Determining the human ability to judge inertia during a dynamic pushing task

TR Number
Journal Title
Journal ISSN
Volume Title
Virginia Tech

The purpose of this research was to study the human ability to perceive inertia and to investigate the relationship between individual sensitivity to mass discrimination and the accuracy of inertia judgment. The study involved two major experiments, a mass discrimination test to provide a measure of kinesthetic sensitivity and a push/aim task to measure inertia perception. The experimental apparatus allowed for the manipulation of an inertial load for both the mass discrimination test and the push/aim task.

The mass discrimination test was based on the method of constant stimuli and involved a standard stimulus of 30 kg and seven comparison stimuli ranging from 24.3 kg to 35.7 kg. It determined Weber fraction values which were used as a measure of sensitivity. The push/aim task simulated a dynamic horizontal pushing task and was restricted to an initial exertion causing the inertial load to travel to a target under its own momentum. Performance was measured by the errors associated with the accuracy of the aimed pushes. These errors were separated into an absolute component representing an overshoot or undershoot, and a relative measure of error representing variability of repeated trials about the absolute error.

Distance to the target (2 or 6 m), amount of load (25, 45, or 90 kg), and apparent size of the load (small or large) were manipulated in the dynamic pushing experiment. Hypothesis testing was used to determine whether performance exhibited a consistent overshooting or undershooting of the target. An ANCOVA was used to measure the effects of distance, amount of load, and apparent size on performance given a covariate of mass discrimination sensitivity. Since the ANCOVA showed no significant effect from the covariate, an ANOVA was used to analyze the effects of the independent variables on both measures of error.

The hypothesis tests revealed that absolute error in the push/aim task was significant by less than zero for the combined data set and at all levels of independent variables, indicating a consistent undershoot of the target. The ANCOVA results il indicated no significant effect from the covariate, mass discrimination Correlation coefficients (R2) were calculated for the relation between sensitivity to mass discrimination and push/aim task performance with respect to absolute and relative error. These values were 0.004 for the absolute error and 0.008 for the relative error. The ANOVA results revealed that weight, distance, and the weight x distance interaction had significant effects on the absolute measure of error. Increasing the levels of either weight or distance resulted in significant increases in the amount of absolute error. Increasing the distance resulted in significant increases in the amount of relative error. The presence of a distance effect and the absence of a weight effect indicates the importance of rate of force over amount of force in the relative error associated with the push/aim task. Apparent size of the load did not have a significant effect on either absolute or relative error.