A Probabilistic Walk Path Model Focused On Foot Landing Points And Human Step Measurement System

This paper describes a model of human walking and the development of the step measurement system. We have used a probabilistic method to make the model that focuses on the relationship between the orientation of the foot placement and where the other foot will land. We make a hypothesis that the foot turns to where someone wants to go in walking naturally to make the model. An experiment using a motion capture system (Vicon Motion Systems, 200 Hz) was conducted to prove the validity of the hypothesis and shows that we can predict the next landing point using the orientation of the foot. Also it indicates we can generate likely walk paths stochastically. The step measurement system we developed does not require anything but a motion capture system. Experiments show that this system can detect the timing of heel contact and toe lift-off with high accuracy.