Portfolio / 2026Athens, Georgia

Robotics · Human–AI trust · Intelligent manufacturing

I study what happens when people and intelligent machines share the work.

I’m JB Frericks, an engineering PhD candidate building physical, virtual, and AI-enabled systems that make human–machine trust observable.

Virtual search environment with robot partners
Field study 04VR search environment
01SystemsBuilt for real interaction
02SignalsBehavior to physiology
03TrustMeasured in context
Build the environmentObserve the interactionMeasure the signalDesign better trust

Technology is rarely just technical.

“The most interesting part begins when a machine’s decision changes a person’s next move.”

I work across robotics, virtual reality, simulation, and intelligent manufacturing to understand how people monitor, question, and rely on AI-enabled teammates.

The work is deliberately hands-on: Unitree robots, Unity environments, ROS/Gazebo stacks, edge vision-language models, experiment interfaces, and synchronized multimodal data.

Read the full résumé
NowEngineering PhD candidateUniversity of Georgia
FocusHuman–AI collaborationTrust, reliance & teaming
MethodBuild to understandPhysical + virtual testbeds
NextDecember 2026Expected completion

Working systems, not just ideas.

Each project is a purpose-built environment for asking a specific question about intelligent machines and the people working alongside them.

01 / Embodied trust2025–Present

VR Human–Robot Search Team

A multi-building virtual world where participants plan, search, compare evidence, and make consequential choices with robot partners.

  • Unity + Meta XR
  • Gaze, speech & decision logging
  • Target: CHI 2027
Interior clue room in the VR human-robot study Inside the experiment / 01
02 / Controlled simulationIEEE CogMI 2024

Human–AI Trust Testbed

A ROS/Gazebo environment built to study trust across a sequence of robot navigation, puzzle, maze, and wire tasks.

  • ROS + Gazebo
  • Repeatable task control
  • Best Paper Award
Gazebo human-robot trust experiment interface Puzzle target from the trust experiment Controlled environment / 02
03 / Industrial robotics2026–Present

Autonomous Factory Assessment

A Unitree Go2 sensing stack that turns spatial and motion data into repeatable measures of factory-floor organization.

  • Point clouds + costmaps
  • Entropy-based metrics
  • Run-to-run comparison
Autonomous factory assessment research poster Innovation Factory / 03
04 / Edge intelligenceASME 2026

AffordRAG Factory

An explainable framework that translates ambiguous worker requests into grounded actions for a mixed team of robots.

  • Edge vision-language models
  • Knowledge graph grounding
  • Executable robot skills
05 / Published system SCOUT: Autonomous RFID inventory robot Manufacturing Letters

One connected practice

Build. Observe. Translate.

01

Build the encounter

Design the robot, virtual world, task, and interface around the behavior the research needs to reveal.

02

Observe the whole person

Bring behavior, speech, gaze, physiological signals, system telemetry, and self-report into one timeline.

03

Translate the evidence

Turn experimental findings into design principles for safer and more legible autonomous systems.

Recent signals.

Awards, papers, and work currently moving toward publication.

2025
Award

David Dornfeld Manufacturing Vision Award

Manufacturing Empathy: Sensorial AI for Rewriting Human–AI Collaboration

Winner
2024
IEEE CogMI

Trust and Collaboration Testing in Controlled Human-Robot Environments

First author

Best paper
PUB
Manufacturing Letters

SCOUT: An autonomous UHF RFID-equipped robot dog for flexible inventory monitoring

Second author

Published
WIP
Multimodal trust calibration

A Psychophysiological Framework for Trust Calibration in Human–AI Teaming

Heart rate · GSR/EDA · EEG · eye tracking

Active

Open to the next hard problem

Have a machine that needs to work better with people?

I’d be glad to talk about applied AI, robotics, intelligent manufacturing, or the human side of autonomy.