Projectile Resistance Coefficient Identification Based on Extreme Learning Machine

Abstract

Aerodynamic parameters play a decisive role in the ballistic characteristics of the projectile. Accurate acquisition of projectile aerodynamic parameters is the key to reducing the spread of drop points and improving strike accuracy in the development process of uncontrolled projectiles. In order to further improve the identification accuracy of the projectile resistance coefficient, based on the mass point ballistic equation, the ballistic data generated through numerical simulation, and the extreme learning machine was used to identify the ballistic resistance coefficient under three kinds of noise conditions. By randomly generating input weights and hidden layer neuron thresholds, the randomly generated input weights and hidden layer neuron thresholds are independent of each other and do not require iterative update, the method reduces the identification time and improves identification accuracy of traditional identification methods. Based on the least square principle, the Moor-Penrose generalized inverse matrix of the hidden layer output matrix was solved to determine the optimal output weight of the network, and then the projectile resistance coefficient was accurately identified. Comparing the results of extreme learning machine with the traditional BP neural network method and Maximum likelihood method, the results show that the proposed method has higher identification accuracy, faster convergence speed and stronger anti-interference ability. It can effectively identify the projectile resistance coefficient under the high-noise environment, which can meet the practical needs of engineering.