Using trunk compensation to model head stabilization during locomotion

Head and trunk movements were measured in eight subjects performing a complex locomotor task. The purpose was to determine if, even in a complex scenario, the head is stabilized during walking. The experimental conditions were Free Gaze (FG) and Anchored Gaze (AG). In AG condition, subjects were asked to always gaze on one of two Anchor Points placed in the scenario during walking. Despite the linear and rotational motion of trunk, the head pitch orientation has little variation (9.2°) in AG condition respect to FG (23.4°). Moreover, the head pitch rotation relative to the trunk in AG condition appeared to compensate for trunk pitch rotation. In fact, an anti-phase correlation between these two signals occurred in the AG condition trials. In this perspective, we design a model in which the head is stabilized and the orientation is simulated for a locomotor task. The model we propose considers the trunk rotation as a disturbance and allows following an input reference head rotation, compensating the trunk rotation. This model, suitable for a future implementation on a robotic platform, is based on an inverse kinematics controller and tested on real human data.