NCAlt: Alternatives and Difference Visualizations for Behavior Trees in Game Development Learning

When learning how to develop AI behavior it is common for students to test different ideas before settling on a desired outcome. This functionality is not available in modern behavior tree systems beyond the traditional methods of duplication and conditional execution. We present NCAlt - a visual be...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of the ACM on human-computer interaction 2022-10, Vol.6 (CHI PLAY), p.1-31, Article 245
Main Authors: Hossain, Md. Yousuf, Zaman, Loutfouz
Format: Article
Language:English
Subjects:
Applied computing
Computer games
Computers in other domains
Context specific languages
Empirical studies in HCI
Graphical user interfaces
Human computer interaction (HCI)
Human-centered computing
Interaction paradigms
Personal computers and PC applications
Software and its engineering
Software notations and tools
Visual languages
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:When learning how to develop AI behavior it is common for students to test different ideas before settling on a desired outcome. This functionality is not available in modern behavior tree systems beyond the traditional methods of duplication and conditional execution. We present NCAlt - a visual behavior authoring framework for exploring multiple game AI behaviors. The system also allows selective merging of nodes between multiple behavior tree alternatives and their visual differencing in the game scene and behavior tree views. We present two worked examples which demonstrate how NCAlt can improve the workflow of game development students. We have conducted a user study where NCAlt was compared against NodeCanvas for creating and using alternatives with moderately skilled game development students. Participants rated NCAlt 15% higher than NodeCanvas on the System Usability Scale (SUS). NCAlt was also rated higher on a self-developed questionnaire. We also conducted a semi-structured interview and a detailed longitudinal follow-up and obtained mostly positive feedback for NCAlt. The results suggest NCAlt has a potential to improve the types of workflows with behavior trees that involve exploration for game development students and facilitate learning, as a result. NCAlt was built on top of NodeCanvas - a plugin for the Unity Engine.
ISSN:2573-0142
2573-0142
DOI:10.1145/3549508