Model of driver's braking moment in virtual car following collision avoidance scenes

To explore the impact factors of driver's braking moment in car following under emergency conditions, the virtual driving behavior tests of several drivers were conducted on a driving simulators. Based on the driver's braking onset data under different scenarios, the impact factors of driver's braking moment were analyzed, and a new threat assessment measure, time to maximum deceleration level braking, was proposed. This measure can directly quantify the danger or threat level of the current dynamic situation. Then two braking moment forecasting models were established by applying linear regression and BP neural network respectively. Results show that the driver's age and gender, the motion state between two cars are the important impact factors. Comparisons show that the BP neural network forecasting model predicts more accurately, that is in agreement with the human natural judgment on the urgency and severity of threat.