Data Analysis of wearable technologies autonomous systems sensor in physiotherapy, Conducted a comprehensive data analysis on Xsens MTx sensor data
Conducted a comprehensive data analysis on Xsens MTx sensor data, employing advanced statistical analysis and visualization techniques to develop a fully autonomous system. This system can detect sequential executions of various exercise types during a session, classify each exercise type, evaluate the correctness of each execution, and identify specific error types and idle time intervals. Additionally, the project included an analysis of market data to provide economically viable recommendations for enhancing competitiveness in the market.
The dataset was produced as part of a study on wearable motion sensors for automated detection of physical therapy exercise executions. A physical therapy specialist collaborates in determining the workouts and the trial protocol. This dataset comprises data from wearable inertial and magnetic sensors during physical therapy exercises, focusing on eight types of exercises executed in three different manners (correct, fast, and low-amplitude) by five subjects. Each subject wore five MTx sensor units from XSens, each equipped with tri-axial accelerometers, gyroscopes, and magnetometers, sampling data at 25 Hz. Detailed methodology and the purpose behind the dataset are elaborated in referenced studies, with one emphasizing the automated detection and evaluation of these exercises. The collection process received ethical approval from Bilkent University, ensuring participant consent and anonymity. This dataset includes data from wearable magnetic and inertial sensors collected while performing physical therapy activities. Physical therapy exercises come in eight varieties, with three ways to perform them: accurate, rapid, and low-amplitude.
Yurtman,Aras and Barshan,Billur. (2022). Physical Therapy Exercises. UCI Machine Learning Repository. https://doi.org/10.24432/C5JK60.