Method for Predicting Impact Point of Trajectory Correction Projectile Based on IBKA-BP-AdaBoost

To improve the accuracy of external ballistic trajectory prediction to meet the needs of modern battlefields, a framework for optimizing the BP neural network combined with AdaBoost ensemble learning algorithm using the improved Black Kite algorithm is proposed. First, the four-order Runge-Kutta method is used to solve the six-degree-of-freedom external ballistic equations of rigid bodies, and the flight characteristics of the projectile are analyzed to establish a dataset. In order to simulate the complex environment of the real battlefield, Gaussian noise and uniform noise with different amplitudes are added to the dataset. Then, the projectile impact point information is used as the output index, the flight information of the projectile as the input index, and after Unscented Kalman Filter, the improved model is used for projectile impact point prediction. At the same time, it is compared with the BKA-BP-AdaBoost, WOA-BP-AdaBoost, RIME-BP-AdaBoost, and GWO-BP-AdaBoost models. The simulation results show that the proposed algorithm has higher accuracy in impact point prediction than the other four algorithms, and shows better robustness and reliability in complex noise datasets, which can provide support for the research on projectile impact point prediction and precision firepower strike.