Analyzing Trigger-Action Programming for Personalization of Robot Behaviour in IoT Environments

Abstract

The rising spread of humanoid robots in various settings of human life, and their increasing affordability, as well as the massive adoption of the Internet of Things (IoT) in various scenarios have made End User Development (EUD) for robotic and IoT applications an interesting research direction. In particular, in the EUD field, trigger-action rules have become popular for their simple structure, which enables users to create rules to implement their desired personalization. Such rules can be a precious source of information for various goals: understanding the aspects people are most interested in, the types of routines they would like to have, the kind of support/automation they would expect from the robot, and the environment in which the robot is immersed. However, since the number of rules that could be generated using such EUD tools could be significant, manual analysis of rules does not seem a viable solution. In this paper we discuss how a visual analytics tool supporting filtering, exploration and analysis of data generated by a EUD tool for programming humanoid robots immersed in IoT environments can be helpful for deriving relevant information associated with the personalization that users express through rules. The analysis can provide designers and developers of EUD tools and associated customizable applications with useful insights for improving the tools and the robotic applications themselves, and facilitate their adoption.