| Objective:To determine the optimal combination of wearing parts and activity types to provide a basis for the reasonable wearing of accelerometers in practice,to improve their accuracy in energy measurement;to help the accelerometer system application in scientific research,reasonable for fitness preferences Use accelerometers to provide a basis.Methods:A total of 33 students from Sichuan Normal University who were not healthy and trained were selected to wear the GT3X accelerometer in seven parts:arm,wrist,hip,thigh and ankle.With k4b2 as the calibration standard,there are 12 test items in three categories:resting class,physical exercise class and housework class.For each test item,five minutes,the second,third and fourth minutes are selected for statistical analysis.Results:(1)There is a certain correlation between the energy consumption value measured by the accelerometer of each wearing part and the value measured by k4b2.The correlation coefficient is between 0.68 and 0.83,and the energy consumption measured by the accelerometer of the hip is the most relevant.The number reached 0.83,and the lowest correlation coefficient was the accelerometer worn on the right wrist with a correlation coefficient of 0.68.The accelerometer's energy consumption measurement accuracy is very different depending on the wearing position.The hip and thigh energy measurement accuracy is higher than the wrist and ankle.In other words,the farther the wearing part is away from the subject's body center of mass.The lower the measurement accuracy.(2)Linear regression model of seven wearing parts,R~2 of the hip is 0.683,R~2 of the thigh is 0.530,R~2 of the ankle is 451,R~2 of the upper left arm and the upper right arm.They were 0.526 and 0.564 respectively;the R~2 of the left wrist and the right wrist were 0.474 and 0.442,respectively.(3)Equation verification:The consistency between the predicted values of the seven wearing parts and the measured values of k4b2 is relatively good,except that there is a slight difference in the average of the differences,but they are all within the 95%confidence interval,wherein the hips The consistency between the predicted and measured values is better than the predicted value of the thigh and the predicted value of the right wrist.(4)Overall RMSE and BIAS coefficients for each wearing site:RMSE of the hip,thigh,and left upper arm were the lowest,1.78,1.78,and 1.75,respectively,followed by the upper right arm,with an average RMSE of 1.94,and the two wrists and ankles.The RMSE is 2.46,2.71,and2.95,respectively.For BIAS,the thigh,right wrist,and right upper arm were 0.12,0.19,and 0.14,followed by the left upper arm and thigh with a BIAS of 0.21,again the hip average of 0.22,and the left wrist's BIAS of 0.26.(5)RMSE and BIAS coefficients of each wearing part under different activity categories:RMSE has the highest resting class,the average is above 2.00,followed by housework,and finally physical exercise,the average of hip,thigh and arm.Both are below 2.00,showing a decreasing trend.There is no obvious difference in the three activity categories of BIAS.In the wearing part,the RMSE of the hips in the resting activities is higher,reaching 2.03,and the housework and physical exercise are 1.57 and 1.76 respectively,thigh,The arm and hip were not significantly different under the three activity categories,and the difference between the ankle and the two wrists and the other four wearing parts was significantly above 2.00.There is little difference in the BIAS mean.When performing different types of physical activity,in addition to selecting the wearing part of the hip,the corresponding wearing part is appropriately selected to improve the accuracy of energy consumption measurement. |
PDF Full Text Request