UCAV maneuvering trajectory prediction based on PSO-CNN

Abstract

To the problem of low accuracy of unmanned combat aircraft maneuver trajectory, a particle swarm optimization convolutional neural network prediction method is proposed. Firstly, establish a three-degree-of-freedom model of Unmanned Combat Aerial Vehicles (UCAV) with constraints to solve the problem of trajectory source. The structure of the convolutional neural network is analyzed, and the particle swarm optimization algorithm (PSO) is used to replace the backpropagation algorithm to update the internal weights and biases. The PSO is compared with multiple algorithms, and the results show that the PSO updates the weights fast and has small errors. Finally, the prediction is made on a relatively complex and cluttered maneuvering trajectory. The method proposed in this paper is compared with three traditional prediction methods, and the result shows that the method proposed in this paper has small prediction errors.