Human Experience & Agent Teamwork Lab

Research

Our mission is to create techniques for human-robot teaming: the flexible coordination of complex, interrelated activities in shared spaces, so that humans and agents together accomplish what neither could alone.

Human-robot teamwork

How should robots and AI agents share work with people? We build scheduling approaches for mixed human-robot teams and test them with real people and real robots, measuring trust, fluency, and who should take the lead.

Past projects

CHAI: Cultivating Human-Agent Interactions

As collaborative robots become commonplace, creating schedules for human-robot teamwork will be crucial. This project tested scheduling algorithms developed for robot teams on human-robot teams, using Rethink Robotics' Sawyer to study how people and robots complete tasks together, and how human autonomy affects a team's productivity.

Students: Brenda Castro, Montana Roberts

• Who takes the lead? Automated scheduling for human-robot teams (2017)

Human-Robot Trust

How much a person trusts a robot shapes how they interact with it. Using a coin entrustment game (a variant of the prisoner's dilemma) and our NAO robot, we measured trust and cooperation as separate phenomena, comparing how people treat robots versus other humans.

Students: Jane Wu, Erin Paeng, Kari Linder

• Trust and Cooperation in Human-Robot Decision Making (2016)
• Exploring human-robot trust and cooperation (2016)
• Human-Robot Trust and Cooperation Through a Game Theoretic Framework (student abstract) (2016)

MAD about Fluency

What makes a human-robot collaboration feel fluent? We built MAD-TN, a tool for measuring fluency, and used it to study human temporal uncertainty and the fluency of human-robot interactions.

Students: Seth Isaacson, Gretchen Rice, Maya Abo Dominguez, William La, Emily Weiss, Zeneve Jacotin, Amy Yuan

• MAD-TN: A Tool for Measuring Fluency in Human-Robot Collaboration (2019)
• Modeling Human Temporal Uncertainty in Human-Agent Teams (2020)
• Analyzing the Fluency of Human-Robot Interactions (2023)

Improvisational Teamwork

Simple temporal networks that let people and agents coordinate on the fly, plus analysis of how dynamic control holds up when uncertainty is skewed.

Students: Malia Morgan, Julianna Schalkwyk, Yuki Wang, Hannah Davalos, Ryan Martinez, Vibha Rohilla, Rosy Chen, Eryn Ma, Ingrid Wu

• Simple Temporal Networks for Improvisational Teamwork (2022)
• Simple Temporal Networks for Improvisational Teamwork (2022)
• Sensitivity Analysis for Dynamic Control of PSTNs with Skewed Distributions (2023)

PaWPal: Productivity and Wellness Pal

College students often struggle to balance work with wellness. PaWPal, an Android app, used experience sampling to model when users are most likely to experience flow and to nudge them toward more productive, immersive work. It was the precursor to the campus-wide Workload and Health at Mudd (WHAM!) study.

Students: Priya Donti, Honey Rosenbloom, Alex Gruver

• Predicting the quality of user experiences to improve productivity and wellness (2015)
• Exploring active and passive team-based coordination (2014)

See the papers →

Plans that survive the real world

Robots slip and people run late. We develop the math and algorithms behind temporal plans that stay flexible, durable, and controllable when things don't go as expected.

Past projects

Probably-in-Control

Sometimes scheduling uncertainty outstrips an agent's ability to control for it. We developed analytical tools for the “degree of controllability”: the likelihood that an agent, such as a robot, can still control for uncertainty caused by slippage or localization error.

Students: Shyan Akmal, Savana Ammons, Maggie Li, Michael Gao, Lindsay Popowski

• Quantifying degrees of controllability in temporal networks with uncertainty (2019)
• Quantifying controllability in temporal networks with uncertainty (2020)

Ctrl-Alt-Repeat

Dynamic control strategies for probabilistic simple temporal networks: deciding when to act as uncertain events unfold.

Students: Michael Gao, Lindsay Popowski

• Dynamic Control of Probabilistic Simple Temporal Networks (2020)

PolyBots & Acrobotics

The constraints in a temporal plan define a polytope bounding the space of possible schedules, and its geometry says something about the plan. PolyBots used geometric properties to quantify a plan's flexibility; Acrobotics introduced “durability,” a measure of a schedule's resilience to disturbances.

Students: Amy Huang, Liam Lloyd, Joon Lee, Viva Ojha

• New perspectives on flexibility in simple temporal planning (2018)
• Measuring and optimizing durability against scheduling disturbances (2019)

See the papers →

Multi-robot coordination

When a team of robots shares a plan, every delay ripples. We study how robots can schedule robust interactions, and how much communication and rescheduling that really takes.

Past projects

Robot Brunch & DREAM

We want multi-robot plans that adapt to events beyond the robots' control, like slippage or sensor failures. This project introduced robustness, the likelihood that a multi-robot plan succeeds, and DREAM, a single streamlined algorithm that trades robustness against computational overhead, including multi-agent communication.

Students: Jeb Brooks, Emi Reed, Alex Gruver, Kyle Lund, Sam Dietrich, Scott Chow, Liam Lloyd, Jordan Abrahams

• Robustness in probabilistic temporal planning (2015)
• Robust multi-robot scheduling (2016)
• Robust execution strategies for probabilistic temporal planning (student abstract) (2016)
• Robust execution of probabilistic temporal plans (2017)
• Trade-offs between communication, rescheduling, and success rate in uncertain multi-agent schedules (2018)
• Dream: An algorithm for mitigating the overhead of robust rescheduling (2019)

See the papers →

Alumni

Led by Prof. Jim Boerkoel. These are the students who made HEATlab what it is, and where their projects took them.

