Identification of state-space models by modified nonlinear ls optimization method

The problem of estimating parameter of LTI state-space system is addressed in this paper. A modified nonlinear least square(NNLS) method is proposed to the optimization of output error cost function. The nonuniqueness of the fully parameterized state-space is considered and the parameter update is restricted to the directions that change the input-output behavior of the system. Simulation studies show the improved performance of the proposed method.