Spectral analysis of level crossing sampling scheme
Résumé
This paper takes place in the context of developing a new class of signal processing chains integrated on a single chip (SoC or System on Chip). The system must be low-cost, reduced-sized, low noise and especially low-power to be powered by batteries or remotely powered. Thus a new kind of sampling scheme called level-crossing sampling is considered achieving non-uniformly spaced out in time sampled points. This sampling scheme is analyzed for periodic signal proving that the sampled signal spectrum is aliased. By weighting the spectrum, the paper shows how to improve the accuracy of the spectral analysis and prove how a Continuous Time signal can be reconstructed under conditions in spite of aliasing.