Gait Prediction of Swing Phase Based on Plantar Pressure

For the lower limb exoskeleton system, gait prediction is of great significance for human-machine collaborative control. This paper proposes a method for predicting the gait of lower limbs based on plantar pressure. A set of lower limb assistive exoskeleton is used to collect the subject’s gait data and plantar pressure at different walking states. A gait database is established, and the step rate and step length are used as indexes to query the database for the predicted gait trajectory. The feasibility of this query method is verified using the dynamic time warping algorithm. The feature quantities of plantar pressure are extracted, and BP neural networks are established and trained to predict gait features. The process of training and prediction is applied to several subjects. The experiment results prove that this method can effectively predict the gait of the swing phase.