California Partners for Advanced Transportation Technology

An important perceptual task during driving is the ability to detect and avoid collisions. Failure to accurately perform this task can have serious consequences for the driver and passengers. The present research developed and tested a model of car following by human drivers, as part of a general model under development of a human driver. Unlike other car following models that are based on 3D parameters (e.g., range or distance) the present model is based on the visual information available to the driver. The model uses visual angle and change in visual angle to regulate speed during car following. Human factors experiments in a driving simulator examined performance in car following to speed variations defined by sine wave oscillations in speed, sum of sine wave oscillations, and ramp function. In addition, using real world driving data the model was applied to 6 driving events. The model provided a good fit to car following performance in the driving simulation studies as well as the real-world driving data, accounting for up to 96% of the variability in speed for the real world driving events.