teachers
• Three Teachers identified six students that could benefit from working with the robots
• Parent consent was gained
• Overall learning objectives drawn for every student
• 1 drop out due to lack of participation
• 1 drop out due to illness
5
Method
6
Method
Participant Gender and Age Overall learning objective
CC M 15 Improve listening skills
LW F 16 To understand and communicate directions KT M 14 To make puroposeful choices MH M 7 To improve English speaking skills
TH To interact with the robot and make choices - - -
• NAO
• Lego Mindstorm
7
Method
• Students were allowed time to play with the robots
• Eight tasks were designed to help achieve the students’ learning objectives
• Tasks were kept the same for both of the robots
• 10mins per task recorded on video
• Four tasks were carried out over a four week period in ‘ABAB’ style
• Fifth week was used to top up any missed sessions
8
Method
• Video data analysed for percentage engagement and percentage errors
• percentage engagement measured using eye contact
• Percentage of errors measured by counting erroneous attempts out of total attempts
• Data for both robots entered into SPSS
• Statistical tests of significance carried out
• T-test and Mann Whitney U test used
9
Method
• Engagement- No significant difference between the robots in two participants
• In two participants the non-humanoid mindstorm had significantly higher percentage of engagement
10
Results
0
20
40
60
80
100
120
LW CC KT MZ
Nao
Lego
• Percentage errors - no significant difference between the two robots in all the participants
11
Results
0
5
10
15
20
25
30
LW CC KT MZ
NAO
Leggo
• Non-humanoid robot can be just as good or in some cases even better at engaging pupils with intellectual disabilities than the Humanoid robot
• Non-humanoid robot was just as good at helping students achieve their learning goals as the Humanoid robot.
• First study to compare humanoid and non humanoid robots- opening doors to more research in the field
• Study carried out in a natural classroom setting, Tasks between the robots kept the same
• Tasks carried out at various times of the day
12
Discussion
• Small sample size
• Eye gaze is a questionable method for measurement of engagement but the only non-invasive method
• Results could be due to age differences: younger students prefer NAO
• Lego effect – putting pieces together
13
Discussion
• Non-humanoid Lego robot can be just as good or in some cases even better at engaging pupils with intellectual disabilities.
• Non-humanoid just as good at helping students achieve their learning objectives
• A larger more varied sample with a diverse learning objectives should be tested in future research
14
Conclusion
Thank you
Any questions?
Recommended