Procedimento para detecção de falseamento via amostragem não uniforme

Abstract

The sampling theorem determines a lower bound for the sampling frequency necessary for the perfect representation of a continuous signal by a countable set of uniformly spaced time samples. If such a bound is not satisfied the resulting spectrum is said to be aliased. Nonuniform sampling is not bound by the sampling theorem, although it is influenced by some practical limitations. This work investigates some of these limitations and proposes a procedure to check if a signal with unknown spectrum is aliased after uniform sampling. Such procedure is evaluated by means of simulation and is discussed in this work in the context of an implementation in digital signal processors with memory restrictions. Results show that alias detection is possible, with a high rate of success, if sample size and signal to noise ratio requirements are met