Abstract

Early gait rehabilitation training in stroke patients, often beginning within days of the stroke, is considered a promising method to correct hemiplegic gait patterns. Delaying the training may produce further impairments that can lead to joint damage and lessen the patient's quality of life. This work presents a Smart Knee Brace (SKB) that has been developed and refined to not only allow a physical therapist to pursue early gait rehabilitation in a patient, but also provides a flexible platform from which to develop effective rehabilitation methods with biofeedback.

This thesis contains a description and discussion of the physical hardware, the sensors and electronics, and the software implemented in the SKB system. The current system includes a locking mechanism for both the flexion and extension directions of knee motion. The locks may be used to induce proper knee motion and muscle use as well as to prevent the patient's knee from collapsing in flexion. There are various sensors to provide kinematic data about the motion of the leg which are used to control this locking mechanism. The real time control of the system is made possible by customized software developed in Matlab and Simulink using the XPC Toolbox.

The real time algorithm behind the control software is based solely on the kinematic data retrieved from the sensors on the SKB and was evaluated on a healthy patient wearing the SKB and through simulation with previously taken stroke patient data. It is shown to accurately detect the initial contact and toe off gait events.