A study on the operation of rehabilitation interfaces in active rehabilitation exercises for upper limb hemiplegic patients: Interfaces for lateral and bilateral exercises

Abstract

For implementing autonomous rehabilitation exercises for upper limb hemiplegic patients, interfaces and a rehabilitation scenario that allow lateral and bilateral motions in a rehabilitation exercise robot are proposed. The proposed method measures the motion information generated from the unaffected part and projects it to an affected part in which the affected part expresses motions of the unaffected part. Both the accelerometer and gyro data were merged for estimating the motion information of the unaffected part. Also, HDR and complementary filters were applied to improve measurement errors in a data merging process. For verifying the proposed method, a device, which is similar to a human body joint, was fabricated. Then, the angular values estimated by using an inertial sensor and the encoder values from the device were compared. In addition, a camera analysis was used to verify the proposed rehabilitation scenario by applying the rehabilitation interface proposed in this study to an exo-skeleton robot arm. It is possible to apply the method proposed in this study to the control variables in different upper limb rehabilitation exercise robots. Thus, it is expected that patient centered active lateral/bilateral rehabilitation exercises can be performed through this interface method.