Intelligent Rehabilitation Devices
Variable Resistant Exercise Devices
There is a broad interest in variable resistance devices for rehabilitation applications. We have developed novel MR fluid rehabilitation exercise devices for knee and other joints. The device provides variable resistance as a function of the joint angle. The resistance profile is determined by the physical therapist for each patient. The device is programmable to allow the therapist to input the resistance profile for the patient. An intelligent controller has been developed to implement the variable resistance profile while accounting for the muscle fatigue. The hypothesis of the research states that for a joint, there is an optimal resistance profile for muscle strengthening and rehabilitation of injured muscles. Clinic evaluation of the device is underway.
Smart Knee Brace for Stroke Patients
We are developing intelligent knee braces for early gait training of stroke patients. The hypothesis of the research states that the early rehab and exercise of patients soon after having stroke will provide much better recovery of the ability to conduct normal physical activities in the long run. The smart knee brace (SKB) allows physical therapists to provide gait rehabilitation and to study gait patterns of stroke patients.
The brace is instrumented with various motion sensors, a fixed extension lock, a variable angle flexion locking mechanisms, and controllable damper. A micro-processor based controller will be developed with advanced real-time gait event detection algorithms. To control the locking mechanism, the SKB control algorithm requires information based on knee kinematic data and gait events. The necessary kinematic data is obtained through sensors attached on the brace. Currently, gait events are determined through a foot switch array wired directly to the brace. A real-time gait detection algorithm is being developed based solely on kinematic sensors directly attached to the brace. This algorithm would eliminate the need for the foot switches in hope of having a self-contained brace with no external wiring.
The above projects require a sound knowledge on modeling of human-machine interactions, nonlinear adaptive controls, sensor and actuator studies, and signal processing. The work is in collaboration with Dr. Katherine Rudolph from the Department of Physical Therapy, and is supported by National Institute of Health.