Analyzing Stop Time Phase Leading to Congestion Based on Drivers’ Behavior Patterns

Traffic oscillation, stop and go traffic, is created by different reasons such as: sudden speed drop of leader vehicle. Stop and go traffic commonly is observed in congested freeways results in traffic oscillation. Many theories had been presented to define congestion traffic based on laws of physics such as: thermodynamics and fluid. But, these theories could not explain the complexity of driving responses in different situations of traffic especially in traffic jams. Unfortunately, because trajectories data are very scarce, our understanding of this type of oscillations in congested traffic is still limited. When the leader vehicle of a platoon drops speed, deceleration waves are released from downstream to upstream. Follower vehicles reacts different behavioral reactions based on personal characteristics. In this paper, behavioral patterns of follower driver were classified based on asymmetric microscopic driving behavior theory and traffic hysteresis in NGSIM trajectories. They were four patterns in deceleration phase and two patterns in acceleration phase. Then, two parameters of last deceleration wave leading to congestion, time and space parameters, τ and δ, were calculated based on Newell’s car following model. Time of two phases, stop and congestion phases, were identified based on follower vehicle trajectory. In order to calculate time of two phases, two points were identified: point of receiving stop wave leading to congestion and point of entering to congestion. Artificial neural network models were developed to analyze the relationship between the microscopic parameters and time of two phases. Analysis results present spacing difference of follower between stop and congestion phase based on under reaction-timid pattern and spacing difference of follower between deceleration and congestion phase based on over reaction-timid pattern and spacing of leader vehicle at the wave diffusion point are most effective parameters in stop time leading to congestion. One of the main practical applications of this paper can be the addressing one of the main problems of micro simulation soft wares (like Aimsun) due to behavioral patterns.

کلیدواژه ها:

Stop time lead to congestion ، Stop and go traffic ، NGSIM data trajectory ، Behavior patterns ، Artificial neural networks