Expert Knowledge-Assisted Drone Combat Strategy via Intention Interpretation

In this paper, a drone game strategy generation model is designed in a denied environment. A typical nonlinear model of drone formation dynamics is first constructed. A behavior prediction algorithm based on neural network is then utilized to predict the trajectory and behavior characteristics of the attack drones, and a neural network algorithm achieves intelligent interpretation of the attack drones intention based on the result of the prediction. A game strategy generation algorithm assisted by expert knowledge and game experience library is finally proposed using the result of the intention interpretation. These algorithms overcome the difficulty of obtaining the game strategy under the deceptive and elusive characteristics of attacking drones and realize high real-time and pertinence of decision-making. Simulation results of typical scenarios are provided to verify the effectiveness of the proposed algorithm model.