The use of inertial measurement units associated with various algorithms has been proposed and developed to evaluate functional abilities and kinematics for stroke patients. In previous research, complex mathematical models were adopted successfully to clarify and to validate the functional results from different sensors. However, only a few algorithms stemmed from the process of motor recovery after a stroke or the way to administer clinical assessment scales. Based on the recovery process or how to conduct the assessment scales, the algorithmsensor based module is supposed to accurately classify clinical motor recovery status and to provide additional kinematics in stroke survivors. In this study, only one sensor is affixed on the dorsum of the affected foot to reduce the burden on a weak extremity. A special movement while in seated, extend their knee and then dorsiflex their feet, based on the motor recovery process after stroke is proposed and tested to classify Brunnstrom stages for lower extremities. After analyzing 24 participants and adopting suitable threshold values for different Brunnstrom stages, the overall accuracy is 86.8%. The ability to distinguish Brunnstrom stage II from others can even reach a 100% accuracy. The accuracies for distinguishing Brunnstrom stage III, stage IV, and stage V are 86.6%, 94 %, and 92.8%, respectively. We also analyze these misclassified data and investigate why the errors occurred. The results reveal the feasibility of the kinematics-based algorithm even using a single sensor.
|期刊||Journal of Physics: Conference Series|
|出版狀態||已出版 - 17 7月 2020|
|事件||2020 5th International Conference on Precision Machinery and Manufacturing Technology, ICPMMT 2020 - Auckland, New Zealand|
持續時間: 3 2月 2020 → 7 2月 2020
指紋深入研究「Adopting a Single Inertial Sensor and Designed Motion to Classify Brunnstrom Stages for Lower Extremities on Post-stroke Patients」主題。共同形成了獨特的指紋。
- 1 已完成