Simulated control of unilateral, anticipatory locomotor adjustments during obstructed gait

Abstract

Anticipatory adjustments of our locomotor patterns are necessary in order to negotiate our uneven daily environments. Recent work (McFadyen and Winter 1991) has shown the re-organization of lower limb mechanics for obstacle avoidance during level walking. The present work describes a model which sets the ground work for predicting how such re-organized motor patterns might be generated from stereotypic unobstructed patterns. Pattern-generating algorithms use an estimation of future contacts with obstacles to create weighting functions that modify joint angle trajectories towards new patterns capable of clearing the obstacle. Feedforward/feedback control is then used to generate the necessary joint torques. The results show that model parameters can be found to generate not only kinematic but also energetic patterns for obstacle clearance that mimic experimental results. The validity of the model with respect to human locomotor control is discussed.