SCyPSCenter for Smart Cyber-Physical Systems, UMass Lowell
Make a Gift

Research thrusts

People in the loop

When a person and an AI share a decision about a system that can hurt someone, who decides what, and how do they learn to trust each other?

The center's newest thrust, and its growth direction. Every other thrust produces a system that eventually hands a decision to a person: a grid operator during an attack, a driver beside an autonomous vehicle, a clinician reading a model's suggestion. This thrust studies that handoff: trust between people and machines, the ethics of algorithms that act on infrastructure, decision-making under pressure, and how to train the people who will run these systems.

What the group builds

Trust between people and robots

How trust in a robot forms and breaks, why overtrust is as dangerous as distrust, and what a machine should do when it has to be repaired.

Ethics of algorithms that act

What an algorithm may decide about a person in an autonomous vehicle or a hospital, and who is responsible when it is wrong, including the ethics of AI in warfare and health.

Decision-making under pressure

How operators and commanders decide with AI beside them in the moments that matter, and where decision inertia and least-worst choices come from.

Education and workforce

Education research and evaluation for the center's training programs, so the next generation is measured, not assumed.

Recent papers

The newest work from the faculty on this thrust. 32 papers since 2019 carry one of their names.

  1. R. Perkins, B. Berkovich, P. Robinette
    The Cobra Effect in Trust Repair: Unintended Consequences of Rebuilding Trust in Human-Robot Collaboration
    Lecture Notes in Computer Science, pp. 664-670, 2026
  2. T. N. Nagy, Z. Rezaei Khavas, M. R. Kotturu, B. Liefooghe, P. Robinette, M. De Graaf
    A Robot Should Compensate for Its Mistakes: An Exploration of the Dynamics of Trust Violation and Repair Strategies in Human-Robot Collaboration
    ACM Transactions on Human-Robot Interaction, vol. 15, no. 1, pp. 1-34, Oct. 2025IF 7.0 (2025)
  3. E. Huynh, P. Robinette
    Modality Matters: A Sim-to-Real Study of Sonar-Based Object Detection and Tracking
    OCEANS 2025 - Great Lakes, pp. 1-6, Sept. 2025
  4. R. Perkins, P. Robinette
    Beyond Scripted Apologies: Calibrating Trust with Dynamically Generated Responses
    34th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), pp. 2503-2509, Aug. 2025
  5. A. Nickelson, R. Perkins, A. J. London, P. Robinette, K. Tumer
    Multi-objective reinforcement learning framework for beneficent artificial intelligence
    Neural Computing and Applications, vol. 37, no. 30, pp. 24773-24791, June 2025
  6. E. Hedlund-Botti, J. Schalkwyk, N. Moorman, C. Yang, L. Seelam, S. Waveren, R. Perkins, P. Robinette, M. Gombolay
    Learning Interpretable Features from Interventions
    Robotics: Science and Systems XXI, June 2025
  7. Z. R. Khavas, A. Majdi, S. R. Azadeh, P. Robinette
    The Role of Drone Appearance and Capability in Human Trust: A Comparative vs. Isolated Analysis
    22nd International Conference on Ubiquitous Robots (UR), pp. 320-327, June 2025
  8. K. Lowandy, S. Kelliher, D. Le, C. Molinari, I. Harris, L. Unger, R. Fink, C. Shemelya, P. Robinette
    Convolutional neural networks for engineering design validation
    Applications of Machine Learning 2024, art. 17, Oct. 2024

All publications

Other thrusts

Most projects cut across two or three of them.

Resilient energy and infrastructureNext-generation networks and trusted devicesCyber-physical AI