Assessment of daily physical activities with sensors attached to the upper limbs in healthy adults using a manual wheelchair

This study examined the energy expenditure (EE) of healthy adults during typical use of a manual wheelchair by attaching sensors to the subjects’ upper limbs. The aim was to determine whether the measured EE values depend on the sensor attachment site and whether the addition of angular velocity information to the acceleration value was advantageous to the EE assessment. Subjects were 11 males and 10 females. Their wrists and mid-upper arms were fitted with sensors to monitor their daily physical activities. Triaxial acceleration, triaxial angular velocity, and EE were measured while performing activities with a manual wheelchair. Coefficients of determination for estimating EE at each sensor location ranged from 0.66 to 0.79 (based on gender, the calculated three axis value of acceleration and angular velocity) and from 0.65 to 0.78 (based on gender and the calculated three axis value of acceleration, without angular velocity). Furthermore, angular velocity was not selected as a significant explanatory variable for estimating EE at the wrist. The average percent error for estimating the EE of daily physical activity in healthy adults using a manual wheelchair and factoring in gender and the calculated three axis value of acceleration at each sensor location ranged from 5.2 to 7.2%. Angular velocity information added to the calculated three axis value of acceleration at the upper arms slightly improved the estimation of EE. In addition, it was found that there was no difference in the assessment at different sensor locations.