An AI Tool to Predict Engagement and ‘Meltdown’ Events in Students with Autism Erasmus Project

Project Title

An AI Tool to Predict Engagement and ‘Meltdown’ Events in Students with Autism

Project Key Action

This project related with these key action: Cooperation for innovation and the exchange of good practices

Project Action Type

This project related with this action type : Strategic Partnerships for school education

Project Call Year

This project’s Call Year is 2020

Project Topics

This project is related with these Project Topics: Access for disadvantaged; Disabilities – special needs; ICT – new technologies – digital competences

Project Summary

CONTEXT – Prevalence of students with autistic spectrum conditions (ASC) varies in European countries, from 4.4 – 19.7 per 1,000, aged 7-9 years.
The UK Govt. autism strategy – ‘Fulfilling and rewarding lives’ (2014) has a particular focus on promoting innovation in service provision. Legislation provides for new duties and services including a disability support system that covers education (e.g., the ‘Children and Families Act 2014’).
Up to 71% of students with ASC attend mainstream schools, and it is important to provide them with an environment adjusted to their needs, and free of any emotional trauma. However, research shows that mainstream schools are frequently fully neither educated nor equipped to deal with the needs of an autistic child, and give the necessary support. Without individualised support, it is highly unlikely that a pupil with ASC will make the academic and social progress that they should.
60% of the teachers in England do not feel they have had the adequate training to teach children with autism, and that this situation is not improving (‘Ambitious About Autism’). 40% of children with ASC have been bullied (Autism Education Trust). A staggering 70% of parents are not satisfied with the level of understanding of autism in their child’s school.
‘Autism Europe’ states that Education must be adapted, to allow: assessment to determine needs and strengths, fully informing and involving families, training to help pre-empt or manage challenging behaviour, individual education and support plans, opportunities to learn skills and engage in socially inclusive situations, and education and support to help prevent secondary consequences of their difficulties such as anxiety disorders and depression.

OBJECTIVES
To address these challenges, our project will transfer and further refine an innovative AI-driven tool to help teachers, support staff, parents and carers to better understand the behavioural and cognitive processes related to the emotional state of students with ASC in classrooms.
Research has shown that engagement increases when activities are tailored to personal needs and learners’ emotional states, and that effectiveness of the learning process is directly correlated with learners’ engagement in learning activities.
Our innovative digital tool will infer the level of engagement of students with autism, and predict and pre-empt ‘meltdown’ moments. Teachers, support staff, parents and carers can then use this information (presented via a well-designed dashboard) to:
• Provide personalised learning pathways and support according to the specific needs of each child with autism.
• Improve the understanding of autism both within school, and in the wider community.
• Provide a learning environment free of emotional trauma.
• Increase the likelihood that students with autism will make the academic and social progress that they should.
• Decrease incidents of challenging behaviour, leading to improved mental wellbeing of individual students, and decreasing instances of absenteeism and bullying.
• Improve access to teaching of academic, independent living, adaptive behaviours and social skills.
• Help prevent secondary consequences of their difficulties such as anxiety disorders and depression.

METHODS and RESULTS: Using mobile devices, the level of engagement in learning activities of students with autism, and their behavioural characteristics can be assessed in a non-intrusive manner using sensor data such as eye gaze, body posture and movements. Labels to train machine learning algorithms to predict engagement and challenging behaviour will be applied to these data, gathered by working with students with autism throughout the schools in our partnership. Once trained, our innovative user-friendly AI-tool (IO2/App) will allow teachers, parents, and carers to personalise learning based on individual needs, and pre-empt challenging behaviour. This App will be supported by an Online Database of latest autism-related methods (IO1), a Handbook (IO4) and a Pedagogical Framework (IO5) acting as a ‘blueprint’ for school staff to optimally introduce AI technology into their classrooms.

CONTEXT – target groups & numbers
– Teachers, teaching assistants, and special education needs coordinators in mainstream education, inclusive settings, and special educational settings (30/country).
– Beneficiaries: students with ASC, and their families (100 in total).
– Stakeholders: policy makers, pedagogical experts, schools, school authorities and their umbrella structures all over Europe. Service providers supporting students with ASC; Local education authorities at all levels, Associations for teachers and teacher unions.

IMPACT – personalised learning solutions tailored according to the needs of students with autism, increased likelihood of academic and social progress, improved mental webbing – fewer incidents of challenging behaviour, and decreased absenteeism, smoother transition to the world of work.

EU Grant (Eur)

Funding of the project from EU: 441295 Eur

Project Coordinator

THE NOTTINGHAM TRENT UNIVERSITY & Country: UK

Project Partners

• PHOENIXKM BVBA
• Nottingham City Council
• SDRUZENIJE NA NA RABOTESHTITE S HORA S UVREZHDANIYA
• SU PAISIY HILENDARSKI
• DIEYFTHYNSI DEYTEROVATHMIAS EKPAIDEYSIS N. KARDITSAS
• SoftQNR D.O.O.
• PANEPISTIMIO THESSALIAS