"Simulated abnormal breathing signals "

"The datasets include simulated respiratory patterns segmented into 15-second windows, with a simulated sampling frequency of 17 Hz, as would be obtained from a radar sensor. Specifically, the dataset includes eupnea, Kussmaul, and Cheyne\u2013Stokes patterns.These patterns were generated following the approach proposed by Paterniani et al. (doi: 10.1109\/JPROC.2023.3244362), in which each breathing cycle is divided into inspiration and expiration phases, modeled by two separate equations. By varying the breaths-per-minute (bpm) value and the absolute maximum chest displacement associated with breathing, the different characteristics of abnormal respiratory patterns can be represented.Specifically, for eupnea (i.e., normal respiration), a random value between 12 and 20 bpm was selected, while the maximum displacement was set to 1.1 cm.For Kussmaul respiration, the bpm range was set between 28 and 40 bpm, with a maximum displacement of 2.2 cm.Finally, Cheyne\u2013Stokes respiratory patterns were simulated using a bpm range from 15 to 23, which is therefore closer to the eupnea range, with a maximum displacement of 1.1 cm. Moreover, to reproduce the characteristic Cheyne-Stokes pattern, a sine function with a period equal to twice the ventilation time was multiplied by the generated breathing patterns, while the apneic intervals separating the crescendo and decrescendo phases were modeled by multiplying the signal by zero.In all three cases, Gaussian noise and a baseline shift were added to simulate the distortion introduced by the radar during data collection."