HUDDLE: Human Autonomy Teaming in Uncertain and Dynamic Environments

The goal of this project is to develop and demonstrate theoretical foundations, models, and principles to facilitate effective HAT design. We propose dynamic, temporally-situated conceptual and computational models of HAT, and adopt a multi-pronged approach to teaming, incorporating cognitive, behavioral, and social factors. Our approach is trans-disciplinary, drawing on concepts from robotics, AI, cognitive science, organizational psychology, and human factors. We will explore how teams reason, coordinate, and trust, particularly with regard to how teams change over time.

This project will address key questions in HAT, including:

  • Modeling and operationalizing team mental models, for example, looking at how human and robotic teammates can collaboratively achieve a shared understanding of knowledge of tasks and the environment.

  • Understanding how team dynamics change over time, including when robots take actions to calibrate and maintain their trust in one another, when they jointly act together, and when they engage in shared decision making. Researchers will study teaming temporally, both in real time and over time.

  • Exploring how two key mediators, coordination and team cognition, affect team performance. Both are understudied in HAT research, yet very important to team outcomes.

This research will advance the state-of-the-art and state-of-the-science in HAT. By creating new methods that dynamically and temporally model team evolution and scientifically establishing how autonomy affects human teammates, this work will yield new insights for HAT in uncertain environments. Ultimately, this project will: (1) Create innovative, cognitively-informed conceptual frameworks and computational models of teaming, (2) Establish core elements of a new interdisciplinary science of human agent teams, (3) Accelerate the advancement of HAT in mission critical environments, (4) Pioneer rapid prototyping and rigorous field testing of new guidelines and interventions for HAT design.


Laurel Riek (PI), Computer Science and Engineering, UC San Diego

Susan Simkins, Psychology, Penn State University

Ericka Rovira, Enginering Psychology, West Point

Tom Griffiths, Psychology and Computer Science, Princeton

Francesco Bullo, Mechanical Engineering, UC Santa Barbara

Angela Yu, Cognitive Science, UC San Diego

Vaibhav Srivastva, Electrical and Computer Engineering, Michigan State University

Vijay Gupta, Electrial and Computer Engineering, Purdue University