The Development and Verification of Automated Adult Physical Fitness Measurement System

• Main Authors: Chen-Yu Lin, 林宸羽
• Other Authors: Wen-Hsu Sung
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/z38mjd

碩士 === 國立陽明大學 === 物理治療暨輔助科技學系 === 107 === Background and objective: In recent years, people’s lifestyle has changed gradually. Most of the leisure time is in static activities, resulting in insufficient physical activity and physical function decline. Physical fitness is an indicator that detects whether physical activity is enough. If People have good physical fitness, they can reduce the risk of chronic diseases and sports injuries, and improve the quality of life and work efficiency. The present measuring fitness way is a testing item that requires equipment and needs a larger site. However, the place where registration can be applied for testing is also limited. The development of the automated physical fitness measurement system is not perfect relative to the method of traditional manual measurement. And the prices of the ready-made system that can be found are quite expensive at present. The somatosensory device is a new measuring method that can track the body's skeleton without contact and marking points. It can be used to perform the different items of physical fitness by designing suitable software. The objective of this study is to verify the feasibility of using a somatosensory device to measure adult physical fitness. Methods: This study predicts that for automatically determines their performance when they perform physical fitness testing, using a Kinect V2 sensor tracks the subjects’ skeleton and measures their heart rate. The items of the physical fitness measurement include sit forward, 1-mm sit-up and 3-mm step test. The 23~65 years of the 21 healthy adults were recruited for this study. Each subject needed two separate occasions approximately 7 days apart to evaluate the test-retest reliability. And the validity of comparing the results of three items using Kinect V2 to the results of the manual measurement. The Pearson correlation coefficient and the consistency coefficient were used to validate the physical fitness measurement with a somatosensory device using this system. P <0.05 had a significant correlation. Intraclass correlation coefficient(ICC) tests were calculated to determine the test-retest reliability for this system. The alpha level of statistical significance was set to 0.05. And the Bland-Altman difference map with the 95% consistency line shows whether there were significant differences due to system or exercise effects. Results and Discussions: The measuring results in this system and manual measurement were highly correlated and consistent (r=0.986, p<0.01, rc=0.978) in 1-mm sit-up, and highly correlated and moderately consistent (r=0.942, p<0.01, rc=0.933) in sit forward. But the 3-mm step test was not significant (r=0.225, p=0.328, rc=0.127) in this system. The three measurements of the test-retest reliability was 0.910(p<0.01), 0.934(p<0.01) and 0.453(p=0.113), respectively. There had errors in the posture interpretation because of the variability of body movement patterns for each subject. And the program of this system didn’t perform individual correction in the heart rate measurement, resulting in measurement errors. Conclusions: The results in this system for the 1-mm sit-up and sit forward of the adult physical fitness measurement is feasible, and the 3-mm step test should reduce measurement error by the individual correction. In the future, it can provide a method for automatically measuring adult physical fitness measurement with non-contact and low-cost somatosensory sensors.