Abstract:

Verification, validation, and accreditation (VV&A) is a key means to ensure the credibility of simulation models, and model validation is the core link. In view of the unavailability of reference data, various sources of reference data, and strong subjectivity of expert validation in the result validation of the missile flight simulation model, a result validation method for the missile flight simulation model based on piecewise feature extraction of time series was proposed. Specifically, a comprehensive piecewise linear method for time series was first proposed. The method consisted of a linear piecewise algorithm based on the second-order derivative for extracting trend edge points and a top-down linear piecewise algorithm based on extreme point optimization, so as to achieve an efficient linear piecewise representation of the missile flight simulation data and reference data. Then, based on the segmentation results obtained above, features such as mean, variance, and slope of each time series could be extracted to assist experts in reducing subjectivity in validation. The Theil inequality coefficient (TIC) method, dynamic time warping (DTW), and other methods could also be used for objective consistency analysis. By taking advantage of piecewise features of time series, the results of the missile flight simulation model in various scenarios can be validated. Finally, a result validation case of the missile model demonstrates the feasibility and effectiveness of the proposed method.

Key words: result validation of missile flight simulation model, reference data, objective validation, piecewise feature extraction, time series