Playground Social Interaction Analysis using Bespoke Wearable Sensors for Tracking and Motion Capture

Abstract

Unstructured play is considered important for the social, physical and cognitive development of children. Traditional observational research examining play behaviour at playtime (recess) has been hampered by challenges in obtaining reliable data and in processing sufficient quantities of that data to permit credible inferences to be drawn. The emergence of wearable wireless sensor technology makes it possible to study individual differences in childhood social behaviour based on collective movement patterns during playtime. In this work, we introduce a new method to enable simultaneous collection of GNSS/IMU data from a group of children interacting on a playground. We present a detailed description of system development and implementation before going on to explore methods of characterising social groups based on collective movement recording and analysis. A case study was carried out for a class of 7-8 year old children in their school playground during 10 episodes of unstructured play. A further 10 play episodes were monitored in the same space following the introduction of large, loose play materials. This experimental design allowed us to study the effect of an environmental intervention on social movement patterns. Sociometric analysis was conducted for comparison and validation. This successful case study demonstrates that sensor based movement data can be used to explore children’s social behaviour during naturalistic play.

Using sensors to study the social dynamics of outdoor play.

“…An important next step for research involving this type of sensor data is therefore 20 comparison and validation against play types observed using traditional methods. In our 21 previous work with these data we identified 'hang out play', 'focal play' and 'pursuit play' as 22 movement-based cues to play types and we plan to investigate whether these play types can be 23 automatically detected in the data (Heravi, et al 2018). The ability to classify more abstract 24 concepts such as fantasy play, competitive play and games with rules would be useful future 25 additions to enable a deep understanding of how play activities link to social development.…”

Mining Exceptional Social Behaviour

“…Analyzing social interactions in the playground can be of utmost importance. Social group structure and dynamics are believed to be strongly related to the child well-being and yet has been poorly understood and studied [16].…”

“…One dataset, playgroundA [16], has the geographic position of 18 7-8 year-old children (9 girls) over time, during approximately 45min. It also includes the personal attributes (gender, age, emotional stability etc).…”

“…Furthermore, as people make more and more use of new technology, a great amount of data from their behaviour is being collected [33]. In addition, deliberately gathering data from social and ubiquitous environments through sensors (of proximity or geo-localization) is also being used to study the behaviour of people without interfering with their actions [16], e. g., movement data [23], on a more naturalistic basis. More specifically, it was applied to the domain of social interactions [26], where the authors analyze the properties (mainly focusing on gender phenomena) of five-year-old kids' social interactions.…”

“…The data was collected with the use of location sensors during the school breaks. One dataset, playgroundA [16], has the geographic position of 18 children over time, in 10 different sessions and personal attributes (gender, age, emotional stability etc). The other dataset, playgroundB [26] has the position of 16 children and sociodemographic attributes, such as gender and age.…”