Minimization of number of gait trials for predicting the stabilized minimum toe clearance during gait using artificial neural networks

Artificial neural networks (ANN) have been increasingly used in gait analysis. Back-propagation neural network has been widely used because of its good predicting power in supervised training mode for gait data analysis. In this paper an artificial neural network was used to model relationships between minimum toe clearance (MTC) characteristics derived from fewer gait trials and that derived from gait data during a 30-minute continuous treadmill walking. The ANN was separately trained and tested with nine statistics calculated from 10 different data segment lengths as inputs, and the mean and standard deviation of MTC data calculated from 30 minutes gait trials as outputs. The results suggest that a trained ANN is able to accurately predict stabilized MTC data, even a 5-gait cycles' data predicted with about 80% accuracy and the prediction accuracy was seen to improve with increase in the length of input data segment.