Car-following Research Articles

A State-of-the-Art Review of Car-Following Models with Particular Considerations of Heavy Vehicles

Car-following (CF) models are fundamental in the replication of traffic flow and thus they have received considerable attention. This attention needs to be reflected upon at particular points in time. CF models are in a continuous state of improvement due to their significant role in traffic micro-simulations, intelligent transportation systems and safety engineering models. This paper presents a review of existing CF models. It classifies them into classic and artificial intelligence models. It discusses the capability of the models and potential limitations that need to be considered in their improvement. This paper also reviews the studies investigating the impacts of heavy vehicles in traffic stream and on CF behaviour. The findings of the study provide promising directions for future research and suggest revisiting the existing models to accommodate different behaviours of drivers in heterogeneous traffic, in particular, heavy vehicles in traffic.

Incorporating human-factors in car-following models: A review of recent developments and research needs

Over the past decades there has been a considerable development in the modeling of car-following (CF) behavior as a result of research undertaken by both traffic engineers and traffic psychologists. While traffic engineers seek to understand the behavior of a traffic stream, traffic psychologists seek to describe the human abilities and errors involved in the driving process. This paper provides a comprehensive review of these two research streams.It is necessary to consider human-factors in CF modeling for a more realistic representation of CF behavior in complex driving situations (for example, in traffic breakdowns, crash-prone situations, and adverse weather conditions) to improve traffic safety and to better understand widely-reported puzzling traffic flow phenomena, such as capacity drop, stop-and-go oscillations, and traffic hysteresis. While there are some excellent reviews of CF models available in the literature, none of these specifically focuses on the human factors in these models.This paper addresses this gap by reviewing the available literature with a specific focus on the latest advances in car-following models from both the engineering and human behavior points of view. In so doing, it analyses the benefits and limitations of various models and highlights future research needs in the area.

The root locus method: application to linear stability analysis and design of cooperative car-following models

A global framework for linear stability analyses of traffic models, based on the dispersion relation root locus method, is presented and is applied taking the example of a broad class of car-following (CF) models. This approach is able to analyse all aspects of the dynamics: long waves and short wave behaviours, phase velocities and stability features. The methodology is applied to investigate the potential benefits of connected vehicles, i.e. V2V communication enabling a vehicle to send and receive information to and from surrounding vehicles. We choose to focus on the design of the coefficients of cooperation which weights the information from downstream vehicles. The coefficients tuning is performed and different ways of implementing an efficient cooperative strategy are discussed. Hence, this paper brings design methods in order to obtain robust stability of traffic models, with application on cooperative CF models.

Extending the Cell Transmission Model to Multiple Lanes and Lane-Changing

Macroscopic or flow-based dynamic traffic assignment (DTA) models normally treat traffic in each direction on a roadway as a single lane and, since they do not consider multiple lanes, they can not consider lane-changing behaviour. To investigate how the results may be affected by explicitly considering lanes and lane changing, we consider a road link that consists of two adjacent homogeneous lanes. We assume that traffic entering each lane already knows in which lane it wishes to exit at the end of the link, whether it wishes to exit in the same lane or in the other lane. We model the traffic flows in each lane using a cell transmission model but adapt it to allow for traffic moving from cells in one lane to cells in the other lane. The CTM is used because it handles the modelling of queues and their spillback in an intuitive and widely accepted manner, and our extensions of it allow congestion in one lane to spill back into adjacent lanes. In particular, we investigate how lane-changing and congestion are affected by varying the assumptions concerning two key behavioural parameters, namely the locations at which drivers wish to change lanes and the vehicle spacing needed for lane changing (gap acceptance) as compared to the spacing needed when staying in the same lane (car following). We conclude that there are many situations where modelling a link as a single lane will give a poor approximation to the underlying multi-lane behaviour, or be unable to capture issues of interest, and for those situations multi-lane modelling is appropriate.

Exploring a Local Linear Model Tree Approach to Car‐Following

AbstractBecause car‐following (CF) models are fundamental to replicating traffic flow they have received considerable attention over the last 50 years. They are in a continuous state of improvement due to their significant role in traffic microsimulations, intelligent transportation systems, and safety engineering models. This article uses the local linear model tree (LOLIMOT) approach to model driver's CF behavior to incorporate human perceptual imperfections into a CF model. This model defines some localities in the input space. These localities are fuzzy and have overlaps with each other. Specific models for each of the localities are then defined and combined in a fuzzy manner to predict the final output. The model was developed using real world dynamic data sets. Three different data sets were used for training, testing, and validating the model. The performance of the model was compared to a number of existing CF models. The results showed very close agreement between the real data and the LOLIMOT outputs.

