Evaluation of New Gyro-Sensor and Accelerometer Device to Estimate Physical Activity

Abstract

The purpose of this study was to assess the accuracy of a new motion sensor wristband (ViM sports memory: ViM), consisting of an accelerometer and a gyro-sensor, by comparing the accuracy with those of indirect calorimeter (IC) and a commonly used accelerometer (Lifecorder: LC). Twenty-five participants (13 males; 12 females) walked at 3.6, 4.8, and 6.0 km · h ؊1 and ran at 7.2 and 9.6 km · h ؊1 on a treadmill for 5 min. Then, another 10 males performed static stretching and hopscotch for 5 min each. Measured energy expenditure (EE) by the IC and estimations of the LC and ViM were compared by repeated measures ANOVA. During walking, differences between the IC and ViM (24 to 74%) were lager than those between the IC and LC (؊16 to 0%). During running, differences between the IC and LC (؊35 to ؊21%) were larger than those between the IC and ViM (؊17 to 14%). During static stretch and hopscotch, differences between the IC and ViM (stretching, ؊21%; hopscotch ؊40%) were smaller than those between the IC and LC (stretching, ؊40%; hopscotch, ؊66%). The ViM is more suitable than the LC for the estimation of EE during running or static stretching; however, the accuracy of the ViM was far inferior to that of the LC during walking.

Improvement of New Gyro-Sensor and Accelerometer Device Used for Walking and Running

“…The classiˆcation algorithm is based on the relationship between the amplitude of arm swing increases and walking speed (Donker et al, 2001;Ford et al, 2007;Wagenaar and van Emmerik, 2000). Takahashi et al (2009) showed that the ViM could distinguish walking speeds approximately correctly, whereas Takahashi and coworkers also reported that the accuracy of the ViM classiˆcation was inferior to that of a conventional accelerometer placed on waist. Because diŠerences in arm swing motion among individuals were larger than those of acceleration at waist, standard deviations of W1, W2, and W3 became large in the previous study.…”

“…In results of the previous study (Takahashi et al, 2009), there were obvious technical biases in the ViM outputs during the treadmill exercise: 1) there were few records of IR1, IR2, and IR3, because the arm swing during walking and running was a periodic rotation movement; and 2) the diŠerence between W2 and W3 became unstable among each individual, because the diŠerences in the arm swing among individuals were large at very low to moderate speeds. These biases probably resulted in problems with the calculation of regression coe‹cients of the recalibration equation.…”

“…Four ViMs were used in the exercise tests. The reliability and objectivity of the ViMs were reported in the previous study (Takahashi et al, 2009 ) was calibrated with a 3.0-L calibration syringe, and the O2 and CO2 analyzers were calibrated with room air and a calibration gas of known O2 (15.94z) and CO2 (3.97z) composition. The breath-by-breath data were averaged every 30 s. _VO2 measured by the IC were converted to METs and EE (kcal･min -1 ) by the following equations (Dale et al, 2002):…”

“…This device includes a uniaxial accelerometer and a gyro-sensor. The ViM classiˆes physical activity into 10 patterns, based on angular velocity and vertical acceleration (Takahashi et al, 2009). The cause of the problem lies in the estimation equation model of EE used by the device.…”

“…In a previous study (Takahashi et al, 2009), we examined the validity and accuracy of a new type of motion monitor, the ViM sports memory wristband (Microstone Co., Saku, Japan). This device includes a uniaxial accelerometer and a gyro-sensor.…”