Unmanned Aerial-Ground Systems for Wildfire Detection, Global Localization and Suppression

This paper presents a new and comprehensive autonomous framework for wildfire detection, global localization, and suppression leveraging Unmanned Aerial Vehicles (UAVs) and Unmanned Ground Vehicles (UGVs). This work significantly expands upon existing research focusing solely on UAV-based wildfire mitigation. By synergistically combining the capabilities of UAVs with those of UGVs, our system enhances situational awareness and operational effectiveness. Notably, this approach integrates cutting-edge localization technologies, including the Simultaneous Localization and Mapping (SLAM) system for accurate global mapping of fire spots to enable near-meter-level precision tracking of wildfires even when obscured from view by vegetation or debris. To further enhance accuracy, we employ a motion model and an Iterated Extended Kalman Filter (iEKF) incorporating both Global Navigation Satellite Systems (GNSS) data and semantic vision information. Extensive simulations in cluttered environments confirm the efficacy of our proposed framework in supporting timely and accurate wildfire detection and suppression.