Abstract
Large-scale crowd management systems are used to monitor and manage crowds in various industries aspects by utilizing relevant innovative technologies. In order to overcome the shortcomings of traditional CCTV equipment in shooting angle and deployment, some scholars propose to use unmanned ariel vehicle (UAV) carried appropriate optical sensory equipment to perform aerial scene surveillance. However, UAV flight missions have problems such as poor adaptability of single-mode path planning to site conditions space and complex cluster scheduling systems. Therefore, we combine the improved particle swarm optimization(PSO) algorithm, the optimized artificial potential algorithm, path exploration switching mode and energy-based task scheduling mechanism to propose a joint global and local path planning optimization for UAV task scheduling towards crowd air monitoring (JGLPP-UTS). In this model, the PSO algorithm is improved based on mutation mechanism and iterative number dependent adaptive inertia weight, we add a path smoothing mechanism. Then, we optimize the artificial potential algorithm for the problem of the target point unreachable and the local minimum. The proposed model switches the path planning mode according to the global and local obstacle environment. Finally, our model comprehensively considers the information of the site to realize the surveillance task scheduling of the UAV. Experiments show that our proposed algorithm can effectively improve the ability of global and local path planning. Compared with the standard PSO path length, the global path is reduced by 8.92%, and the adaptive value is reduced by 82.9%. After the smoothing operation, we also report that the path length can be further reduced. Moreover, the task scheduling strategy can realize the effective use of airborne resources.