Summers Lab · Multi-Robot Systems · UTD

Autonomous
Fleet Systems

A research project and student team building a world-class testbed for swarm robotics and intelligence — where collective behavior emerges from local interactions, and where engineering and mathematics are equally at home.

Two questions nobody has fully answered

How does useful collective behavior emerge from simple local interactions among autonomous agents? And how do we deliberately design communication architectures and control structures that make reliable, scalable coordination possible?

These bottom-up and top-down perspectives are mirrors of each other. The most interesting science lives at their intersection — and this lab is built to probe it experimentally, with over 100 real robots.

Bottom-up: Local rules, emergent global behavior, distributed intelligence.

Top-down: Designed architectures, formal guarantees, principled coordination.

Lab photo — coming soon
100+
Robot platforms
24
OptiTrack cameras
1,300
sq ft high bay
6
Research tracks

Theory and practice treated as equals

01

Feedback Control & Stability

Designing controllers with formal guarantees — stability, robustness, and performance — under real-world disturbances and model uncertainty.

02

Motion Planning

From graph search to trajectory optimization: finding paths that are safe, dynamically feasible, and optimal with respect to a task objective.

03

State Estimation & Perception

Fusing noisy sensor measurements into accurate, uncertainty-aware state estimates — the foundation every other stack layer depends on.

04

Multi-Agent Coordination

Graph-theoretic and distributed optimization tools for designing protocols that produce coherent collective behavior at fleet scale.

05

Learning & Adaptation

Integrating ML and classical engineering with honest attention to interpretability, safety guarantees, and when learned components help or hurt.

06

Cross-Disciplinary Perspectives

Biology, music, choreography, and social science each offer insight into collective behavior that pure engineering cannot generate on its own.

Real engineering problems, real stakes

Coordinated multi-robot systems are moving from research labs into the world at speed. The theory and infrastructure built here connects directly to consequential applications.

🌬️

Wind Energy Infrastructure

Autonomous inspection of turbines and transmission lines — including the large-scale wind infrastructure right here in Texas.

🚗

Autonomous Vehicle Fleets

Coordinated multi-vehicle navigation without central controllers — intersection negotiation, lane merging, shared road space.

🌊

Environmental Monitoring

Distributed sensor networks for ocean chemistry, atmospheric composition, wildfire behavior, and climate variables at scale.

🏥

Service Robotics

Coordinated robot fleets in hospitals, warehouses, and buildings — delivery, navigation, and human-safe autonomous operation.

🛡️

Reconnaissance & Surveillance

Coordinated UAV teams for search, mapping, and logistics in contested environments where decisions must be made at machine speed.

🌾

Precision Agriculture

Ground and aerial robot coordination for crop monitoring, irrigation management, and disease response at field scale.

A living, performing laboratory

Alongside the research infrastructure work, the team pursues a motivating long-horizon goal: a lab that never sleeps. Robots that perform, recharge, and perform again — autonomously, indefinitely.

Formation flight demo — coming soon
  • Sem 1 · Wk 3–5

    Milestone 0 — First Light

    Three or more Crazyflies in the air simultaneously with light decks active. The moment the lab becomes real for the first cohort.

  • Sem 1 · Wk 8–12

    Milestone 1 — Static Formation Show

    8–12 Crazyflies holding a geometric formation under OptiTrack feedback with synchronized light sequences.

  • Sem 2 · Wk 1–10

    Milestone 2 — Dynamic Choreography

    15+ robots in a 2–3 minute choreographed show: formation transitions, spelling in mid-air, synchronized light cues.

  • Sem 2 · Wk 10–16

    Milestone 3 — Heterogeneous Show

    Crazyflies and Turtlebots performing together — aerial and ground robots coordinated in a single unified show.

  • Semester 3+

    Milestone 4 — The Lab Screensaver

    Generative, continuously varying performance running autonomously whenever the lab is idle. Robots cycle through charging and flight indefinitely.

Faculty leads

The project is led by three faculty members from the Erik Jonsson School of Engineering and Computer Science at UT Dallas.

Photo coming soon

Prof. Tyler Summers

Principal Investigator

Associate Professor, Mechanical Engineering. Research in control, optimization, and learning in complex dynamical systems.

Faculty page ↗

Photo coming soon

Prof. Waseem Abbas

Co-Director

Assistant Professor, Systems Engineering, Research in network controls, graph machine learning, and multi-agent systems

Faculty page ↗

Photo coming soon

Prof. Justin Koeln

Co-Director

Associate Professor, Mechanical Engineering. Research in controls, model predictive control, and energy systems.

Faculty page ↗


Graduate Coaches & Students

Graduate student coaches and project members will be listed here as the first cohort is formed.

Join the first cohort

Imagine a fleet of drones coordinating a disaster response — each one making local decisions that add up to something intelligent and purposeful. In Autonomous Fleet Systems, you will tackle these problems as part of a close-knit, intergenerational team where freshmen, seniors, and PhD students work side by side.

No prior robotics experience is required — just curiosity and a willingness to contribute. Each semester builds on the last, and the relationships you build along the way are as much a part of the experience as the engineering itself.

Track 0

Theory & Foundations

Math-first. For students who want to understand the proofs.

Track 1

Platforms & Sensing

Hardware, sensors, OptiTrack pipeline.

Track 2

Control & Planning

Single-agent autonomy stack, trajectory generation.

Track 3

Multi-Agent

Coordination protocols, distributed algorithms.

Track 4

AI & Learning

Learned components, system identification, safety.

Track 5

Visualization & Data

Unity integration, experiment design, analysis.

Track X

Cross-Disciplinary

Biology, music, choreography, social science — all disciplines welcome.

Enrollment

Students enroll for course credit through the ECS RIDE program. Effort scales with credit hours — about 3 hours of work per credit hour per week.

1 credit · ~3 hrs/wk 2 credits · ~6 hrs/wk 3 credits · ~9 hrs/wk

Start at 1–2 credits and increase as you become more deeply involved. Multi-semester participation is strongly encouraged.

Who should apply

Students at any level — freshman through senior — from any major. Engineering, CS, math, physics, biology, music, art, and social science students have all found meaningful roles in projects like this one.

The project is designed to be accessible at multiple levels of preparation. You do not need to know ROS or control theory before you start.

Interested? Fill out the interest form or reach out directly.

Express Interest