Prediction model of UAV damage efficiency based on GRU neural network

In the potential international conflicts of the 21st century, the construction of anti-drone systems is urgently needed. Possessing a reliable anti-drone system<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</sup> is of significant guidance for developing reliable drone penetration algorithms and for launching concentrated strikes against hostile drones. Starting from the existing state information of drones and their damage states, this paper proposes a drone damage effect prediction model based on GRU<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[2]</sup>, including the design of the network structure, network training, and the implementation of prediction process algorithms. Considering the volatility of output data in drone damage effect prediction, the loss function is further refined into an Elastic Net regression<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[3]</sup> model based on the traditional GRU neural network to enhance the model’s computational capabilities and reduce overfitting. The GRU neural network proposed in this paper has been verified to have strong applicability and higher accuracy in predicting drone damage effects.