Ankle active rehabilitation based on Intent recognition

Abstract

Aiming at the problem that most of the current ankle rehabilitation robots lack active patient participation, this paper proposes an intention-aware ankle active rehabilitation method. A plantar pressure sensor on the 2-SPU/RR ankle robot collects pressure data and uses an support vector machine(SVM) classifier to process the data to sense the patient's movement intention, using principal component analysis to reduce the dimensionality in order to improve the speed of classification recognition. In the first step, the rehabilitation trajectory is predefined with reference to the physician's treatment advice. In the second step, active following movement guided by movement intention is achieved based on plantar pressure, and the patient's active trajectory data is recorded. In the third step, a rehabilitation evaluation model is established through the trajectory deviation data and other indicators to objectively assess the participants' ankle rehabilitation status and provide theoretical reference for the physicians' diagnosis. The experimental results showed that the proposed ankle active rehabilitation strategy was effective, and the rehabilitation evaluation model quantitatively analysed the participants' ankle functional status based on the recorded indexes, which provided a reference for the formulation of the next step of rehabilitation training.