On Car-Following Models Of Traffic Flow And Their Applications

Intersections are the high-risk areas of urban traffic congestion.Traffic signals are the essential means to alleviate the intersection congestion.Most of the existing signal timing designs adopt a static constraint method rather than a dynamic one in the con-trol method.And they always take the vehicle queue length as the control index,and then cause a problem that the inconsistency between releasing time and the releasing numbers of vehicles.Therefore,based on the real situation that vehicles passing the intersection,we propose several traffic flow models as dynamic constraint models for the timing optimization,analyze the relationship between the signal timing and re-leasing numbers of vehicles,and discuss the intersection signal timing optimization problem on basis of the proposed traffic dynamic constraint models.The following related subjects are investigated in this dissertation:1.As the important carrier to realize traffic,road has an important influence on traf-fic flow.By using the car-following model and the lattice model,the corresponding models are established to investigate the influence mechanism of gradient highway and curved road on traffic instability and the evolution of traffic jams,respectively.(1)Based on the fact that the safety distance between the vehicles,which driving on the gradient highway,is affected by the slope,and the relative velocity has an important effect upon the stability of traffic flow,the effect of gradient highway on the stability of traffic flow has been studied by using the car-following model.The results of numeri-cal simulations and theoretical analysis show that compared with flat road,the safety distance between vehicles in uphill situations becomes smaller,while downhill cases becomes bigger;In uphill situation,traffic flow becomes stable with the increase of s-lope and downhill case is just the opposite;Relative speed has a positive effect on the stability of traffic flow.(2)Considering topography conditions,economic factors and driving safety,in real traffic,a road may be built as curved road.Traffic flow on curved road is different from the one on straight road.Lattice model is established on the ba-sis of car-following model.By the use of the existing lattice model,an extended lattice model for traffic flow has been established with the consideration of curved road.Dif-ferent from the previous work,we not only consider the effects of curvature radius of curved road and friction coefficient on traffic flow,but also study the impacts of angle going into curved road,the radian and length of curved road upon the stability of traf-fic flow.The results of numerical simulations and theoretical analysis show that traffic flow is greatly affected by the angle going into curved road,the radian and length of curved road.The maximal theoretical flux and velocity of traffic flow are influenced by the above factors as well.We further analyze the two models which reflect the impact of road conditions on traffic flow using the measured data,respectively.Compared with experimental data,the results show that the improved models are better than the original model in reflecting the measured data.2.Two new car-following models have been proposed to investigate the effect of driver-s on traffic flow.(1)Based on the fact that drivers need a certain reaction time to react to the headway,velocity and other factors,an extended optimal velocity difference mod-el with the consideration of driver's reaction time-delay(OVDD)has been proposed.The results show that the occurrences of unrealistic high deceleration and negative velocity at low density region,which exist in the previous work,have been eliminat-ed effectively.The model also can describe the traffic instability,evolution of traffic jam and stop-and-go traffic.(2)In the process of driving,there often has a deviation when the drivers,especially the unskilled drivers estimate their own speed.Aiming at this phenomenon,a modified car-following model has been proposed with the con-sideration of velocity deviation which represents the inexact judgment of velocity.The proposed model is successful in anticipating the kinematic wave speed at jam density and the delay time of car motion,and it also can successfully predict the processes of acceleration and deceleration.The unrealistic negative velocity at low density region,which exists in the Full velocity difference(FVD)model,has been eliminated effective-ly.Meanwhile,the analytical results show that the proposed model can describe the effect of initial small disturbance on the stability of traffic flow and the mechanism of the occurrence of different nonlinear density waves.Meanwhile,we analyze the two models which reflect the impact of drivers on traffic flow using the measured data,respectively.3.Pedestrian flow and traffic flow belong to the self-driven many particle systems,in this sense,they have some similarities.A central challenge in pedestrian flow model-ing consists in predicting the level of service(LOS)of walking facilities.As the typical LOS indicator,pedestrian density is an important factor for safety and customer sat-isfaction.Unlike the previous work,based on the lattice model and considering the effect of density difference,an extended lattice hydrodynamic model for bidirectional pedestrian flow is proposed.The proposed model can describe the effect of density difference upon the evolution of pedestrian jam and instability.The results show that pedestrian flow will become stable with increasing of density difference.The maximal theoretical flux and velocity will become small with increasing of density difference.4.Vehicles usually take the following behavior to go through a signal intersection.Us-ing the measured data,the factors influencing on the number of vehicles going through the intersection have been analyzed.And the possibility of using car-following mod-els to design signal time for intersection has been confirmed.Furthermore,by using OVDD model,the signal timing design for signal intersection has been investigated.The simulation results show that using OVDD model,the number of vehicles going through the intersection is more than the ones in the observed situation in the same signal cycle.The designing scheme of the signal cycle time and split have been given.The results show that signal timing plan of signalized intersection has been improved by using OVDD model with the consideration of response sensitivity of the signal of the leading vehicle,starting speed of the leading vehicle as well as the following close-ness of vehicles.