Unified approach to measurement signal reconstruction

Inversion of the mathematical model is intrinsic to any measurement process, and the numerical complexity of this operation grows with the functional flexibility, accuracy, and speed of measurement. A special case of model inversion, namely, measurement signal reconstruction, is studied, using a linear model of the relationship between two scalar signals. Being numerically ill-conditioned, the reconstruction requires regularization, and reconstruction are classified into six groups according to the mechanism of regularization on which they are based. The six groups are: direct, variational, probabilistic, iterative, parametric, and transform methods