BeatriXR is a comprehensive and adaptive Virtual Reality (VR) toolkit developed in Unity using MRTK3 (Mixed Reality Toolkit 3). It is designed to simplify the creation and visualization of direct feedforward—information that informs users about the expected result of an action before they perform it.
BeatriXR combines a modular VR component library mapped to established feedforward design spaces with an LLM-powered decision-support layer. It assists designers at design time in configuration selection and adapts to trainee behavior and environmental noise at runtime to offer context-aware guidance.
For more information about this project please refer to our paper: BeatriXR: Comprehensive and Adaptive Feedforward Support for Guidance in Virtual Reality
The toolkit consists of the following key Unity scripts:
- ProcedureEngine.cs: Manages step-by-step procedural execution, tracking task subjects, interactable targets, and modalities.
- ContextEngine.cs: Analyzes MRTK3 hand tracking, controller connectivity, and environmental noise events (ContextEvents) to compute alternative task pathways.
- WeightedStepPriority.cs: A priority-ranking algorithm that weighs interaction distance, modality transition costs, and sensor feedback to determine the best interaction modality.
- FeedforwardSettings.cs: Holds the active parameters mapping directly to the Triggering, Previewing, and Exiting design spaces.
- ChatGPT.cs: Manages requests to the OpenAI API to dynamically adjust feedforward presentation style based on user context.
- UIHandler.cs: Renders interactive 3D menus, target indicators, comparison plates, and options scroll lists in MR.
- RewindableAvatar.cs: Controls the humanoid helper avatar, playing, looping, or rewinding task animation previews.
- SensorEmulator.cs: A developer-focused utility to trigger noise or controller failure states for system debugging.
If you use BeatriXR in your research, please cite our paper:
@article{10.1145/3816419,
author = {Artizzu, Valentino and Luyten, Kris and Rovelo Ruiz, Gustavo Alberto and Spano, Lucio Davide},
title = {BeatriXR: Comprehensive and Adaptive Feedforward Support for Guidance in Virtual Reality},
year = {2026},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3816419},
doi = {10.1145/3816419},
journal = {Proc. ACM Hum.-Comput. Interact.},
series = {EICS 2026}
}