Evolving Autonomous Navigation: A NEAT Approach for Firefighting Rover Operations in Dynamic Environments

This paper introduces a novel approach for automating rover navigation in dynamic environments, explicitly focusing on firefighting applications, using the NeuroEvolution of Augmenting Topologies (NEAT) [1]. The research leverages NEAT's capability to evolve intricate neural networks, aiming to create a sophisticated rover navigation system. This system is designed to adeptly maneuver through complex terrains, avoid obstacles, and locate humans in highrisk fire scenarios. The motivation for this research is rooted in the urgent need for advanced autonomous technologies in lifesaving operations, particularly firefighting, where a rapid, precise response is crucial. The system's performance is rigorously assessed through comprehensive simulation experiments, showcasing its ability to achieve superior navigation efficiency and convergence. Additionally, the research provides an in-depth analysis of the rover's evolved behaviors using various visualization tools such as charts, statistics, and heatmaps. The approach we have used, the NEAT-based solution, demonstrates promising potential, offering a more adaptable and practical approach for real-world firefighting and rescue missions.