Modeling urban road risky driving behaviors in China with multi-agent microscopic traffic simulation

Typical driving behaviors such as car-following and lane-changing can be described based on common concepts. But these behaviors could be different from driver to driver, from nation to nation due to different individual influencing factors (e.g. age, gender, driving age, mood) and situational influencing factors (e.g. weather, congestion, respect for law). Studies show that drivers who have higher level of "driving discourtesy" (i.e rudeness and aggressiveness) have higher probability of performing risky driving behaviors including traffic rule violations. In this paper, we propose a model named Driving Discourtesy Model (DDM). In this model, a new indicator is defined to measure the "driving discourtesy". With a probability distribution method, we are able to estimate the probability of performing risky driving behaviors of each vehicle based on a vehicle's individual influencing factors and situational factors. A multi-agent traffic simulation is developed to test DDM. The experiment results show that risky driving behaviors including speeding, lane-changing for taking speed advantages and driving on hard shoulder can be simulated effectively using DDM.