Development of Two-Lane Car-Following Model to Generate More Realistic Headway Behavior

2차로 양방향 도로(이하 2차로 도로)의 주요한 특징은 화물차량 등의 저속차량으로 인하여 발생하는 차량군과 차량추월 행태이다. 차량추종과 차량추월모형이 결합된 2차로 도로 교통류 모형을 개발하기 위해서는 2차로 도로에 적합한 차량추종모형의 개발이 필수적이다. 2차로 도로에서 고속차량은 저속차량의 후미에서 차량추종을 수행함과 더불어 추월을 수행하게 되며, 이 과정에서 짧은 차간거리가 발생하게 된다. 또한 저속차량을 추종하는 차량은 대항교통류의 차두시간분포 내에서 적정 공간을 이용하여 차량추월을 결정하게 된다. 따라서 2차로 차량추종모형은 짧은 차간거리에서 차량주행과 더불어 차두시간분포를 설명해야 한다. 추가적으로 국내의 2차로 도로규모를 고려하면 대규모 도로망을 모의실험할 수 있는 모형이 필요한 실정이다. 국내의 경우, 2차로 차량추종모형에 대한 연구는 거의 보고되고 있지 않다. 본 연구에서는 대규모 가로망에 적용이 가능하면서 2차로 도로의 차량추종행태를 보다 현실적으로 구현할 수 있는 모형을 개발하였다. 개발된 모형의 실험적 평가 결과, 혼잡 교통류의 특성 중 하나인 가다서다(stop-and-go) 현상과 차두시간분포를 효과적으로 설명하는 것으로 나타났다. 따라서 본 연구에서 제시된 차량추종모형이 차량추월모형과 결합된 경우, 보다 현실적으로 2차로 도로 교통류를 설명할 수 있을 것으로 판단된다. The key characteristics of two-lane-and-two-way traffic flow are platoon and overtaking caused by low-speed vehicle such as truck. In order to develop two-way traffic flow model comprised of CF(car-following) and overtaking model, it is essential to develop a car-following model which is suitable to two-way traffic flow. Short distance between vehicles is caused when a high-speed vehicle tailgates and overtakes foregoing low-speed vehicle on two-way road system. And a vehicle following low-speed vehicle decides to overtake the front low-speed vehicle using suitable space within the headway distribution of opposite traffic flow. For this reason, a two-way CF model should describes not only running within short gap but also headway distribution. Additionally considering domestic two-way-road size, there is a on-going need for large-network simulation, but there are few studies for two-way CF model. In this paper, a two-way CA model is developed, which explains two-way CF behavior more realistic and can be applied for large road network. The experimental results show that the developed model mimics stop-and-go phenomenon, one of features of congested traffic flow, and efficiently generates the distribution of headway. When the CF model is integrated with overtaking model, it is, therefore, expected that two-way traffic flow can be explained more realistically than before.

Risk illusions in car following: Is a smaller headway always perceived as more dangerous?

Evaluating risk perception in car following is important for understanding driver behavior in close following. A new model of risk perception in car following is proposed in this study, in which the risk perceived by the driver of a following car is divided into two parts: the tailgating risk involved in following the leading vehicle, which is relatively stable from the driver’s viewpoint, and the intuitive risk perception arising from the remaining field of view of the surrounding environment, with oncoming vehicles rushing toward the driver or stationary objects being passed at a relatively high speed. We proposed that the driver does not always think that the leading vehicle is dangerous; on the contrary, a driver may feel safer when he/she gets closer to the leading vehicle if he/she perceives obvious risks from the environment. A driving-simulator experiment was designed to validate the proposed model, in which environmental risk was increased by adding a number of cars travelling in the opposite direction in the adjacent lane. Adjustments of the distance headway by 48 subjects were recorded. The results provided positive support for the proposed model to a certain extent. It was found that when vehicles came in the opposite direction, drivers showed a tendency to decrease their distance headways, but this tendency actually increased the risk. Additionally, data were collected on comfortable and minimum safe distances that the drivers kept, and a significant interaction effect of the speed and the size of the leading vehicle on distance headway was observed.

