Attention Analysis in Robotic-Assistive Therapy for Children With Autism.

Children with Autism Spectrum Disorder (ASD) show severe attention deficits, hindering their capacity to acquire new skills. The automatic assessment of their attention response would provide the therapists with an important biomarker to better quantify their behaviour and monitor their progress during therapy. This work aims to develop a quantitative model, to evaluate the attention response of children with ASD, during robotic-assistive therapeutic sessions. Previous attempts to quantify the attention response of autistic subjects during human-robot interaction tasks were limited to restrained child movements. Instead, we developed an accurate quantitative system to assess the attention of ASD children in unconstrained scenarios. Our approach combines gaze extraction (Gaze360 model) with the definition of angular Areas-of-Interest, to characterise periods of attention towards elements of interest in the therapy environment during the session. The methodology was tested with 12 ASD children, achieving a mean test accuracy of 79.5 %. Finally, the proposed attention index was consistent with the therapists' evaluation of patients, allowing a meaningful interpretation of the automatic evaluation. This encourages the future clinical use of the proposed system.

Development of an Interactive Total Body Robot Enhanced Imitation Therapy for ASD children.

Autism is a neurodevelopmental disorder in which the available therapies target the improvement of social skills, in order to ensure a high quality of life for the child. The use of Social Assistive Robots offers new therapeutic possibilities in which robots can act as therapy enhancers. IOGIOCO project emerges in this framework: it aims at the development of a Robot- Assisted Therapy protocol for the treatment of Autism Spectrum Disorder, through gesture training. The definition of these gestures and their recognition by the robot are parameters that directly affect the engagement of the children. However, the design of a protocol becomes harder in a highly unconstrained environment. Therefore, the current work aims at expanding the gesture set and improving the gesture recognition algorithm available in the IOGIOCO platform. More specifically, total body gestures have been added to the available upper limbs movements, and a custom Activity Detection method has been developed, which allows the identification of the time window in which a gesture is performed. The insertion of this method on a recognition algorithm based on a ResNet, a particular kind of Convolutional Neural Network, improved its F1-score from 57% obtained with the previously-available version, in a dataset of ASD children, to 76%, demonstrating the effectiveness of the Activity Detection method. Furthermore, the expansion of the interaction possibilities to total body movements was positively evaluated by the clinical staff, increasing the engagement of patients and the set of possible trained skills. Therefore, the results of the current work are encouraging. To reinforce the conclusions drawn, the proposed algorithm should be tested in real time on several autistic children within a complete Randomized Clinical Trial, also to study the effectiveness of this type of treatment. From the technical point of view, further improvements of the developed methodology should tackle the remained issues, such as further increasing the recognition capability, especially in the transitions from sitting to standing, that proved to be a hard task for the developed method.