Prediction of driving performance using Nonlinear Causal Resource Analysis

This research examines and demonstrates the potential of Nonlinear Causal Resource Analysis models in predicting driving performance. Motor-control, sensory, and central processing performance resources measurements were collected from a group of subjects (Group A (n=60) and Group B (n=9)), demonstrating a wide range of driving and performance resource availability. In car driving performance was determined by a skilled driving evaluator. Ten performance resources from subjects in Group A were used to construct NCRA models of driving. Model based predictions for Group B correlated (r=.83) with expert rater evaluations. For each prediction, a limiting resource was identified. The NCRA models, constructed with a modest number of subjects and basic performance resources, were able to well predict the level of driving performance, and identify the limiting resource for each subject.