Driver Risk-Taking Behavior as a Function of Congestion Level: Analysis Using Adopted Headways in Traffic Stream

At low traffic levels, there is little interaction between vehicles and the headways are random. However, when the traffic flow rate increases to a certain level, there are higher number of vehicles on a unit length of road and the headways between the vehicles are no longer random. At high traffic flow level, drivers can choose to maintain headways that are large enough to be safe in case the vehicles in front abruptly decelerate or stop. Alternately, they may also choose to maintain an headway. The objective of this research is to determine the adopted headway selection approach by the drivers and to test whether the adoption of the unsafe headways is a function of speed, vehicle type and traffic flow. A multinomial logit model based on traffic level, vehicle type, and vehicle speed was derived which predicted the number of drivers traveling unsafely at prevailing traffic conditions. It was found that there is no relationship between choosing safe headways and road environment when the traffic density is low (less than 18 vehicles per mile per lane). When the density of vehicles on the road is high (greater than 35 vehicles per mile per lane), there is a direct relationship between the drivers inclination to take risk (have headways that are unsafe) and prevailing traffic conditions (speed and volumes). When the density is intermediate (greater than 18 vehicles per mile per lane but less than 35 vehicles per mile per lane) there is some relationship but the prediction rate of multinomial model is not good, probably due to the mixture of random headways and interactive headways.