Date: Wednesday, November 19, 2025

Time: 11:00AM - 12:30PM ET

Location: Love Building, Room 295

Virtual:  https://gatech.zoom.us/j/5675939668

Rui de Gouvea Pinto

Robotics Ph.D. Student

George W. Woodruff School of Mechanical Engineering

Georgia Institute of Technology

Committee:

Dr. Jonathan Rogers (Advisor)

Daniel Guggenheim School of Aerospace Engineering

Georgia Institute of Technology

Dr. Sarah H. Q. Li

Daniel Guggenheim School of Aerospace Engineering

Georgia Institute of Technology

Dr. Anirban Mazumdar

George W. Woodruff School of Mechanical Engineering

Georgia Institute of Technology 

Dr. Matthew Hale

School of Electrical and Computer Engineering

Georgia Institute of Technology

Dr. Sean Wilson

Robotics and Autonomous Systems Division

Georgia Tech Research Institute

Abstract:

This research develops a framework for autonomous multi-agent coordination to support wildfire containment under realistic operational constraints. The objective is to enable teams of unmanned aerial vehicles (UAVs) to perform backburning and monitoring tasks adaptively in response to evolving fire behavior, while accounting for resource and communication constraints. The work integrates three key components: a feedback-oriented wildfire simulation environment (EMBRS), a decentralized task-allocation algorithm that accounts for battery and payload limitations (RA-CBBA), and a connectivity-aware extension that preserves network integrity during distributed operations.

EMBRS provides the closed-loop environment needed to model fire spread, simulate firefighting actions, and generate fire behavior forecasts. RA-CBBA enables UAVs to plan tasks while incorporating required replenishment events, achieving feasible assignments in large, dynamic environments. A connectivity-maintenance module assigns mobile relay agents to preserve network connectivity during execution. These components are integrated into a global firefighting planner that synthesizes ignition tasks, monitoring tasks, and real-time predictions into an adaptive multi-agent backburning strategy.

The proposed framework advances autonomous wildfire research by embedding realistic fire dynamics, operational constraints, and communication challenges into coordinated multi-agent planning. This work aims to provide a foundation for safe, reliable, and scalable autonomous backburning operations in future wildfire management systems.