Functional Task based Assistance during Walking for a Lower Extremity Assistive Device

In this paper, we propose a functional task based assistance controller to aid user in the walking task with our Lower Extremity Assistive Device (LEAD). Firstly, a gait period detector, which utilizes a Gaussian Mixture Model (GMM), is developed to estimate the user's current gait period among the six major periods. Then, an impedance based controller is used to apply assistive torques to the hip and knee joints of the user based on the functional task intended at the current gait period. To validate the above control scheme, preliminary experiments have been performed with one healthy subject walking on a treadmill. The results show that the gait period detector can effectively detect each gait period for the whole cycle. Based on measurements of the heart rate, the proposed assistance method has shown that it can effectively assist a human user in walking at speed of 1 km/h.