High Speed Target Trajectory Prediction Based on Machine Learning

The air combat posture is constantly changing, and accurately predicting the movement trajectory of targets in air combat is of great significance for high-precision tracking and attacking targets. Convolutional neural network can effectively obtain air combat data. This article proposes the use of CNN models for predicting air combat trajectory data, utilizing one-dimensional convolution to effectively extract temporal features of air combat target trajectory data. Through convolution operations, short-term features of the different three axis are extracted, and feature preservation and reinput are performed in the pooling layer to obtain prediction results. Then, the effectiveness of the proposed model is verified through comparison and error analysis with common models. Provide reference for the neural network application in air combat precision strike.