Classification and Engineering Analysis of Solid Propellant Missile Telemetry Data

What is it about?

This work utilizes the first 30 seconds of missile telemetry data to rapidly classify missiles. Neural Networks are used to classify the missiles with high performance and the Shapley Additive Explanation (SHAP) method are used to derive the most important features of the model. Using SHAP we can statistically show the most important features are the earliest data points in the telemetry and are most useful for classification.

Featured Image