Error minimization and detection limit improvement in crossing-based sampling and spectrum analysis

Abstract

The occurrence of stochastic resonance in sinusoid crossing-based signal and sampling spectrum analysis, is studied under two extreme cases: (I) large input signal s(t) where not all |s(t)| values are less than the amplitude A of the reference sinusoid r(t) = Acos(2πWt), and (II) s(t) is a weak undetectable sinusoid. A sinusoid crossing occur whenever s(t) - r(t) = 0. In case I, no crossing exists between s(t) and r(t) in halfperiod intervals of r(t) whenever |s(t)| > A. At these intervals, s(t) can not be known because it has no representation. In case II, sinusoid s(t) is weak such that the locations of its crossings with r(t), are identical with that of a straight line s(t) = 0.