Subspace identification for continuous-time stochastic systems via distribution-based approach (Q5953541)

This paper describes an approach for system identification of continuous-time stochastic state space models from random input-output continuous data. The basic idea is to adopt the concept of random distribution in the sense of Itô-Schwartz or the generalized random function by Gel'fand and Vilenkin. The authors describe the derivatives of (non-differentiable) stochastic processes in the sense of distributions and construct the input-output algebraic relationship by using the differential information produced in the distribution sense. A comparison with other approaches employing the filters is discussed.