Combining Simulation With Flight Test Through Bayesian Inference

摘要

The number of test points performed in a flight test is often limited due to many factors, including cost and schedule. Having only a small number of open-air-range test points can make statistically relevant data analysis challenging. Using simulation results as prior information, Bayesian techniques can generate statistically significant results from only a few flight-test data points. This article presents fictitious simulation data from electronic-attack test points and representative flight-test data for test points run at the same conditions. The simulation and flight-test data are then combined using Bayesian techniques to generate a posterior probability density function. The posterior probability density function shows the effects of electronic attack and the associated credible intervals (somewhat similar to a frequentist confidence or prediction interval). The results demonstrate that Bayesian techniques can provide a way to obtain meaningful results, even with limited flight-test data.