Search

• Car following behavioral stochasticity has been analyzed based on two sets of field experimental trajectory data. • Wave travel time does not follow a consistent pattern. • Changing rate of wave travel time follows a mean reversion process. • A new stochastic Newell type model is proposed. • Stochastic factors exhibit common feature from different perspectives. This paper analyzes the car following behavioral stochasticity based on two sets of field experimental trajectory data by measuring the wave travel time series τ ˜ n (t) of vehicle n. The analysis shows that (i) No matter the speed of leading vehicle oscillates significantly or slightly, τ ˜ n (t) might change significantly; (ii) A follower's τ ˜ n (t) can vary from run to run even if the leader travels at the same stable speed; (iii) Sometimes, even if the leader's speed fluctuates significantly, the follower can keep a nearly constant value of τ ˜ n (t). The Augmented Dickey-Fuller test indicates that the time series ξ n (t) = d τ ˜ n (t) / d t follows a mean reversion process, no matter the oscillations fully developed or not. Based on the finding, a simple stochastic Newell model is proposed. The concave growth pattern of traffic oscillations has been derived analytically. Furthermore, simulation results demonstrate that the new model well captures both macroscopic characteristic of traffic flow evolution and microscopic characteristic of car following. [ABSTRACT FROM AUTHOR]

• A thorough analysis of the trajectories of the car-following experimental data supports the proposed instability mechanism. • The time series of acceleration and the speed difference exhibit striking similarity, and the speed difference plays a more important role on car-following dynamics than spacing. • The stochasticity follows a mean reversion process. • A heuristic analysis to explain why oscillation grows in a concave pattern was proposed. • A new hybrid, stochastic car-following model incorporating speed adaptation and stochasticity was proposed. Understanding the mechanisms responsible for the emergence and evolution of oscillations in traffic flow has been a subject of intensive research by numerous scholars. In our previous work, we proposed a new mechanism to explain the generation of traffic oscillations: traffic instability caused by the competition between speed adaptation and stochastic effects. By conducting a closer examination of car following data obtained in a 25-car platoon experiment, we discovered in this work that the speed difference plays a more important role on car-following dynamics than the spacing, and when its amplitude is small, the growth of oscillations is mainly determined by the stochastic factors that follow the mean reversion process; when its amplitude increases, the growth of the oscillations is determined by the competition between the stochastic factors and the speed difference. An explanation is then provided, based on the above findings, to why the speed variance in the oscillatory traffic grows in a concave way along the platoon. We also proposed a mode-switching stochastic car-following model that incorporates the speed adaptation and spacing indifference behavior of drivers, which captures the observed characteristics of oscillation and discharge rate. Finally, our sensitivity analysis shows that the influence on oscillation growth and discharge rate is small by the reaction delay but large by the indifference region boundary. [ABSTRACT FROM AUTHOR]

This paper proposes an improved cellular automaton traffic flow model based on the brake light model, taking into account that the desired time gap of vehicles is larger than one second. Although the hypothetical steady state of vehicles in the deterministic limit corresponds to a unique relationship between speeds and gaps in the proposed model, the traffic states of vehicles dynamically span a two-dimensional region in the plane of speed versus gap, due to the various randomizations. We show that our model reproduces (i) the free flow, synchronized flow, wide moving jams, as well as transitions among the three phases; (ii) the evolution features of disturbances and the spatiotemporal patterns in a car-following platoon; (iii) the empirical time series of traffic speed obtained from Next Generation Simulation data. Furthermore, the basic feature of time headway distribution is also qualitatively reproduced. Therefore, we argue that a model has the potential to reproduce the empirical and experimental features of traffic flow, provided that the traffic states can dynamically span a 2D speed-gap region. [ABSTRACT FROM AUTHOR]

