A study of cognitive development in infants based on sensor data from smart toys

Aim: We intend to conduct a longitudinal study in which we will recruit 4-5 children and use multiple smart toys and a chatbot system to record sensorimotor skill data daily from 3 months to 3 years of age.

Social/scientific motivation: Parents are always worried about the growth of their babies, as well as the development of their personality. With this toy, parents can not only know how often their children play with the toy, but how the child interacts with it and develops. Parents can chat with their children by message apps such as Telegram that sends commands to the smart toy.

Scientific background: Longitudinal data with a high temporal resolution, following the same child over years, is rare. Currently, we are only aware of Deb Roy’s pilot study where he recorded 230,000 hours of audio-video recordings spanning the first three years of his child’s life at home in order to study how language develops. We plan to study how children learn sensorimotor skills by recording their interaction with our smart toy almost every day during the first three years. With this understanding of the human cognitive development and we will implement a new learning mehtod for artificial neural netwoks, that requires less data and mimics the way kids learn.

Alan’s motivation: I think this project is the beginning of IoT. We start by collecting and analyzing the data from toys. This project can let me know how to build a IoT system, IMU algorithm and children behavior.

Butters’ motivation: This is the first smart toy in the world that connect with Chatbot, using message apps to control and receive the signals of the toy. I think this is a state of art IoT system and worthy of making it bigger and wider, to show the world Chatbot can do more things rather than just chatting.

Members: Alan Yu, Butters Wu, Torbjörn Nordling