Design and Optimization of Drone Assisted Wildfire Fighting System
Year: 2025
Type: article
Abstract: The utilization of autonomous agents such as multiple UAVs has been continually increasing to recognize spot fires and screen out of control fire hazards moving toward a structure, fence, forest, or firefighting crew via remote sensing. Wildfires have caused catastrophic losses as their economic and regional impact were not reduced. The monitoring and suppression of wildfire with drones using the Voronoi Partitioning Algorithm is proposed in this report. Robot Operating System was used to deploy quadcopters in an area using Centroidal Voronoi Tessellation. By the algorithm, a plane with n foci is divided into convex polygons such that every polygon contains precisely one generating point. In Voronoi Partitioning, UAVs are situated in the centroid of the partitioned area hence the area covered by each drone is almost equal, and an unbiased search is followed through. The MATLAB simulation of the Voronoi algorithm was run with ‘n’ number of drones to see the configuration firsthand and see how every drone occupied an equal area density to carry out the specific application, making the detection process smoother and more efficient. Hector Quadrotor was simulated in Gazebo environment and related packages were configured to emulate it. Rviz was used to check the function of the cameras for fire detection and was run alongside the Hector Quadrotor. A strategy that could be used for forest firefighting by using multi drone systems is elaborated in this report. A literary review was done to discuss the various available path planning techniques and drone systems to detect fires. Using Voronoi-Tessellation in MATLAB, the path for the robots’ search was developed. Separately, the drone used was simulated in a virtual environment called Gazebo. By using combinations of different drones and thermal cameras in the simulation, multiple alternatives have been recognized. Further, an addition of a thermal attribute to the environment to simulate a real-world scenario and systemize the communications between various instances of the drones were made to detect wildfire affected areas accurately.
Institutions: Manipal Academy of Higher Education, Northeastern University
Language: English
Subfield: Safety, Risk, Reliability and Quality
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7160888185