This paper firstly shows that a recent model (Tian et al., Transpn. Res. B 71, 138–157, 2015) is not able to replicate well the concave growth pattern of traffic oscillations (i.e., the standard deviation of speed is a concave function of the vehicle number in the platoon) observed from car following experiments. We propose an improved model by introducing a safe speed and the logistic function for the randomization probability. Simulations show that the improved model can reproduce well the metastable state, the spatiotemporal patterns, and the phase transitions of traffic flow. Calibration and validation results show that the concave growth pattern of oscillations and the empirical detector data can be simulated with a quantitative agreement. [ABSTRACT FROM AUTHOR]

This paper firstly show that 2 Dimensional Intelligent Driver Model (Jiang et al., 2014) is not able to replicate the synchronized traffic flow. Then we propose an improved model by considering the difference between the driving behaviors at high speeds and that at low speeds, which is in the framework of three-phase traffic theory. Simulations show that the improved model can reproduce the phase transition from synchronized flow to wide moving jams, the spatiotemporal patterns of traffic flow induced by traffic bottleneck, and the concave growth pattern of traffic oscillations (i.e. the standard deviation of the velocities of vehicles increases in a concave/linear way along the platoon). Validating results show that the empirical time series of traffic speed obtained from Floating Car Data can be well simulated as well. [ABSTRACT FROM AUTHOR]

In driving situations, if drivers have advanced access to downstream traffic conditions ahead of time, they may modify their driving behavior accordingly. This proactive approach can help reduce frequent acceleration and deceleration of vehicles and minimize traffic congestion. This paper considers the average headway of downstream vehicles to predict the traffic conditions of downstream roads. Based on the full velocity difference (FVD) model, we propose a car-following model that incorporates spatial expectation effect (SEFVD). Firstly, a linear stability analysis is performed on the SEFVD model, and the results indicate that the critical value of the SEFVD model is relatively small, and the stabilization region increases gradually with the number of vehicles downstream. Secondly, employing perturbation amplitudes on a virtual circular road, a numerical simulation of the SEFVD model is carried out and compared with the FVD model. The results demonstrate that the SEFVD model has the ability to suppress the formation of traffic oscillations and stabilize vehicle formation more quickly. Simulation results for the 11 cars starting from a traffic signal indicate that the delay time and kinematic wave speed of the SEFVD model are within reasonable limits. To evaluate the changes in energy use, we also carried out energy consumption experiments. The experimental findings show that the SEFVD model uses less energy compared to the FVD model. Finally, the parameters of the models are calibrated and verified by using NGSIM data. The total calibration error of the SEFVD model (m = 2) decreases by 42.97% compared to the FVD model, and the total validation error decreases by 29.3%. The results show that the proposed model can fit the actual data well, and the fitting error is smaller. [ABSTRACT FROM AUTHOR]

• This study conducted a laboratory experiment in a mixed traffic with autonomous vehicles (AVs) and manually-driven vehicles (MVs) to uncover travel mode preferences. • The result showed that more than half of the travelers chose AVs. The result came close to Nash equilibrium, but did not reach social optimum. • Those who were provided more information will consider the cost difference between the two modes, which resulted in losses of both individual and social benefits. • Subsidies that eliminated the cost inequality can make the system more efficient. The effect of information on social welfare depended on whether providing subsidy or not. • Learning model simulation showed that the main factors affecting travelers' decisions were inertia, experiences, and inequality. An increase in autonomous vehicles (AVs) would result in a decline in traffic congestion; however, the travel cost associated with AVs is always higher than that of manually-driven vehicles (MV). This situation is interpreted as a so-called multi-player social dilemma. This study designed an economic experiment to investigate the effect of AVs on mode choice in mixed traffic flows. Participants were informed about the cost function in both modes and were asked to choose the travel mode for more than 60 rounds. In full information (FI) treatment, participants received information about the travel costs of the AV and MV modes at the end of each round. In the partial information (PI) treatment, participants received information only about the travel cost of the mode they chose. We found that participants were sensitive to cost differences in the FI treatments. Based on inequality aversion models, we proposed a perceived cost that could better explain the experimental equilibrium. A monetary reward was provided to encourage participants to take AVs and solve social dilemmas. The results demonstrated that the reward mechanism reduces traffic congestion and increases social benefits, especially in the FI treatment. Finally, a learning model that considers inertia and perceived cost is proposed to explain the decision-making process of the participants during the experiment. The findings have implications for traffic forecasting in the mixed flow of MVs and AVs and provide insights and policy suggestions for AV management. [ABSTRACT FROM AUTHOR]

