Aerodynamic Parameter Identification of Projectile Based on Improved Extreme Learning Machine and Ensemble Learning Theory

The firing accuracy of the projectile has a positive relation with aerodynamic parameters. Due to complex dynamic characteristics projectiles, there is an overfitting risk when single extreme learning machine (ELM) used identify parameters projectile, and identification results oscillate transonic region. To obtain accurately, parameter model based on ensemble theory ELM optimized by improved particle swarm optimization proposed. algorithm (IPSO) adaptive update strategy optimize weight threshold ELM. Combined theory, neural network regarded as weak learner generate strong learner. structural were continuously through training, obtained. simulation show that introduction IPSO enables exhibit excellent generalization ability. proposed can accurately describe variation Mach number.