Uncertainty management in space station autonomous research: pattern recognition perspective

Abstract

Various space station autonomous operations where research is being conducted to support unmanned missions are discussed. The problem of representation and management of uncertainties involved in image analysis and recognition tasks of those operations is particularly addressed. Various tools based on fuzzy set theory and probability theory that reflect different kinds of ambiguity, uncertainty, and information in an image are listed. Their usefulness in providing soft decision and quantitative indices for autonomous operations is described along with the uncertainty in membership function evaluation. The merits of incorporating fuzzy set theory in neural networks and in genetic algorithms for efficient handling of uncertainties are also addressed.