Estimating Hand-rim Force of Wheelchair Propulsion Using Portable Sensors

Shoulder pain and injury are highly prevalent among manual wheelchair users with spinal cord injury (SCI) [18]. The consequences of upper extremity (UE) pain are significant because manual wheelchair users rely on their arms for mobility and independence. Specific biomechanical parameters of wheelchair propulsion such as cadence, magnitude of force, and pattern of the hand during the no propulsive part of the stroke have been associated with risk of upper extremity injury. People who push with greater force have higher risk of developing nerve dysfunction and UE pain. Therefore, forces experienced at the shoulder during wheelchair propulsion should be reduced [9]. This paper presents a study where three tri-axis accelerometers placed on the upper arm, and under the wheelchair’s seat together with a wheelchair rotation monitor (WRM) were used to estimate the force applied during propulsion. Twenty six participants with SCI were asked to push their chair equipped with a SmartWheel (SMW) on level and sloped surfaces. The estimated force obtained by using a regression model based on bagging decision trees was compared with the force collected from the SMW. Mean absolute errors (MAE) and mean absolute percentage of error (MAPE) were calculated between the estimated and criterion values. Pearson’s Correlation Coefficients (R) and R were also calculated between the criterion and the estimated force. Results showed reasonable accuracy in determine force trend with a correlation coefficients ranging from 86 to 88%, but limited accuracy in determining the force value with mean absolute percentage of error ranging from 20 to 22.4%.