On the basis of statistical analysis of acoustic noise realizations recorded in rail (railway and subway) transport the noise model was elaborated as normal ergodic centred locally stationary random process. The numerical estimates of local stationary intervals were obtained. The simulation model for adaptive noise canceller was proposed which is consistent with the noise model. The noise cancellation simulation was held resulted in the estimates of the noise cancellation performance. It is found that the system developed cancels the noise effectively and can be used for personal protection of passengers and staff in railway and subway trains.
rail transport noise, correlation interval, locally stationary random process, adaptive filtering, simulation