Human Steering Control Under Unpredictable Disturbances

Abstract

There are multiple methods for controlling dynamical systems under external disturbances, with some emphasizing robust performance and others prioritizing efficiency. Here, we characterized the control strategies of human steering behavior under unpredictable disturbances. Using a self-steerable motion platform, participants (n = 22) steered a vehicle along roads of different widths (narrow, medium, and wide) while experiencing random, time-varying physical perturbations. Steering control strategies were assessed in terms of the frequency-dependent gain and phase of compensatory steering for the different road widths. Participants' road-keeping performance depended on the road width. When transitioning from narrow to wide roads, participants carried over their existing control strategy, suggesting a control policy that prioritizes robustness. This persistence was reflected in the vehicle spending more time within the road boundaries of the wide road than on the narrow ones. When moving from wide to narrow roads, participants again largely maintained their control strategy, albeit with some modest modulation by road width. In the frequency domain, road order did not affect the gain but did modulate the phase of control, suggesting that any control policy adjustments occurred primarily in the timing between perturbation and response. Together, these findings suggest that feedback gains for steering under random perturbations reflect a neural control strategy mainly tuned for robustness, with only a modest influence of efficiency.