Automating Google Workspace (GWS) & more with Apps Script
Development and evaluation of a social robot platform for therapy in
1. Development and evaluation of a social robot platform for therapy inautism Daniele Mazzei, University of Pisa EMBC11Boston 29th August - 3thSept 2011
2. ASDs and Robotics One of the main difficulties in subjects with autism spectrum disorders (ASDs) is their inability to understand and analyze the emotional state of their interlocutor. Recent research shows that ASDs perceive robots not as machines, but as their artificial partners
3. Our Hypothesis A therapist guided interaction between a robot able to express emotions and ASDs can help them to improve social skills. The FACE of Autism, Mazzei et. All, ROMAN2010, Viareggio Italy, Sept. 2010
6. The FACE ‘mood’ Emotion Cartesian Space (ECS) based on the Posner theory called ”The circumplex model of affect”
7. FACET platform FACE android Gaze tracking hat Sensorized T-Shirt Face Control Unit (Bodymap and brain) Face tracking and facial expression recognition Inertial platform Room motorized cameras Face configurator and control interface Environmental microphone
9. Physiological signal acquisition Based on e-textiles HR, HRV, RR, skin conductance, skin temperature, respiratory rate and accelerations All this signals can be correlated with subject emotional states Made by Smartex Srl, Italy Valenza, G et all., Towards a smart glove: Arousal recognition based on textile ElectrodermalResponse, EMBC10
10. HATCAM Eye Tracker Wearable and specifically designed to investigate early attention disorders in infants Hat or Band mounted system fast and easy calibration procedure Mirror allows to use images of both eyes (3D gaze point identification) A.Armato, Implementation of new assistive technologies for people affected by Autistic Spectrum Disorders ASDs", Pisa May 2011. (Patent: TO2008A000862)
12. The FACET protocol Phase A:Familiarization with the FACET room and the android Phase C:Conversation with psycologist about FACE Subject – Robot interaction 20 min 0 min Phase B:Exposition and inteprretation of FACE’s expressions Phase D:Conversation with FACE (if starts)
13. Preliminary results At the moment this therapy/paradigm has been tested with 4 ASD subjects and 2 controls No subjects showed discomfort or fear during the therapy increased cardiac sympathetic activity was observed in ASDs.
14. Conclusions Well accepted by ASDs Could have a wide range of applications in the study of non-verbal communication and expression. Can become an innovative therapeutic approach aimed to gradually increase social cognitive skills of ASDs.
15. Future Works On going experiments on control and ASDs More tests of the FACET hardware/software infrastructure Using of the FACET platform to perform generic studies about human-robot empathic link
16. Future Works Complete and integrate the face adaptive control algorithm trying to close the loop
17. Thanks For Your Attention Questions? www.faceteam.it mazzei@di.unipi.it
Notes de l'éditeur
This talk is about an innovative acquisition platform for Human Robot interaction studies used in Children with autism therapy
One of the main difficulties in people with Autism Spectrum Disorders is their inability to understand and analyze the emotional state of their interlocutor and consequently to correctly manage interpersonal relationships and common life scenarios.Recent research shows that ASD subjects perceive robots not as machines, but as their artificial partners.There are various famous works about robot and autism. for example: Robota a robotic dols developed by Dautenhahn group or Kismet a robotic head developed by Breazeal at MIT and other works like the one of Picard, of Plaisant and othersAll this projects are focused to investigate human social development with special focus on interpersonal and human-robot communication and interaction
Our hypotesis is that Therapy where a therapist drive an emotional interaction between a children with autism and a robot able to change itsbehavior (simulating a human ”mood”) can help Authistic subjects to improve their social skills…and more importantly can help them to be able to better interpret other people mental states which is one of the main problem of Autism.We think that robot Facial expressions are more easily accepted by autistic patients than human actors expressions because they are simple, repeatble and have a stereotypical nature.
But in order to test our hypotesys we need a system able to vehiculate emotion as facial expression But These innovative therapeutic approach also require an observation method that allow the analysis of the subject’s state and behavior during interaction with the robot. In order to understand if the therapy is able to “activate” the subject
FACE is an android female face used as emotion conveying system in the FACET platform.It has been made, by Hanson Robotics and is actuated by 32 servo motors moving the artificial skin through cables acting as tendons, whichallow human facial expressions to be re-created.The android has also a camera in the right eye used for the subject face tracking.
FACE expressions are represented as emotions or ”moods” usinga emotion rapresentationcartesian space tha has been developed by posner in his theory called ”The circumplex model of affect” In the ECS are represented as point of a cartesian space where the x coordinate represents the Valence and the y coordinate the arousal.This mood representation approach has been previously used also in other social robots as Kismet, Eddie and Probo allowing This approach allows also, starting from the six basic Ekman Emotions, to fill the entire ECS space interpolating and creating a cotniuousexpressions space
FACET Is a modular acquisition platform based on a multi sensory room in which the guided therapy with the FACE android takes place.This room is Equipped with various sensors such us motorized cameras, directional microphones, sensorized shirt, eye tracker and interfaces for controlling the robot and supervising the therapy.
As previously mentioned another important part of the FACET equipment is the Sensorized shirt. This shirt is an innovative physiological signal analysis system that allows to acquire ECG, respiratory rate, skin conductance and temperature of the subject during the therapy.This shirt is made using an innovative woven technology that allows to embed the electrodes directly in the fabric with a consequent minimal subject discomfort that is very important in autistic children therapy.Electrodes Signals are acquired and sent to a computer through a dedicated wireless unit that is also placed on the shirt
Another important element of the FACET platform is the eye tracker HATCAM. Hatcam is a bi-ocular eye tracking system based on a hat equipped with a wireless camera a mirror and an inertial platform.The system allows to acquire the 3Dimensional subject gaze point and not only the gaze direction as in the case of the mono-ocular eye tracking systems.Eye tracking systems are very important in autism studies because there are evidences that Autistic have Atypical gaze patterns…..In particular ASD subject tends to reduced eye-region fixation time in favour of an increased focus on mouths and objects
this therapy is based on a stepwise protocol were common life scenarios are reproduced and the subject is guided to interpret and understand the android mood
At the moment we are working on collecting data using the FACET platform and the android with the ECS facial generation system.Unitl now this therapy/paradigm has been tested with 4 ASD subjects and 2 controls and no subject shown any sign of disconfort or fear during the therapyDue to the very low number of subjects and to the wide range in ages, a statistical analysis was impossible and these data have to be considered as proofof the platform’s ability to perform HRI studies on children with autism.We are analysing data in order to find correlation between physiological signal and therapy phases and also differences between ASDs and controls.In particular HR time-domain parameters analisys shown that the mean HR values is reduced in ASDs subjects with respect to controls during the entire session, showing that patients were not disturbed by the presence of the android.the comparison between the values of LF/HF ratio of the two groups showed a trend of higher LF values in ASD subjects and this result are in accord with literature in which an increased cardiac sympathetic activity was observed in ASDs.
In conclusion.....is well accepted by children with autismThis system can have consequently a wide range of use also in context different from the autism.For example could be used as generic platform for human-robot interaction studiesOr as robotic behaviour testing systemObviously we need to test more the adaptive control algorithm and also the therapeutic protocol with more autistic and also control subject......