Research Sandbox – Work-in-Progress Excursions

Work-in-progress interactive excursions built by Theodore Pavlic to support ongoing research projects, many developed together with graduate students while working toward a paper or testing a hypothesis in the field.

For the more polished, course- and topic-oriented explainers, see the companion Topic Visualizers collection.

These are experimental demonstrations and explorations rather than finished tools – expect rough edges, and expect them to change. I built each one to communicate ideas of my own and to give students, collaborators, and myself a place to experiment with them. They do not disclose any student's unpublished work, though they may show how much of that later work was grounded in ideas I had put forward here.

Sharing these openly is not a release of the underlying ideas into the public domain. The MIT License covers the page code, not the research ideas it illustrates. I retain authorship of those ideas and ask that they not be reused without attribution. If you intend to build on any idea here that has not yet been published, please reach out to me first as a professional courtesy.

Status:

© 2026 Theodore P. Pavlic. Page code released under the MIT License; the underlying ideas are not.

Collective Behavior

  • Y-Tree Trajectory Detrending and Spectral Analysis preview

    Y-Tree Trajectory Detrending & Spectral AnalysisWIP

    Detrend and spectrally analyze the trajectories of collective-transport groups (weaver ants) moving loads vertically up trees, including forked trees, by projecting motion onto the tree.

  • RSF Explorer, Subnest Destination Choice preview

    RSF Explorer – Subnest Destination ChoiceWIP

    Test field hypotheses about how distance interacts with other nest properties in the navigational choices of collective-transport groups and individual ants in polydomous colonies, framed through a resource selection function (RSF) for subnest destination choice.

Multi-Robot Systems

  • V2V Energy-Safe Swarm Explorer preview

    V2V Energy-Safe Swarm ExplorerWIP

    Explore strategies for battery-safe vehicle-to-vehicle (V2V) energy sharing that extend the endurance and capabilities of multi-robot teams, with a related-work review exportable as Markdown and BibTeX.

Reinforcement Learning & Decision Models

  • Q-Learning Social Cue Model preview

    Q-Learning Social Cue ModelWIP

    Explore a connection between conventional reinforcement learning (Q-learning) and a psychological social-cue assay, as a candidate model for the phenomena.

More from Ted Pavlic