Symbolic identification of dynamical systems: Theory and experimental validation

This paper proposes a syntactic method of system identification in dynamical systems. The underlying concept involves abstraction of a qualitative description of the dynamical system by state-space embedding of the output data-stream and discretization of the resultant pseudo-state and input spaces. The task of system identification is achieved through grammatical inference techniques, and the deviation of the plant output from the nominal estimated language generates a measure of anomaly in the system. This system identification technique has been experimentally validated for detection of anomalous behavior on a laboratory apparatus of a permanent magnet synchronous motor (PMSM) undergoing gradual demagnetization.