Prediction of Red-Light Running using an Artificial Neural Network

Twenty-one percent of total highway fatalities occurred at or within an intersection, in 2017. When a driver is approaching a signalized intersection, and the traffic signal changes from green to yellow, the driver has to make a decision whether to safely stop or attempt to proceed through the intersection. Unlawfully passing through the signalized intersection can result in a dangerous traffic accident. The driver's decision can possibly result in a rear end crash (if driver stops abruptly) or a right-angle crash (if driver illegally runs red-light). An astounding 96% of drivers surveyed stated that they have a fear being hit by a red-light runner. Variables that have an effect on driver's decision is drivers distance from the stop line, drivers sex, and drivers age. In addition to making this decision of whether to proceed through the intersection, there are other factors (cell phone utilization) that can impact the decision of the driver. Cell phone utilization can be broken down into 2 task: incoming and outgoing call, and further broke down into 3 different interactions (Hands Free, Hand Set, and Hand Held). Previous studies indicate that cell phone utilization has a negative impact on driver's decision due to the distraction it creates. We propose to use an Artificial Neural Network (ANN) to predict whether a driver will run a red light, based on various features including features involving distracted driving.