Modeling child behavior in simulation can be a valuable tool for testing robot planning strategies before real-world deployments but creating accurate models is difficult due to a lack of data and unique behaviors across children. We analyzed child attention from our previous study to create two models of infant behavior based on toy interest. The linked github shows images of the behavior tree models we created based on our data annotation. These models and this process could be used to create future models that would be data-driven and able to be used for simulating child attention to toys and robots.

https://github.com/shareresearchteam/infant_simulation_behavior_trees

Associated Papers:

Ameer Helmi, Kristen Koenig, and Naomi T. Fitter, "A Model Child? Behavior Models for Simulated Infant-Robot Interaction," Proceedings of the International Conference on Social Robotics (ICSR), Florence, Italy, 2022.

Ameer Helmi and Naomi T. Fitter, "Using motion expert feedback to design models for infant-robot interaction," Proceedings of the 2020 ICSR Workshop on Child-Robot Interaction, International Conference on Social Robotics (ICSR), Golden, CO, USA (remote), 2020.