Cellular automaton model for traffic flow based on safe driving policies and human reactions

This paper proposes a new single-lane cellular automaton model for traffic flow. The model takes into account normal drivers’ spacing policies and transportation engineering practices to guarantee that microscopic vehicle behavior is more in line with vehicular movement in the real world. As a result, drivers’ reactions are based on a safety analysis that determines the most appropriate action for a vehicle to take. Hence, the model introduces a new set of simple rules to change the speed of vehicles that incorporates three important thresholds required by the follower vehicle to accelerate, slow down or maintain its speed. Thus, the space gap, relative speed and limited acceleration/deceleration capabilities are introduced into simulations. Simulation results obtained from a system with periodic conditions show that the model can smooth the speed drop when vehicles approach the upstream front of the traffic jam. Therefore, the model avoids unrealistic deceleration behavior found in most previous cellular automata models. Besides, the model is also capable of reproducing most empirical findings including the three states of traffic flow, the backward speed of the downstream front of the traffic jam, and different congested traffic patterns induced by a system with open boundary conditions with an on-ramp. Moreover, the new model preserves the computational simplicity of the cellular automata models.