Power Transmission from Friis Equation and Machine Learning

The well-known Friis equation has been derived directly from a proposal of perceptron in an artificial neural network. Thus, while input is given by the transmitted power, gains are seen as weights. Indeed, some probabilistic ingredients are considered. Although not any variation to known formulation, some parameters have been added. Thus for example, noise modeled by normalized number would take part of perceptron formulation. Finally as intrinsic agent modeled by a Gaussian pulse would constitute a necessary element that gives account the electromagnetism part of Friis equation. Simulations about numerical values of perceptron output and variations of received power are presented. From results, one might to opt by the present Machine Learning formulation of Friis equation when high contamination of random variables would degrade the received power in real scenarios.