Fixed Cameras Assisted UAVs Monitoring: Enhancing Monitoring Efficiency and Safety

Haihan Zhang, Haipeng Dai

Abstract: Traditional monitoring tasks involve monitoring objects using fixed cameras with limited field of view, making it challenging to deploy in certain scenarios and achieve optimal monitoring positions. Due to the flexible mobility and wide field of view of unmanned aerial vehicles (UAVs), they have been widely employed in various monitoring tasks in many domains. However, if UAVs are too close to monitoring objects, it may pose safety risks. Consequently, regulations have been established by various countries specifying that UAVs should fly at a certain distance from objects. This makes it difficult to monitor certain objects using UAVs. Fortunately, nearby fixed cameras can be rented for monitoring these objects. However, given a limited number of UAVs and finite budget for renting cameras, determining the monitoring strategies for the UAVs and cameras to maximize monitoring utility becomes a challenging problem. To address this challenge, we first discretize the monitoring space and propose a strategy extraction algorithm to extract a finite number of representative candidate strategies from an infinite number of candidate strategies in continuous space. Then, we formulate the problem as a submodular maximization problem with two knapsack constraints and one matroid constraint. To the best of our knowledge, under these constraints, the known approximation algorithm has an approximation ratio of . We further propose an approximation algorithm to obtain the final monitoring strategies and prove that the solution provided by this algorithm outperforms the known algorithm. Moreover, in a special case of the problem, we demonstrate that the proposed algorithm has a better approximation ratio than the known algorithm. Finally, we conduct extensive experiments to validate the effectiveness of the proposed algorithms.