Interpolating Missing Data of Projectile Trajectory Using BP Neural Networks

摘要

Exterior ballistic tests of projectiles and rockets usually use radar tracking measurement system. Sometimes, there some missing data are in the acquisition data, which make it difficult to further analysis and process data. This article discussed the use of BP neural network as a way to reproduce the missing values including preprocessing data and selecting the number of neurons. A simulation of trajectory and missing data was conducted based on Mass Point Ballistic Model for a projectile. Comparing traditional polynomial and cubic spline interpolation results, the interpolation efficient of BP Neural Networks is much better for the treatment of missing values. This study provides a novel method to infill the missing data of trajectory for projectile in test.