Smart shoe for predicting knee abduction moment

This research aims to develop a smart shoe to measure a knee abduction moment (KAM) while walking. Previous research shows that activities with the high KAM are re- lated to create the future knee pain in elders. Various kinds of floors propose to reduce the risk of osteoarthritis and injury of falls in elderly persons. KAM is a factor in determining a supporting property of the floor, and it is usually measured from the optical tracking sys- tem and force plate in the laboratory that can not be moved. To overcome above issue, our research proposes a smart shoe with a portable sensor to estimate KAM in an open environ- ment. The smart shoe consists of an inertial measurement unit (IMU) sensor and six force sensors connected to the microcontroller to collect forces and kinematic data while walking. Then, a multi-layer perceptron regressor (MLP Regressor) model issue for predicting KAM. The two steps are used for this purpose. The first step is to predict the maximum KAM from the stance phase. The second step is to predict the KAM in a full human gait cycle. Finally, The proposed shoe can estimate the peak KAM with an accuracy of 86.95 percent at the stance phase and 94.3336 percent of the full human gait cycle compared to the actual KAM measured in the laboratory.