This paper investigates the revenue-neutral tradable credit charge and reward scheme without initial credit allocations that can reassign network traffic flow patterns to optimize congestion and emissions. First, we prove the existence of the proposed schemes and further decentralize the minimum emission flow pattern to user equilibrium. Moreover, we design the solving method of the proposed credit scheme for minimum emission problem. Second, we investigate the revenue-neutral tradable credit charge and reward scheme without initial credit allocations for bi-objectives to obtain the Pareto system optimum flow patterns of congestion and emissions; and present the corresponding solutions are located in the polyhedron constituted by some inequalities and equalities system. Last, numerical example based on a simple traffic network is adopted to obtain the proposed credit schemes and verify they are revenue-neutral. [ABSTRACT FROM AUTHOR]

• Transfer the insights on the rather special situation/observation of moving bottleneck to the ubiquitous case of fixed bottleneck. • The role of stochasticity: Intra-driver stochasticity is needed to describe the phenomena. • The nature of the goodness-of-fit (microscopic or macroscopic) should reflect the nature of the observations to be calibrated. • The high-quality data in itself can be used to test traffic flow models and theories. This paper aims to address a simple but fundamental question, how the traffic oscillations grow along the road. Firstly, we conduct an empirical study on the growth pattern of traffic oscillations on US-101 Freeway and German-A5 Freeway. Then we perform an experiment to study the growth pattern of traffic oscillations on a single lane with a fixed bottleneck of speed limit. Both empirical and experimental results show that traffic oscillations grow in a concave way along the road. This finding is consistent with the previous one that traffic oscillations grow concavely along the platoon following a slower leader taking on the role of a moving bottleneck, which is not to be expected a priori for a fixed bottleneck. Finally, we use the finding to test three typical car-following models. The test indicates that whereas the intelligent driver model (IDM) fails to reproduce the observed growth pattern, the 2D-IDM and the stochastic speed adaptation model could. These findings are expected to improve our understanding of the role of stochasticity in car following, and the high-quality data in itself can be used to test traffic flow models and theories. [ABSTRACT FROM AUTHOR]

At a non-signalized intersection, some vehicles violate the traffic rules to pass the intersection as soon as possible. These behaviors may cause many traffic conflicts even traffic accidents. In this paper, a simulation model is proposed to research the effects of these behaviors at a non-signalized intersection. Vehicle’s movement is simulated by the cellular automaton (CA) model. The game theory is introduced for simulating the intersection dynamics. Two types of driver participate the game process: cooperator (C) and defector (D). The cooperator obey the traffic rules, but the defector does not. A transition process may occur when the cooperator is waiting before the intersection. The critical value of waiting time follows the Weibull distribution. One transition regime is found in the phase diagram. The simulation results illustrate the applicability of the proposed model and reveal a number of interesting insights into the intersection management, including that the existence of defectors is benefit for the capacity of intersection, but also reduce the safety of intersection. [ABSTRACT FROM AUTHOR]

