Frequency-domain identification of the human controller.

Area of Science:

• Human-Computer Interaction
• Control Theory
• Cognitive Science

Background:

• Human control systems are complex and can be modeled using control theory.
• Previous models often assumed continuous human control, neglecting intermittent control strategies.

Purpose of the Study:

• To objectively test three control-theoretical models of the human control system: non-predictive, predictive, and intermittent predictive control.
• To investigate the suitability of system identification techniques for analyzing human-in-the-loop data with intermittent control.

Main Methods:

• A two-stage system identification approach was applied to experimental human-in-the-loop data.
• The method involved deriving closed-loop frequency response from periodic data, followed by parametric model fitting.
• This technique was adapted for intermittent predictive control analysis.

Main Results:

• Non-predictive control models showed a poorer fit to the experimental data compared to predictive models.
• Predictive and intermittent predictive control models provided equally good fits, indistinguishable by the applied method.
• Identified model parameters correlated with the specific control strategies (position or velocity focus) adopted by participants.

Conclusions:

• Sustained manual control is compatible with intermittent control strategies.
• The findings challenge the necessity of continuous control models for human manual control.
• System identification provides a robust method for differentiating and validating human control models.