Citation
Abstract
This article discusses the advantages and caveats of a statistical link analysis method established by J. Yuen in the 1970s. Since then, the Jet Propulsion Laboratory has adopted this approach as flight principles in order to conduct link analysis for its deep-space missions. This article presents three new results: 1) we show that by invoking Lyapunov’s condition of the central limit theorem, the Gaussian approximation of statistical link analysis is indeed mathematically valid if there is no single link parameter with a variance that is much larger than the others; 2) we discuss a system engineering approach of incorporating expert opinions as a mean to mitigate the lack of statistical knowledge of the link parameters; and 3) we also introduce the use of saddle-point approximation to estimate the tail probability in statistical link analysis in situations when the Gaussian approximation does not apply.
Keywords
Details
- Volume
- 42-183
- Published
- November 15, 2010
- Pages
- 1–16
- File Size
- 318.1 KB