The time-frequency poles of a random system

Nonstationary random processes arise naturally from physical phenomena and can be modelled by using random systems. An effective way to characterize a random system is to transform it to the time-frequency domain. This transformation generates a time-frequency system, whose solution is the time-frequency spectrum of the output of the random system in time. We study the nature of the poles of the time-frequency system. The obtained results show that knowledge of the time-frequency poles allows a deeper understanding of the random system in time, and, in general, of a nonstationary random process.