Alcohol Effects on Driving Behavior and Performance in a Car Simulator

Abstract

A fixed-base simulation has been developed to test the effect of alcohol on driving performance. The simulation includes both lateral steering control and a discrete visual detection, recognition, and response task set up to provide the workload and division of attention typical of real world driving. Measurements of both driver control behavior and driver-vehicle performance were obtained for the steering task, and detection and recognition indexes and reaction time were measured on the discrete task. Preliminary results on scanning behavior as measured with an eye-point-of-regard monitor are also presented. Data are given for eighteen drivers, ranging in age from 21-65, at blood alcohol concentrations (BAC) of 0, 0.06, and 0.11. Alcohol causes larger lane and heading deviations, and increases detection and reaction times on the discrete task. Control-behavior measures show that the driver's control gain decreases, but stability margins are maintained under alcohol, while driver remnant increases. Such effects could be due to indifference thresholds and/or intermittent attention in the control task. Both continuous steering control and discrete peripheral ``sign'' response tasks were performed, singly and combined, to investigate the effects of divided attention. Performance on the steering control task was decreased when both tasks were done concurrently, but the sensitivity to alcohol effects was similar. The driving simulation has proven an efficient tool for alcohol research. It has gained acceptance from subjects as a valid approximation of driving, and the various related measurements have proven to be reliable and sensitive to levels of intoxication.