Aerodynamically Controlled Missile Flight Datasets and Its Applications

Abstract

This paper provides open flight datasets generated by extensive Monte Carlo simulations for an aerodynamically controlled missile under a fixed engagement scenario with the purpose of encouraging data analytics research in the missile application. With the fast advance in the field of data analytics, fueled by the recently developed machine learning and deep learning algorithms, the potential of applying the data analysis techniques to the domain of aerospace engineering is increasingly high, especially in the light of detecting anomalies and uncovering the hidden information in the flight data. A dataset is an essential ingredient in the research of data analytics and the development of new data analysis frameworks. However, since it is almost impossible or highly inefficient to construct a large dataset of guided missiles via physical experiments, a disclosed database is nearly non-existence. Even if this is possible, there is no open source dataset for academic research due to the security issue. Thus, by developing a high-fidelity 6-DOF (degree-of-freedom) simulation program, we generate realistic flight datasets of a guided missile, which can be used publicly. Furthermore, we provide illustrative examples of using the generated datasets for the purpose of demonstrating the potential of the application, that is, to detect abnormal data patterns to determine signs of the control loop instability during the flight.