Pattern Recognition for a Flight Dynamics Monte Carlo Simulation

Abstract

The design, analysis, and verification and validation of a spacecraft relies heavily on Monte Carlo simulations. Modern computational techniques are able to generate large amounts of Monte Carlo data but flight dynamics engineers lack the time and resources to analyze it all. The growing amounts of data combined with the diminished available time of engineers motivates the need to automate the analysis process. Pattern recognition algorithms are an innovative way of analyzing flight dynamics data efficiently. They can search large data sets for specific patterns and highlight critical variables so analysts can focus their analysis efforts. This work combines a few tractable pattern recognition algorithms with basic flight dynamics concepts to build a practical analysis tool for Monte Carlo simulations. Current results show that this tool can quickly and automatically identify individual design parameters, and most importantly, specific combinations of parameters that should be avoided in order to prevent specific system failures. The current version uses a kernel density estimation algorithm and a sequential feature selection algorithm combined with a k-nearest neighbor classifier to find and rank important design parameters. This provides an increased level of confidence in the analysis and saves a significant amount of time.

Tool for Rapid Analysis of Monte Carlo Simulations

“…The authors were able to show that the combined algorithms are useful in identifying influential variables, 3 but the MATLAB version of TRAM was relatively slow in processing large data sets. The current version has been programmed on a GPU and processing times have decreased greatly.…”

“…The authors strived to develop this tool from the perspective of an aerospace engineer that has a solid flight dynamics background but who is not necessarily an expert in the fields of statistics of pattern recognition. The constraints listed below were published previously, 3 but are listed here once again because they were paramount in the selection of the algorithms for TRAM. (b) Algorithms must make no assumptions about input probability density functions.…”

“…This eliminated the need for the neighbors to be sorted according to their distance from a particular graph point. In the previous version of TRAM, 3 only a fixed number of the very nearest samples were used in determining the class of the grid point.…”

“…An original variable comes directly from the given data set, and a compound variable is formed from the difference or the quotient of two original variables. 3 In each type of analysis the highest-ranked variables are those for which there is substantial separation between the passing and failing data points. These two types of analyses are performed separately as determined by the user, but the ranked results can be combined.…”

“…2, 3 The current version of this tool is now named TRAM, which stands for Tool for Rapid Analysis of Monte Carlo simulations. There are two main features that make this tool applicable to almost any Monte Carlo data set.…”