Fast identification of continuous-time Lur’e-type systems with stability certification

In this paper, we propose an approach for parametric system identification for a class of continuous-time Lur’e-type systems using only steady-state input and output data. Employing a quasi-Newton optimization scheme, we minimize an output error criterion constrained to the set of convergent models, which enforces a stability certificate on the identified model. To compute the steady-state model response efficiently, we adopt the Mixed-Time-Frequency (MTF) algorithm. Furthermore, using the MTF algorithm, we present a method to efficiently compute the gradient of the objective function with any user-defined accuracy. Starting with an initial convergent model estimate, the developed identification algorithm optimizes parameter estimates. The effectiveness of the proposed approach is illustrated in a simulation example.