• A Multi-layer Graph Mask Q-Learning algorithm for traffic signals is proposed. • This study constructs a new upper and lower graph structure to extract geometric and spatial features of intersections. • This article introduces an action masking mechanism, enabling the model to be applicable to intersections of arbitrary structures when utilizing any phase as the action. • Model tests were conducted on synthetic road networks with different traffic volumes and two real urban road networks, and the test results showed that the algorithm has strong generalization ability and scalability. Due to its capability in handling complex urban intersection environments, deep reinforcement learning (DRL) has been widely applied in Adaptive Traffic Signal Control (ATSC). However, most existing algorithms are designed for specific road networks or traffic conditions, making it difficult to transfer them to new environments. Moreover, current graph-based algorithms do not fully capture the geometric and spatial features of intersections, leading to incomplete embedding of the agent's environment depiction. Additionally, the actions adopted by these algorithms are inherently based on fixed-cycle phases, limiting the flexibility of traffic signal control. To address the aforementioned issues, this paper proposes a Multi-layer Graph Mask Q-Learning (MGMQ) algorithm for multi-intersection ATSC to optimize traffic and reduce delay. Unlike previous graph-based algorithms, this paper introduces a method for computing multi-layer graphs, dividing the traffic environment into upper-level traffic network-layer graphs and lower-level intersection-layer graphs, and employs the graph attention algorithm and an improved GraphSAGE algorithm for computation. This method not only enables the generation of embedded state for intersections that include geometric and spatial features, but also allows the algorithm to adapt to different traffic conditions and road networks. Additionally, we introduce an action masking mechanism, allowing this algorithm can be adapted to different action spaces. As a result, the algorithm uses arbitrary signal phases as actions to achieve flexible traffic flow control, and can be directly applied to intersections with arbitrary geometry. The final test results demonstrate that a model trained solely on synthetic road networks can be directly transferred to other synthetic network configurations or real-world urban road networks, outperforming current state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

Signalized intersection is the primary bottleneck in an urban traffic system. This paper developed a driving guidance strategy at isolated signalized intersection by considering the pre-stop line from the perspective of cooperation driving, speed guidance, and optimization control method. First, the pre-stop line and pre-signal light are set up to regulate and guide the driving trajectory of Human Vehicles (HVs). Then, the cooperative driving strategy is proposed for heterogeneous traffic flow composed of Connected Automatic Vehicles (CAVs) and HVs. The effect of CAVs market penetration rate on the signalized intersection capacity is also analysed. Finally, based on Gauss Pseudospectral OPtimization software (GPOPs), a multi-stage optimal control model is proposed according to the variations of vehicle trajectories at the pre-stop line. In the numerical experiments, a reasonable pre-signal timing scheme and position of pre-stop line are obtained by optimizing the sub-stages, and then the optimal fuel consumption trajectory of CAVs is obtained by overall optimization. Both theoretical analysis and numerical simulation show that the proposed multi-stage optimal control model can effectively realize the collaborative optimization of capacity and vehicle fuel consumption at signalized intersections under the driving guide strategy with pre-stop line. [ABSTRACT FROM AUTHOR]

As the important spatiotemporal state of traffic flow discovered by Kerner's three-phase traffic theory, the synchronized traffic flow describes a new traffic phase in congested traffic. However, until now most models within the standard traffic theories cannot reproduce it. The average space gap model (ASGM) is a simple cellular automaton model aimed to reproduce various empirical findings discovered by Kerner's three-phase traffic theory by incorporating the influence of the multi-anticipative effect on vehicle's deceleration. However, this paper shows that the simulated synchronized flow by ASGM is not consistent with the reality. To this end, the Multi-Anticipative Model (MAM) based on ASGM is proposed, which describes the influence of the multi-anticipative effect on both the accelerations and the decelerations. Simulations indicate that the empirical consistent synchronized flow and related congested patterns can be well reproduced by MAM. Moreover, MAM can reproduce the speed drop in the car following vehicle platoons reported by the empirical observations. Generally, MAM indicates that the multi-anticipative effect can shed light on the understanding and capturing the complex characteristics of traffic flow especially reported by Kerner's three-phase traffic theory. • A new cellular automaton with multi-anticipative effect has been developed based on ASGM. • The synchronized flow with related empirical findings in Kerner's three-phase traffic theory can be well reproduced. • The empirical phenomenon of sudden speed drop can be captured in car following platoons. [ABSTRACT FROM AUTHOR]