Modeling and analysis of a manual wheelchair coast down protocol

The purpose of the paper is to mathematically model a manual wheelchair on a dynamometer and flat tile floor respectively, using a simplified version of R.A. Cooper´s model (J. Rehabil. Res. Dev., vol. 27, no. 2, p. 151-62, 1990). Then the model is compared to measured data from coast down trials using the pearson-product correlation coefficient. Using a 2nd order polynomial to model the rear wheel angular position as a function of time we obtained pearson-product correlation coefficients which ranged from 0.9988 to 1.0000 with a mean of 0.9998. From the 2^nd order polynomial the authors were able to measure the friction torque which was used to compare the dynamometer to a flat tile floor. The friction torque for the dynamometer and the poor were significantly different for both sides when compared using analysis of variance with a significance value of 0.05. Therefore, a braking system should be added to the dynamometer in order to increase the dynamometer friction torque in order to match that of the flat tile floor. The normalized difference of the side to side friction torque for the dynamometer and the floor were not significantly different. This suggests that the side to side friction torque differences are due to the user and wheelchair setup not the dynamometer