Control and diagnostic systems currently used in rolling stock diagnose malfunctions in wheel and-motor units, bearing defects, wheel-and-motor units, bearing misalignments, mounting defects, imbalance of rotating parts, gearbox defects. They also diagnose leaks in the feed and
brake lines, break valve malfunctions, brake cylinder malfunctions, compressor malfunctions. In this case, parameters of vibration, pressure, force, current, voltage, resistance, pulses, time intervals are used as diagnostic indicators. The control results are primarily reflected on the
driver's monitor. However, these systems do not detect the beginning of the latent period of the onset of a malfunction, which usually precedes an accident. In addition, these systems provide no information on the dynamics of its development. Therefore, the result of control is sometimes belated and does not allow taking appropriate actions in a timely manner. To eliminate this drawback, we propose a technology that makes it possible to register the beginning of a latent period of malfunctions using the estimate of the distribution law of the noise of noisy signals.