Of our alumni, 5 have earned PhDs, 1 has earned a master’s degree, 7 are in PhD programs now, and 3 are in master’s programs.

Class of 2025

• JH

  Julia Hansen

  MAD about Fluency

  MBAn student, MIT Sloan

Class of 2024

• RC

  Rosy Chen

  Improvisational Teamwork

  MS student, Computer Science, Carnegie Mellon

• 

  Zeneve Jacotin

  MAD about Fluency

  Associate Project Manager, Motivo

• 

  Eryn Ma · Pomona

  Improvisational Teamwork

  PhD student, Software and Societal Systems, Carnegie Mellon

• EW

  Emily Weiss

  MAD about Fluency

  PhD student, Computer Science, University of Southern California

• AY

  Amy Yuan

  MAD about Fluency

  MS student, Computer Science, University of Illinois Urbana-Champaign

Class of 2023

• 

  Malia Morgan

  Improvisational Teamwork

  Predoctoral Young Investigator, AllenNLP, Ai2

• 

  Julianna Schalkwyk

  Improvisational Teamwork

  Graduate student, Georgia Tech

• 

  Ingrid Wu

  Improvisational Teamwork

  Research Assistant, MIT CSAIL

Class of 2022

• 

  Maya Abo Dominguez

  MAD about Fluency

  Software Engineer II, Data Engineering, Samba TV

• 

  William La

  MAD about Fluency

  Senior Software Engineer, Meta

• 

  Ryan Martinez

  Prob-in-Ctrl

  PhD student, Mathematics (PDEs), UC Berkeley

• 

  Vibha Rohilla

  Prob-in-Ctrl

  Software Engineer, Google

• 

  Yuki Wang

  Improvisational Teamwork

  Graduate student, Cornell University

Class of 2021

• 

  Hannah Davalos

  Prob-in-Ctrl

  PhD student, Industrial Engineering & Operations Research, UC Berkeley

• 

  Seth Isaacson

  MAD about Fluency

  PhD student, Robotics, University of Michigan

• 

  Lindsay Popowski

  Ctrl-Alt-Repeat

  PhD student, Computer Science, Stanford

Class of 2020

• 

  Savana Ammons

  Prob-in-Ctrl

  Computational Mathematician, Mobius Logic

  PhD, University of Illinois Urbana-Champaign

• 

  Michael Gao

  Ctrl-Alt-Repeat

  AI/ML Engineering Lead, Adobe

  MS, Management Science and Engineering, Stanford, 2022

• 

  Joon Lee

  Acrobotics

• 

  Tiffany Madruga

  MAD about Fluency

• AN

  Athena Paraskevas-Nevius

  MAD about Fluency

  Senior Member of Technical Staff, Salesforce

• 

  Gretchen Rice · Olin

  MAD about Fluency

  Prevention education; former Fulbright English Teaching Assistant, Taiwan

• 

  Megan Robalewski · Scripps

  MAD about Fluency

Class of 2019

• 

  Jordan Abrahams

  Robot Brunch

  Senior Software Engineer, Google

• SA

  Shyan Akmal

  Prob-in-Ctrl

  PhD, Computer Science, MIT, 2024

• 

  Morgan Frisby

  HRI in Retail

  Software Engineer, Avarra

• 

  Maggie Li

  Prob-in-Ctrl

  PhD student, Operations Research, Cornell

• 

  Viva Ojha

  Acrobotics

  Senior Software Engineer, Bloomberg

• 

  Brenner Ryan

  Robot Brunch

  Software Engineer, Carlson Medical Technologies

Class of 2018

• 

  Brenda Castro

  Human-Robot Teamwork

  Licensed Mental Health Counselor, Good Therapy NYC

• 

  Amy Huang

  PolyBots!

  Software Engineer

• 

  Liam Lloyd

  PolyBots!

  Senior Software Engineer, Permanent Legacy Foundation

• 

  Montana Roberts · Scripps

  Human-Robot Teamwork

  Security, Datadog

• 

  Jane Wu

  Human-Robot Teamwork

  Postdoctoral Fellow, EECS, UC Berkeley

  PhD, Computer Science, Stanford

Class of 2017

• 

  Scott Chow

  Robot Brunch

  Principal Software Engineer, Berkshire Grey

  PhD, Robotics, Oregon State, 2024

• 

  Sam Dietrich

  Robot Brunch

  FPGA Engineer, Microsoft

• 

  Kyle Lund

  Robot Brunch

• 

  Erin Paeng

  Human-Robot Teamwork

• 

  Emi Reed

  Robot Brunch · PaWPal

Class of 2016

• 

  Sam Echevarria

  PaWPal

• 

  Kari Linder · Claremont McKenna

  Human-Robot Teamwork

  Principal Design Lead, Cloudflare

• 

  Emma Meersman

  PaWPal

  Software Engineer, Google

Class of 2015

• 

  Priya Donti

  PaWPal

  Assistant Professor, MIT EECS & LIDS

  PhD, Computer Science & Public Policy, Carnegie Mellon

• 

  Alex Gruver

  PaWPal · Robot Brunch

  Co-Founder, Jhourney

• 

  Honey Rosenbloom

  PaWPal

  Machine Learning Engineer, Allegheny County Department of Human Services

Class of 2014

• 

  Jeb Brooks

  Robot Brunch

  Embedded Support Consultant, Green Hills Software