Nonlinear system identification of hystersis-backlash in hammerstein models

摘要

The aim of this work is to identify the two parts of hammerstein systems: linear and nonlinear, separately. The hysteresis-backlash is chosen to be the nonlinear part in these systems, as it includes memory nonlinearities. The identification of linear part is done using an appropriate system parametrization and a specific input signal. The nonlinear part is designed so that it provides persistent excitation and makes the internal signal measurable in the considered parametrization. Once the linear model is identified, the nonlinear element model identification will be easier. This is achieved using two appropriate parameterizations and specific input signals. The proposed identification method showed a satisfactory results.