Identification of modal parameters, when a structure is under operational conditions is termed Operational Modal Analysis (OMA). Current OMA techniques are based on the assumption of linear time-invariant systems, and thus have limited applicability when applied to structures known to violate these assumptions. The present study investigates how the Random Decrement (RD) technique can improve robustness of OMA methods when friction-induced nonlinear damping is present in a system. This is done by estimating the amplitude dependent damping. A friction mechanism is introduced in a model of a structure, and by applying the RD technique at different amplitudes of simulated responses, RD signatures are produced, that represent the system vibrating with these amplitude levels. This allows the modal parameters to be estimated based on RD signatures computed with each amplitude level, using time domain parameter estimation methods, and the amplitude dependency of the damping is identified.