Multi-UAVs for Bushfire Situational Awareness: a Comparison of Environment Traversal Algorithms
Reza Bairam Zadeh, Arkady Zaslavsky, Seng W. Loke, Somaiyeh MahmoudZadeh · 2021
Unmanned Aerial Vehicles (UAVs) are efficient (often Internet-connected) devices with unique capabilities, making them suitable for achieving continuous and real-time Situational Awareness (SA) over disaster areas such as bushfires. Nevertheless, UAVs are resource-constrained with limited processing power and flight time, preventing them from achieving full coverage of the affected area. This research aims at developing a computationally fast approach implementable on resource-constrained UAVs for mission planning and traversing the environment to monitor the fire-affected areas. Thus, this study investigates the performance of optimized Rapidly-exploring Random Tree (RRT*) and modified Genetic Algorithm (GA) for multiple UAVs, facilitating them to efficiently traverse over the affected area and capture information from the spots with over 50% probability of fire existence. The performance of these two algorithms is thoroughly investigated, and a comparative analysis proves the efficiency of GA in satisfying all mission objectives, including short computation time and efficient battery usage.