Identification of linear parameter varying system using flexible least squares

This paper introduces the flexible least squares (FLS) for the identification problem of linear parameter varying (LPV) system. FLS considers the measurement error (output estimation error) and dynamic error (parameter estimation error) simultaneously and is suitable for dealing with identification of parameter varying system. It is deduced that FLS is algebraically equivalent to the Kalman filter whose stability criterion has been well developed, a stability criterion is built for FLS in a similar way. Finally we apply FLS to the identification of LPV system, simulation results shown its effectiveness.