Characterizing Non-Stationary Millimeter-Wave Communication Using Fuzzy Entropy

A fuzzy entropy has been proposed to characterize the non-stationary behavior of the continuous millimeter-wave channel. When the receiver and transmitter's relative speed is low, millimeter-wave channel varies smoothly, and the channel realizations are locally correlated. The correlation between the channels can be determined by measuring the spectral overlap between them. The correlation dimension is directly proportional to the transmitter and receiver mobility and the non-stationary event that the channel experiences. In this paper, first, we characterize the spectral overlap between the channels using Channel Matrix Distance, and the sequence of measurements is constructed accordingly. Then, Fuzzy Entropy (FE) is used to detect the irregularity in the spectral overlap sequence. The FE method assigns a continuous grade between zero and one to the input data, where one means there is no irregularity in the data and zero means the irregularity has been detected in the data. We show that the FE can discriminate the blockage event from the other events that cause the non-stationary behavior. We present the numerical results to evaluate different aspects of the FE method.