Smart Projectile Parameter Estimation Using Meta-Optimization

Abstract

Parameter estimation and system identification of smart projectiles is an important and commonly used industrial tool. Existing methods rely on good initial estimates of the parameters to avoid local minima and ensure proper convergence. New projectile configurations may include highly nonlinear dynamics or unknown control parameters that cannot be known a priori. A new method for projectile parameter estimation is proposed that combines an output error parameter estimation algorithm with meta-optimization. Meta-optimization uses a suite of optimizers in an intelligent manner to reliably minimize a cost function. This new method is applied to the identification of a smart projectile system equipped with microspoilers using simulated spark range data. The method is able to reliably estimate the aerodynamic coefficients of the projectile body as well as the properties of the control mechanism based on a fit of multiple trajectories.