Towards Eliminating Overreacted Vehicular Maneuvers: Part I Model Development and Calibration
Author(s)
Yu, Y
Zou, Y
Qu, X
Griffith University Author(s)
Year published
2019
Metadata
Show full item recordAbstract
Microscopic car following models are of great importance to traffic flow studies and vehicular dynamics reproducing. The Full Velocity Difference (FVD) model is a well-known example with satisfactory simulation performances in most times. However, by analyzing the structure of the model formulas, we find that it can sometimes generate overreacted vehicular maneuvers such as unrealistically strong (overshooting for short) accelerations or decelerations that conflict with normal driver habits or even beyond the actual vehicular acceleration/deceleration performance, especially when the target vehicle encounter a leader cut-in ...
View more >Microscopic car following models are of great importance to traffic flow studies and vehicular dynamics reproducing. The Full Velocity Difference (FVD) model is a well-known example with satisfactory simulation performances in most times. However, by analyzing the structure of the model formulas, we find that it can sometimes generate overreacted vehicular maneuvers such as unrealistically strong (overshooting for short) accelerations or decelerations that conflict with normal driver habits or even beyond the actual vehicular acceleration/deceleration performance, especially when the target vehicle encounter a leader cut-in or move out (leader lane change for short). As Part I of the entire research, this paper corrects the above deficiency of the FVD model by proposing a capped-Full Velocity Difference (capped-FVD) model in which we limit any potential overshooting accelerations or decelerations generated to a reasonable range. Then, all model parameters are also calibrated using field data. Performance comparative analyses to validate the performance improvement of the capped-FVD model are included in the other companion paper serving as Part II of this research.
View less >
View more >Microscopic car following models are of great importance to traffic flow studies and vehicular dynamics reproducing. The Full Velocity Difference (FVD) model is a well-known example with satisfactory simulation performances in most times. However, by analyzing the structure of the model formulas, we find that it can sometimes generate overreacted vehicular maneuvers such as unrealistically strong (overshooting for short) accelerations or decelerations that conflict with normal driver habits or even beyond the actual vehicular acceleration/deceleration performance, especially when the target vehicle encounter a leader cut-in or move out (leader lane change for short). As Part I of the entire research, this paper corrects the above deficiency of the FVD model by proposing a capped-Full Velocity Difference (capped-FVD) model in which we limit any potential overshooting accelerations or decelerations generated to a reasonable range. Then, all model parameters are also calibrated using field data. Performance comparative analyses to validate the performance improvement of the capped-FVD model are included in the other companion paper serving as Part II of this research.
View less >
Conference Title
Smart Innovation, Systems and Technologies
Volume
149
Subject
Urban and regional planning