27577
Cloud Computing Enabled Social Robot Platform for Children with Autism Spectrum Disorders

Poster Presentation
Friday, May 11, 2018: 10:00 AM-1:30 PM
Hall Grote Zaal (de Doelen ICC Rotterdam)
Z. Tan1, B. Xing2, W. Song2, W. Cao1, H. Zhu2,3, L. Yi4 and J. Chen5, (1)South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou, China, (2)Centre for Optical and Electromagnetic Research, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou, China, (3)Child Developmental & Behavioral Center, Third Affiliated Hospital of SUN YAT-SEN University, Guangzhou, China, (4)School of Psychological and Cognitive Sciences and Beijing Key Laboratory of Behavior and Mental Health, Peking University, Beijing, China, (5)School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, Stockholm, Sweden
Background:

Children with autism spectrum disorder (ASD) have a core problem in terms of social communication deficits. Recent advances in social robots, which are capable of interacting with humans by following social behaviours and rules attached to its role, open a way to promote engagement and communicative exchange in children with ASD. Many researches have demonstrated that social robots can offer children with ASD a safe environment, and facilitate diagnosis and assessment. However, it is of key importance that the social robot provides accurate response to children with ASD, which might be difficult to be realized with limited computation power offered by the robot itself.

Objectives:

We are aiming at developing cloud computing enabled social robot platform that engages in developing social communication skills of children with ASD. The platform facilitates fast and reliable connections between the social robots and remote data centres that provide sufficient computation capability in supporting timely and precisely interactions with children.

Methods:

Figure 1 presents the schematic diagram of our developed cloud computing enabled social robot platform. Two inset pictures are the examples to illustrate how the social robots are interacted with the participating child. The cloud enabled platform is composed of a web client, social robots, external equipment for data collection, and data centres. All collected data (that could be from the social robots and/or external equipment) is then sent to data centres, which can be either third-party cloud or operated locally, through WiFi or cable connections. The web client used for management can view the information sent to the data centres and communicate remotely with the robot.

Results:

We have tested automatic voice response that is carried out by our platform. All the measurements were done in Chinese. The test question is “what is your name?”. The question itself lasts 1.39s. 7 different answers, all of which address questions precisely, were received for 7 measurements. The length of the answering audio varies from 1.78s to 6.31s. The total waiting time, which is defined as the time duration from the end of question to the beginning of answering, varies from 3.20s to 5.77s. More than 90% of the waiting time is spent for the social robot uploading the audio file to the data centres. It is because of very limited data rate provided by the network interface embedded in the social robot. To further reduce the latency, the uploading data rate should be improved. On the other hand, compared to our developed cloud enabled platform, the stand-alone social robot performs less stable and is not always able to provide proper answers, which implies it might not be an appropriate solution to interact with children with ASD.

Conclusions:

We developed cloud computing enabled social robot platform, which provides fast and reliable interactions, and assists in diagnosing and assessment, for developing social communication skills of children with ASD. The developed platform is able to provide automatic and precise response to the participants, which is hardly realized by the stand-alone social robots.