基于无人机群飞行轨迹优化的多辐射源追踪技术

In recent years, unmanned aerial vehicles (UAVs) have become popular objects in the military and civil fields due to their advantages of low cost and mobility. Safe and reliable trajectory control is very important in enabling UAVs to quickly, stably, and efficiently accomplish different tasks. First, this paper constructs a trajectory optimization model of the UAV swarm under multiple constraints and proposes the framework of estimation, matching, localization, and tracking. Next, a flight trajectory optimization algorithm of the UAV swarm based on deep reinforcement learning is designed. A deep neural network is specifically used to estimate the channel model, and the mapping relationship between the received signal strength and the distance is obtained. The received signal strength matrix is generated by the interactive method. Moreover, the corresponding distance matrix is calculated, and the matching results between the UAV and the radiation source are obtained. The multisphere intersection method is used to calculate the reference location of the radiation source by combining the mapping relationship between the received signal strength and the distance. Furthermore, the original optimization problem is transformed into the Markov decision process, and the radiation source location information is introduced into reinforcement learning to design an efficient UAV swarm flight trajectory optimization algorithm. In addition, group entropy is introduced to analyze and evaluate the intelligence of the proposed algorithm. Finally, the performance and the intelligence of the proposed algorithm and the UAV swarm trajectories are analyzed using the simulation results, and the effectiveness of the proposed method is verified.