• Autism
• Eye tracking system
• Neural Networks

Individuals with Autistic Spectrum Disorders (ASDs) have impairments in the processing of social and emotional information. The
number of children known to have autism has increased dramatically since the 1980s. This has sensitized the scienti¯c community to the
design and development of technologies suitable for treating an autistic patient in order to broaden the emotive responsiveness, such as
the employment of robotic systems to engage proactive interactive
responses in children with ASDs.
My PhD work focuses on the design and develop of new technologies
for therapy with individual affect by ASD. The main challenge of my
work has been to design and develop a novel control architecture able
to reproduce the brain characteristics in terms of high concurrency processing, flexibility and the ability to learn new behavior. The
main di±culties in implementing Artificial Neural Networks (ANNs)
in hardware in terms of accuracy, gate complexity and speed performance are discussed.
A new wearable eye tracking system able to investigate attention
disorders early in infancy is proposed. Technological choices are emphasized with respect to unobtrusive and ecological to adapt the
device for infants. New algorithms to increase the system robustness under illumination change and during calibration process have
been developed and herein presented. Experimental test prove the
effectiveness of the solutions.
Considerations on the future research directions are addressed, stressing the multiple application fields of the designed device.