Optimal processing of low power signal in the system of Internet of Things

Among different technologies under investigation for Internet of Things (IoT) one of the most promising is Nar-rowband IoT (NB-IoT). Key feature of this technology is ultra narrow band of signal spectrum. This leads to high influence of low frequency interferences, e.g. flicker noise. This interference has 1/f power spectrum, which allows to use fractal model to describe properties of this random process. The article deals with statistical methods for development of fractal models, e.g. the fractal Brownian motion model with fractional dimension or Hurst exponent. An optimal algorithm for detecting signals on the background of sum of fractal flicker noise and thermal noise is developed and its efficiency is investigated in temporal and spectral field as well. It is proved, that increasing of signal duration leads to spectrum transfer to low frequency area with high intensity of flicker noise spectrum. It leads to decreasing noise resistance.