Estimation of Mental Workload during Motorcycle Operation

Abstract

In this paper, we propose a method of estimating the mental workload (MWL) of a motorcycle rider through the application of machine learning methods. The eye movement parameters of the motorcycle rider are measured under two run objectives, the one leads to a high MWL and the other one leads to a low MWL. The parameters are saccade duration, eye fixation (short stop), tracking frequency, saccade amplitude, and most frequency eye movement velocity. They are taken as explanatory variables for a discriminant function. By applying machine learning methods, we find that we can determine rider's MWL under preset running conditions with a high accuracy (≥80%).