Activity Analysis and Detection of Repetitive Motion in Autistic Patients

There is no doubt today that autism is both neurodevelopment and behavioural - a multifactorial disease characterized by impaired social skills and lack of efficient communication. In order to support more efficient development and improve communication, many directions in science have been and still are on clinical trials. We compared repetitive motion patterns with patterns characteristic of normal everyday activities and isolated the key elements of each that allow distinguishing between the two motion patterns. The Principal Component Analysis technique was used to reduce the dimensionality of the motion data collected. A kNN classifier was also used to determine the two classes of motion apart with 100% sensitivity and specificity. We present here a uniquely designed alarm system that can help with autistic children´s difficulties on a real-time basis. The uniqueness of the system is that it can be extended or applied to other situations of even non-autistic children that suffer some health disturbances or particular situations that need a specific assistance. Beside the system´s design description and methodological details, further work on improving it to the higher level is proposed.