Interfaces for alternatives, difference visualizations and collaborative authoring of behavior trees

Abstract

A visual behavior authoring framework for exploring and comparing multiple game AI behavior alternatives is presented. A mixed method study revealed participants rated it higher compared to the control framework. Mostly positive feedback was also obtained in a semi-structured interview and a follow-up qualitative longitudinal study. An extension for collaboration using synchronous and asynchronous modes is also presented. A mixed methods study revealed that collaboration provides improvement over the control condition, but no differences were found between the collaboration modes. In a semi-structured interview, the synchronous mode was revealed to be useful for quick prototyping, while the asynchronous mode – for most other situations. Overall, the results suggest that proposed solutions have a potential to improve the exploration and collaboration for behavior tree development, but further research is necessary.