An optimal-control model of vision-gait interaction in a virtual walkway

The specific aim of this paper is to model the vision-posture coupling behavior, which is important for astronauts to stabilize their locomotion in partial gravities as the National Aeronautics and Space Administration plans for manned missions to the Moon and Mars . As such, an optimal scheme is assumed in postural-control processes to stabilize visual optical flows. An experiment was conducted, in which human subjects attended a visual-gait tracking task. In tracking control, head position errors can be used to regulate inputs so that appropriate compensatory changes can be obtained. The "optimal" scheme describes a compromise between postural adjusting efforts and tracking errors. The results show that the proposed optimal-control model describes the gait tracking process more reliably than McRuer's crossover model of the human-plant compensatory behaviors. In practice, if the tracking goal is to be roughly right rather than precisely wrong, this paper also provides the experimental data regarding the human tolerance and achievable performance under various unloading conditions and tracking difficulties. This information and related experimental setup could also be applied to postsurgery gait rehabilitation.