Blandine PICHON

PhD Student, Polytech Annecy-Chambéry.



Address: LISTIC - Polytech Annecy-Chambéry, BP 80439 - Annecy le Vieux, 74944 ANNECY Cedex, France


Topic: Artificial Intelligence methods to measure sensory disorders of autism by IoT devices

Abstract: Several kinds of research use connected objects to improve the quality of human life, to facilitate it or to interact with the physical world. We propose to take advantage of the Internet of Things (IoT), which interconnects physical objects, such as musical instruments, to measure sensory disorders, particularly in the study of autism (Autism Spectrum Disorder - ASD). As underlined by the Haute Autorité de Santé, sensory particularities have an impact and their knowledge allows an adapted care of people with ASD.
An initial study carried out within our university, in association with the Annecy Conservatory, has already validated that the sensory disorders of autism can be measured using sensors integrated into objects.
The issue is to establish sensory profiles, characterized by hypo and hyper-sensitivities, and to measure their relevance in relation to those traditionally obtained. We also wish to be able to study their evolution over time with weekly sampling, which cannot be done by traditional approaches.
The scientific challenge lies in the cooperation of 2 measurement methods whose only common points are at the level of the fundamental concepts of measurement science. Indeed, the key point of this study is similar to a calibration knowing that: the measured quantity is poorly defined; there is no standard; the uncertainties are large.
The measurement process of the proposed study will consist in rendering objects sensitive by sensors. The system that will be developed during the thesis requires machine learning methods using symbolic representation information and uncertainty management. The challenge is to build this tool which includes the instrumentation of musical instruments, the processing of sensor information fusion and interpretation in order to establish sensory profiles at the output.
The study will have to integrate the measurement of the relevance of the profiles thus obtained and compare them to those obtained using traditional methods, particularly with regard to their interests and the impact during the development of the accompanying project.
The experimentation phase will take place at the Annecy Conservatory of Music, Dance and Theatre, where a music workshop with children with ASD will be held.

Keywords: Artificial Intelligence, ASD, hypo sensitivity, hyper sensitivity, IoT

Supervisor : Mohammed-Reza Salamatian, Stéphane Perrin, Eric Benoit, Sophie Donnadieu

Start date: April 2021

Doctoral school : SISEO " Sciences and Engineering of Systems, Environment and Organizations