A Clinic-Oriented Ground Reaction Force Prediction Method in Gait.

Gait is a feature set to describe the human walking state, so it is one of the important methods for doctors to diagnose, evaluate, and judge the rehabilitation process of patients with neurological diseases in clinic. However, at present, doctors often rely on scales to evaluate patients’ gait performance, which are often biased by patients’ self-perception and doctors’ subjective experience. With the development of MEMS technology, wearable sensors have gradually been applied in clinic practice, but most of wearable sensors in clinic are still limited to the acquisition of kinematics data, and seldom involves dynamic evaluation, which makes the lack of multifaceted evaluation. In view of the above problems, the method proposed in this paper realizes ground reaction force prediction by simplifying the human dynamics walking model and multiple nonlinear regression model based on inertial measurement units (IMUs) attached on shanks. The proposed method has been verified on 6 healthy subjects and 8 stroke patients, and the mean accuracy for all subjects is controlled within 7.4 % , which has good clinical application value.