Identification and Validation of a Car-Following Model with a Differential Control Term for Traffic Congestion at a Motorway Sag Section

Abstract

Vehicle-following behavior in a passing lane was analyzed based on vehicle trajectories estimated by video images synchronously acquired by eleven cameras that were installed for four days at the Yamato sag outbound section of Tomei motorway. Discriminant criteria of lead vehicles of platoons were formulated based on closed-loop behaviors of vehicles on time-headway and relative-speed planes. The parameter values of a vehicle-following model with relative-acceleration as a differential term were optimized for root-mean-square errors of the estimated positions of vehicles. The model was applied with optimized parameters to the simulation of leading uncongested and congested vehicle-platoons. The model successfully demonstrated the advantages of reproducibility to the existing models. Vehicle-following behavior of congested platoons was simulated for four days after the lead vehicles had been replaced by those of leading uncongested platoons. The results reconfirmed a possibility that congestion phenomena were influenced by the reduction in the velocity of the lead vehicles of the platoons at the upgrade section as well as their initial velocity.