Learning waste recycling by playing with a social robot

In this paper we investigate the use of a social robot as an interface to a serious game aiming to train kids in how to recycle materials correctly. Serious games are mostly used to induce motivations and engagement in users and support knowledge transfer during playing. They are especially effective when the goal of the game concerns behavior change. In addition, social robots have been used effectively in educational settings to engage children in the learning process. Following this trend, we designed a serious game in which the social robot Pepper plays with a child to teach him to correctly recycle the materials. To endow the robot with the capability of detecting and classifying the waste material we developed an image recognition module based on a Convolutional Neural Network. Preliminary experimental results show that the implementation of a serious game about recycling into the Pepper robot improves its social behavior. The use of real objects as waste items during the game turns out to be a successful approach not only for perceived learning effectiveness but also for engagement of the children.