| The Special Purpose Dexterous Manipulator (SPDM) Ground Test Facility (SPDM GT) is a 7-degree-of-freedom (DOF) robot. This research investigates system identification of the SPDM GT for the purpose of improving control systems design. Consistent with current control design, the project identifies linear models which best approximate the nonlinear SPDM GT dynamics.; Simulation models of the SPDM GT with various nonlinearities are constructed for generating experimental data. The identification procedure applies a grey box identification methodology that uses a prediction error method (PEM) to identify the parameters of the linear model. The identified linear models are compared against nominal, locally linear models of the simulated plants. It is found that identified models of simulated plants that only contain the configuration-dependent nonlinearity perform well. Additional nonlinearities degrade the quality of the identified models and expose the limitations of some input signal characteristics. |
