Fast Algorithms for Deterministic Non-Equidistant Digital Filtering of Signals in the Time Domain

The article is devoted fast algorithms for deterministic non-equidistant digital filtering of signals in the time domain. The algorithms are based on codes with minimal aperiodicity. Deterministic non-equidistant sampling of signals in the time domain allowed merging the positive properties of both periodic and stochastic sampling processes. The algorithms perform address recirculation to computing of convolution operator and allow the comb characteristics of the filters to be constructed using the thinning procedure for partial products. The algorithms allow obtaining convolution values with significantly smaller hardware and time costs. In addition, the algorithms allow to easily parallelizing the structure of the computing device. The presented algorithms are convenient for hardware calculation of the spectrum because there are required one multiplier block for many channels and one control device, which reduces hardware costs. The algorithms can be used in the task of real time authentication of user electronic devices and cyber-physical devices, including intelligent sensors, based on the individual characteristics of the sensor electrical noise signal.