Freeway traffic oscillations: Microscopic analysis of formations and propagations using Wavelet Transform

In this paper we identify the origins of stop-and-go (or slow-and-go) driving and measure microscopic features of their propagations by analyzing vehicle trajectories via Wavelet Transform. Based on 53 oscillation cases analyzed, we find that oscillations can be originated by either lane-changing maneuvers (LCMs) or car-following (CF) behavior. LCMs were predominantly responsible for oscillation formations in the absence of considerable horizontal or vertical curves, whereas oscillations formed spontaneously near roadside work on an uphill segment. Regardless of the trigger, the features of oscillation propagations were similar in terms of propagation speed, oscillation duration, and amplitude. All observed cases initially exhibited a precursor phase, in which slow-and-go motions were localized. Some of them eventually transitioned into a well-developed phase, in which oscillations propagated upstream in queue. LCMs were primarily responsible for the transition, although some transitions occurred without LCMs. Our findings also suggest that an oscillation has a regressive effect on car-following behavior: a deceleration wave of an oscillation affects a timid driver (characterized by larger response time and/or minimum spacing) to become less timid and an aggressive driver less aggressive, although this change may be short-lived. An extended framework of Newell’s CF model is able to describe the regressive effect with two additional parameters with reasonable accuracy, as verified using vehicle trajectory data.

Anticipation in car following: Modelling expectations for the development of driver assistance systems

The EU-project ISi-PADAS aims to integrate driver models into the development process of driver assistance systems. Driver modeling is used to predict whether the driver is likely to not react adequately to the driving situation and whether the PADAS should intervene. It is based on empirical investigations. An accident analysis showed that in many rear end accidents, driver's inadequate expectations about the lead car may be a major factor in causing the accident. A driving simulator study was conducted to examine this hypothesis. The behavior of a lead car was varied in order to create different expectations. Attention allocation and driver behavior was examined. The results are partially consistent with the hypotheses

Responses to deceleration during car following: Roles of optic flow, warnings, expectations, and interruptions.

More than 25% of accidents are rear-end collisions. It is essential to identify the factors that contribute to such collisions. One such factor is a driver's ability to respond to the deceleration of the car ahead. In Experiment 1, we measured effects of optic flow information and discrete visual and auditory warnings (brake lights, tones) on responses to deceleration during car following. With computer simulations of car-following scenes, university students pressed a button when the lead car decelerated. Both classes of information affected responses. Observers relied on discrete warnings when optic flow information was relatively less effective as determined by the lead car's headway and deceleration rate. This is consistent with DeLucia's (2008) conceptual framework of space perception that emphasized the importance of viewing distance and motion (and task). In Experiment 2, we measured responses to deceleration after a visual interruption. Scenes were designed to tease apart the role of expectations and optic flow. Responses mostly were consistent with optic flow information presented after the interruption rather than with putative mental expectations that were set up by the lead car's motion prior to the interruption. The theoretical implication of the present results is that responses to deceleration are based on multiple sources of information, including optical size, optical expansion rate and tau, and discrete warnings that are independent of optic flow. The practical implication is that in-vehicle collision-avoidance warning systems may be more useful when optic flow is less effective (e.g., slow deceleration rates), implicating a role for adaptive collision-warning systems.

The Effects of Driving Situation on Sleepiness Indicators after Sleep Loss: A Driving Simulator Study

Almost all studies of sleepy driving are carried out in driving simulators and with monotonous road conditions (no interaction with other cars). The present study investigated indicators of sleepy driving in a more challenging scenario after a night awake. 17 participants drove a high fidelity moving base driving simulator experiment while sleepiness was monitored physiologically and behaviourally. Short periods of situations of free driving (no other vehicles) alternated with short periods of following another vehicle (car following) with and without the possibility to overtake. The result showed that a night of prior sleep loss increased sleepiness levels at the wheel (eye closure duration and lateral variability) compared to after a night of normal sleep. Blink duration while overtaking was significantly lower compared to the other situations, it was at the same level as after night sleep. Speed when passing a stopped school bus was not significantly affected by sleepiness. However the warning caused a more rapid reduction of speed. In conclusion, a moderately challenging driving contest did not affect sleepiness indicators, but a very challenging one did so (overtaking). This suggests that it is important to monitor the driving situation in field operational tests of sleepy driving.

