SUMMARY

System identification techniques have been applied to model human sensorimotor systems for research and diagnostic purposes. Accurate human modeling via system identification requires informative experimental data, which further require informative perturbations applied to the human. In the past, designing ideal perturbations has been extensively studied. However, physical realization of the designed perturbations, a process that may degrade perturbation quality, has not been investigated enough. A notable application is the design and use of the pseudorandom sequence. This work aims to: 1) propose performance metrics based on the spectral flatness measure to evaluate perturbation and control system performance, 2) design reference prefilter and feedback control to generate physical pseudorandom perturbations with improved spectral flatness, and 3) validate the proposed control design approach with human system identification experiments. Overall, the proposed performance metrics will facilitate the comparison and optimization of perturbation design and physical realization. The proposed control design approach will allow for separation of prefilter design and feedback control design, simplifying the control system implementation. The outcomes of this work can be applied to a variety of system identification applications requiring high spectral flatness physical perturbations.