Neural network based wireless charging output control strategy for UAVs

Most of today's UAV charging is labour-intensive and in complex environments, it is not possible to ensure that the UAV receives energy safely and stably. So this paper proposes a UAV hovering wireless charging system to enhance the flight endurance in complex environments. This paper adopts the compensation topology that the output current is unrelated to load impedance, and reveals the change rule of the output current of the two compensation networks under mutual inductance and parameter perturbation and the effect of load perturbation on the output current in the actual system. Because the influence of the fluctuation of mutual inductance between the transmitter and receiver side on the transmission performance of the system in the wireless energy transmission system is taken into account, the article proposes an artificial neural network to control the output current to ensure the improvement of the effectiveness and feasibility of the system in the dynamic change of mutual inductance. After that, simulation experiments are carried out to verify the scheme based on the SS compensation topology, and the average error of the scheme is within 3%, and the interference factors are also analysed and considered, and the error is within 5%.