Detection of Emergency Vehicles: Driver Responses to Advance Warning in a Driving Simulator

This research evaluated the effects of an advance warning device (AWD) on the safety of driver interactions with emergency vehicles (EVs). The AWD was intended to provide drivers with advance warning of an approaching on-call EV via visual and auditory warnings when the EV was within a 300- to 400-m radius. Research suggests that drivers can experience difficulty accurately detecting the distance and direction of approaching on-call EV. In-vehicle technology has not previously been explored as a means of overcoming the limitations of existing EV lights and sirens and improving driver detection of EV. An experimental study using an advanced driving simulator examined the effects of the AWD on driving performance in a range of circumstances in which real-world EV crashes and near-misses commonly occur. Each event contained a combination of scenario type (adjacent lane, turning across, car following) and warning condition (control, standard, advance). Data from 22 participants were collected, including measures of speed, braking, and visual scanning. For adjacent-lane and turning-across events, the AWD was associated primarily with reductions in mean speed. The AWD resulted in an earlier lane change to clear a path for the EV in the car-following event. The reduction in speed observed was a positive finding, given the relationship between impact speed and injury severity. Response priming emerged as the mechanism underpinning these effects. Response priming may result in safety benefits in other settings when an advisory warning is presented before the threat can be perceived.

Repeated‐dose effects of mequitazine, cetirizine and dexchlorpheniramine on driving and psychomotor performance

Previous studies have demonstrated that the antihistamines mequitazine, cetirizine and dexchlorpheniramine produce mild sedation after single doses. It is unknown, however, whether acute sedation persists after repeated dosing. Therefore, this study assessed the effects of repeated dosing of these antihistamines on driving and psychomotor performance. Sixteen healthy volunteers were treated with mequitazine 10 mg q.a.m., cetirizine 10 mg q.a.m., dexchlorpheniramine Repetab 6 mg b.i.d. and placebo for four separate 8-day periods. Drug effects were assessed on days 1 and 8 using on-the-road driving tests (highway driving and car following), psychomotor tests (tracking and divided attention) and subjective questionnaires. Dexchlorpheniramine and mequitazine significantly impaired driving performance on the highway driving test on the first day; dexchlorpheniramine increased Standard Deviation of Lateral Position by 2 cm [95% confidence interval (CI) 0.5, 3.8] and mequitazine by 2.5 cm (CI 1.0, 4.3). These effects on driving performance disappeared after 8 days of treatment. No effect of treatment was found on car following, tracking and divided attention. Although subjective ratings confirmed that subjects knew their driving had been impaired in the mequitazine and dexchlorpheniramine condition after completion of the highway driving test on day 1, they did not expect their driving to be affected before the start of the test. Cetirizine did not impair performance on any of the tests. Single doses of mequitazine 10 mg and dexchlorpheniramine Repetab 6 mg cause mild driving impairment. However, when taken over several days, the impairing effect wears off, possibly as a result of tolerance.

A high fidelity traffic simulation model based on cellular automata and car-following concepts

A high fidelity cell based traffic simulation model (CELLSIM) has been developed for simulation of high volume of traffic at the regional level. Straightforward algorithms and efficient use of computational resources make the model suitable for real time traffic simulation. The model formulation uses concepts of cellular automata (CA) and car-following (CF) models, but is more detailed than CA models and has realistic acceleration and deceleration models for vehicles. A simple dual-regime constant acceleration model has been used that requires minimal calculation compared to detailed acceleration models used in CF models. CELLSIM is simpler than most CF models; a simplified car-following logic has been developed using preferred time headway. Like CA models, integer values are used to make the model run faster. Space is discretized in small intervals and a new concept of percent space occupancy (SOC) is used to measure traffic congestion. CELLSIM performs well in congested and non-congested traffic conditions. It has been validated comprehensively at the macroscopic and microscopic levels using two sets of field data. Comparison of field data and CELLSIM for trajectories, average speed, density and volume show very close agreement. Statistical comparison of macroscopic parameters with other CF models indicates that CELLSIM performs as good as detailed CF models. Stability analyses conducted using mild and severe disturbances indicate that CELLSIM performs well under both conditions.

The perception of lead vehicle movement in darkness

Three experimental studies are described on the detection of longitudinal lead-vehicle movement in darkness. Experiment I, performed in the laboratory, isolated the relative horizontal angular motion of the lead-vehicle's taillights as a cue to the detection of relative longitudinal movement. Movement thresholds were determined as a function of initial headway, exposure duration, direction of the movement, and the presence of a background. All these factors except the last had significant effects on the movement threshold. Experiment II, also a laboratory-experiment, isolated changes in apparent size or brightness of the taillights as a cue. Movement thresholds were again found to be a function of the variables investigated in Experiment I. However, thresholds now were much higher than under the isolated angular cue. It was therefore concluded that the relative horizontal angular motion of the lead-vehicle's taillights is the priminent cue in the detection of relative vehicle motion in depth. Experiment III substantiated this conclusion by confirming the parametric outcomes of Experiment I under realistic night-driving conditions. The practical relevance of the estimates made from the data of two important temporal parameters in car following, viz. time until collision and free time after detection.