Robotics & Autonomous Systems

Autonomous F1 CAV CAR located in front of UVA Rotunda with faculty, students, and former university president, Jim Ryan.

Shaping the Future of Intelligent Systems

Link Lab researchers are developing the next generation of robotics and autonomous systems that perceive, learn, and adapt in complex environments. Our interdisciplinary approach integrates AI, machine learning, sensing, and control to create intelligent systems with real-world impact.

FEATURED RESEARCH

  • Flowchart of a closed-loop diabetes management system. A person wears a continuous glucose monitor (CGM) sensor and an insulin pump. Historical blood glucose (BG) levels from the CGM feed into a Bayesian RNN, which predicts future BG levels. Predictions are evaluated by an STL-U Quantitative Monitor and an Uncertainty Calibration module, then passed to an Adaptive Controller that determines insulin dosages delivered through the insulin pump. Dashed arrows indicate training-time processes and solid arrows in

    Autonomous Systems Interacting with Human Operators

    Lu Feng (CS) and collaborators provided a framework for improving the safety of AI systems that interact with people by predicting future outcomes under uncertainty, evaluating those predictions against formal safety requirements, and adjusting control decisions accordingly. Demonstrated in two case studies (Type 1 diabetes management and semi-autonomous driving applications), the approach improved both safety and overall system performance in human-AI interactions.

     

    PUBLICATION: Quantitative Predictive Monitoring and Control for Safe Human-Machine Interaction

  • Diagram of a manufacturing system integrated with multiple energy sources. The top section shows machines, buffers, a source, and a sink representing production demand, while the bottom section shows a utility grid, solar panels, wind turbines, a generator, and a battery connected through energy flows that supply power to the manufacturing process.

    Integrated Energy Optimization in Manufacturing

    Cindy Chang (MAE), Zongli Lin (ECE), and their students examined how microgrids can be more effectively integrated into manufacturing environments by jointly optimizing energy supply and production demands while accounting for energy storage degradation and operational constraints. Using a multiagent reinforcement learning approach, the study developed practical control strategies that improved system performance and demonstrated their effectiveness in a manufacturing case study.

     

    PUBLICATION: Integrated Energy Optimization in Manufacturing Through Multiagent Deep Reinforcement Learning: Holistic Control of Manufacturing, Microgrid Systems, and Battery Storage

  • Diagram illustrating six navigation scenarios for autonomous robots or pedestrians: (a) doorway, (b) intersection, (c) hallway, (d) L-corner, (e) blind corner, and (f) crowded traffic. Colored circles and arrows represent moving agents and their directions, while red X symbols mark potential conflict or collision points.

    Multi-Robot Navigation in Social Mini-Games

    Chandra, Rohan (CS) examines the challenge of enabling autonomous robots to safely and efficiently navigate crowded, confined spaces, such as hallways, doorways, and intersections, where they must interact with people and other robots. It introduces a unified framework for classifying and evaluating existing approaches, helping researchers and practitioners better compare methods and identify key opportunities for future advances in robot navigation.

     

    PUBLICATION: Multi-robot navigation in social mini-games: definitions, taxonomy, and algorithms

     

Afsaneh Doryab

Barry Johnson

Brian Park

Brian Smith

Daniel Quinn

Felix Lin

Frank Lagor

Haibo Dong

Haiying Shen

Homa Alemzadeh

Kun Qian

Laura Barnes

Lu Feng

Madhur Behl

Matthew Bolton

Nicola Bezzo

Qing (Cindy) Chang

Rohan Chandra

Sarah Riggs

Shangtong Zhang

Streit Cunningham

Somayeh Asadi

T. Donna Chen

Tariq Iqbal

Tomo Furukawa

Ye (Sarah) Sun 

Yen-Ling Kuo

Yixin Sun

Zongli Lin

Search publications by keyword, author, year, or journal to learn more about the research emerging from the Link Lab.

Publications

Faculty Spotlight Video

Screenshot from a video interview showing a person speaking in a workshop or garage with a race car blurred in the background. On-screen text identifies the speaker as “Prof. Madhur Behl, Cavalier Autonomous Racing Team Principal,” and video captions appear at the bottom.

Explore More Robotics & Autonomous Systems Research

DAN QUINN | HAIBO DONG

This study examines how vertically aligned flapping surfaces, inspired by fish swimming formations, interact to influence propulsion performance. Through simulations and experiments, the researchers show that specific spacing and motion patterns can improve thrust or reduce energy use by altering vortex behavior, providing insights for designing more efficient bio-inspired propulsion systems.

PUBLICATION: Hydrodynamic interactions of low-aspect-ratio oscillating panels in a tip-to-tip formation. 

Pan, Yu; Zhu, Yuanhang; Westfall, Elizabeth; Quinn, Daniel B.; Dong, Haibo; Lauder, George V. Journal of Fluid Mechanics, 2026 

MADHUR BEHL

This work presents the HALO safety architecture, a framework designed to improve the reliability of high-speed autonomous racing vehicles by identifying and mitigating software and system failures. Validated through real-world autonomous racing trials, the approach uses fault analysis and runtime monitoring to enhance safety across perception, planning, control, and communication systems.

PUBLICATION: HALO: Fault-Tolerant Safety Architecture For High-Speed Autonomous Racing 

Harder, Aron; Kulkarni, Amar; Behl, Madhur. ACM Transactions on Cyber-Physical Systems, 2026  

YEN-LING KUO

This work explores the challenges of developing effective human-AI collaboration in physical environments, where agents must adapt to complex behaviors, dynamic interactions, and shared goals. By introducing the Moving Out collaboration environment, the study highlights limitations in current AI systems’ ability to coordinate, provide assistance, and maintain consistent actions, while identifying opportunities to improve physical reasoning and adaptive collaboration.

PUBLICATION: Towards Physically-grounded Human-AI Collaboration

Kang, Xuhui; Kuo, Yen-Ling, Proceedings of the AAAI Symposium Series, 2025

Robotics and Autonomous Systems News

  • UVA Electrical Engineer Nikos Sidiropoulos Wins IEEE Fourier Award, Among Signal Processing’s Highest Honors

    Electrical engineering professor Nikos Sidiropoulos was recognized for his work on multidimensional data — an area spanning streaming recommendations, cellphone connectivity and medical imaging.

    Read Now

  • Four UVA Engineering Faculty Earn NSF CAREER Awards

    Four early-career faculty at the University of Virginia School of Engineering and Applied Science have earned prestigious National Science Foundation CAREER Awards.

    Read Now

  • Chen’s NSF CAREER Award Supports UVA Engineering Artificial Intelligence Research to Better Forecast Infectious Disease Epidemics

    A UVA Engineering researcher is using AI to develop computational tools that improve infectious disease forecasting, guide public health decisions and strengthen disease surveillance.

    Read Now