This research is intended to develop a prototype digital simulator for the prediction of accident involvement of alcohol-impaired drivers. The simulator will be based on the quantitative results of numerous laboratory research efforts which have studied the effect of alcohol on individual sensory, psychomotor, and decision-making skills important to driving. The software implementation will be based on an object-oriented approach in which the driver is comprised of objects (visual, aural, motor, decision- making) which interact to simulate receptor of stimuli from the traffic environment (roadway, vehicles, obstacles, weather). This environment will also be modeled as objects interacting with the alcohol-impaired driver. The behavior of all simulation objects will be based on first principles (e.g., dynamics of vehicles) or on the basic empirical results of laboratory studies (e.g., sensory, psychomotor, decision-making). The driver's decision-making process will be implemented using a rule-based expert system to provide additional realism. Probabilistic effects will be incorporated into the simulation using Monte Carlo techniques. Successful completion of the Phase I prototype will set the stage for the Phase II development of a comprehensive simulator featuring advanced interactive graphical input and output facilities.