Nonlinear identification of marine thruster dynamics

We present a general identification framework for a particular class of single-input single-output nonlinear dynamical systems consisting of the cascade of a static nonlinearity with a linear dynamical system. Different linearly parameterized decompositions of the static nonlinearity are introduced and compared. A parametric identification problem is formulated and a least square estimation algorithm is used for model recovery. Numerical examples are presented to verify the proposed algorithm, and the algorithm is finally used to recover the dynamics of a marine thruster using experimentally sampled input/output data. Experimental results show the effectiveness of the proposed algorithm