The future of Autonomous Vehicles (AVs) will experience a breakthrough when collective intelligence is employed through decentralized cooperative systems. A system capable of controlling all AVs ...crossing urban intersections, considering the state of all vehicles and users, will be able to improve vehicular flow and end accidents. This type of system is known as Autonomous Intersection Management (AIM). AIM has been discussed in different articles, but most of them have not considered the communication latency between the AV and the Intersection Manager (IM). Due to the lack of works studying the impact that the communication network can have on the decentralized control of AVs by AIMs, this paper presents a novel latency-aware deep reinforcement learning-based AIM for the 5G communication network, called AIM5LA. AIM5LA is the first AIM that considers the inherent latency of the 5G communication network to adapt the control of AVs using Multi-Agent Deep Reinforcement Learning (MADRL), thus obtaining a robust and resilient multi-agent control policy. Beyond considering the latency history experienced, AIM5LA predicts future latency behavior to provide enhanced security and improve traffic flow. The results demonstrate huge safety improvements compared to other AIMs, eliminating collisions (on average from 27 to 0). Further, AIM5LA provides comparable results in other metrics, such as travel time and intersection waiting time, while guaranteeing to be collision-free, unlike the other AIMs. Finally, compared to other traffic light-based control systems, AIM5LA can reduce waiting time by more than 99% and time loss by more than 95%.
•Subjects initiated braking significant earlier in the field than in driving simulator.•Subjects in driving simulator showed significantly higher speed and braked distinctly more intensively than in ...the field.•In driving simulator, the subjects appeared to have reduced risk awareness.•Caution is required when using a driving simulator for designing intersection assistance systems.
Introduction: The majority of accidents involving personal injury occur in urban intersections. As human error is the primary cause of these accidents, it seems reasonable to support drivers by intersection assistants. Studies to evaluate such systems are usually performed in driving simulators. However, this implies driver behaviour validity, which cannot always be taken for granted due to the artificial nature of simulated environments. The paper pays special attention to the validity of braking behaviour in urban intersections. In particular, the effect of the test scenario and selected driving simulator design parameters is considered. Method: The Time to Intersection (TTI) time interval between field and simulator study was evaluated at five simple urban intersections. Results: Independently of the type of intersection and the parameter configuration of the simulator, subjects initiated braking in preparation for a turning manoeuvre significantly earlier in the field than in the driving simulator. In particular, the results of both tests differed more at intersections where the driver did not have right of way than at intersections with other layouts, confirming the impact of the test scenario. Some differences were also found when the driving simulator parameter set was varied. Conclusions: The braking behaviour near urban intersections differs between real and simulated experimental environments. From these results, we conclude that caution is required when designing intersection assistance systems based solely on the braking behaviour of subjects in simulated environments. The test scenario and the design parameters must be considered as factors which influence the results. Practical applications: Depending on the current traffic situation and interactions, multiple different kinds of accident can occur at the considered types of intersection, meaning that closely investigating the driver behaviour validity at these junctions is highly important for further intersection assistance systems. Validity of TTI is for example important when determining the best choice of warning interval for a turning assistance system.
Vehicles are always in an alternating state of going and stopping at oversaturated signalized intersections, which not only reduces the operational efficiency of the intersections but also greatly ...increases the average fuel consumption of vehicles. The purpose of this study is to propose a fuel consumption analysis method that can be adapted to the oversaturated signalized intersections. Considering queue effects, this paper proposed a way to define the judging criteria and the classification for the oversaturated state based on the intersection spacing, traffic demand, and the intersection capacity. The theoretical models were constructed for estimating vehicular fuel consumption towards two different oversaturated states and with/without automatic start and stop systems respectively, which was dependent on two main parameters of the delay and the stop that were confirmed by the definite number theory. The results show that the degree of saturation of the approach and the automatic start and stop systems have a great influence on vehicular fuel consumption at oversaturated signalized intersections, and the shorter spacing is expected for the oversaturated state II because of more deceleration and acceleration maneuvers.
•A vehicular fuel consumption estimate model is proposed for the oversaturated signalized intersection.•The influence of the degree of saturation on fuel consumption is studied.•The influence of intersection spacing on fuel consumption is analyzed.
•Goal is to use connected vehicle technology to better operate intersections.•Algorithms developed to minimize total delay or number of stops holistically.•Algorithms tested with simulation for ...different demand values and demand ratios.•Tests show significant delay reduction with increasing penetration rate.•Tests show slight delay reduction if vehicles controlled autonomously.
Information from connected vehicles, such as the position and speed of individual vehicles, can be used to optimize traffic operations at an intersection. This paper proposes such an algorithm for two one-way-streets assuming that only a certain percentage of cars are equipped with this technology. The algorithm enumerates different sequences of cars discharging from the intersection to minimize the objective function. Benefits of platooning (multiple cars consecutively discharging from a queue) and signal flexibility (adaptability to demand) are also considered. The goal is to gain insights about the value (in terms of delay savings) of using connected vehicle technology for intersection control.
Simulations are conducted for different total demand values and demand ratios to understand the effects of changing the minimum green time at the signal and the penetration rate of connected cars. Using autonomous vehicle control systems, the signal could rapidly change the direction of priority without relying on the reaction of drivers. However, without this technology a minimum green time is necessary. The results of the simulations show that a minimum green time increases the delay only for the low and balanced demand scenarios. Therefore, the value of using cars with autonomous vehicle control can only be seen at intersections with this kind of demand patterns, and could result in up to 7% decrease in delay. On the other hand, using information from connected vehicles to better adapt the traffic signal has proven to be indeed very valuable. Increases in the penetration rate from 0% up to 60% can significantly reduce the average delay (in low demand scenarios a decrease in delay of up to 60% can be observed). That being said, after a penetration rate of 60%, while the delays continue to decrease, the rate of reduction decreases and the marginal value of information from communication technologies diminishes. Overall, it is observed that connected vehicle technology could significantly improve the operation of traffic at signalized intersections, at least under the proposed algorithm.
•Alternative designs are proposed for signalized intersections with high turn demands.•The designs can reduce the number of protected turn phases needed each cycle.•Tests find that the designs can ...significantly increase intersection capacities.•Automated and connected vehicle technologies may enhance the likelihood of deployment.
Protecting left-turn movements on all four approaches to a signalized intersection conventionally requires a minimum of two extra phases per cycle. Losses in capacity often result. Various intersection designs have been proposed to combat those losses. Perhaps the best known of these designs is the continuous flow intersection. It features specially-configured approach lanes and mid-block pre-signals. These enable opposing left-turn and through-moving vehicles to proceed through the intersection free of conflicts, and without need for additional protected-turn phases.
The present paper offers an alternative design for four-way intersections, which to our knowledge has not previously been proposed. The design furnishes lower capacities than do continuous flow intersections, but spares the expense of having to reconfigure approach lanes. Pre-signals store queues and route traffic through the intersection much as in a continuous flow design. The distinguishing feature of the alternative is that it enables all four turn movements to be served during a single protected phase. Only one additional phase is therefore required per cycle. Numerical analysis shows that the plan regularly achieves higher intersection capacities than do conventional designs. Capacity gains as high as 80% are predicted. The proposed design is rather mentally taxing to drivers. Hence, opportunities for deploying the design in real settings are discussed with an eye toward the more connected and automated driving expected in the future.
•Sequential movements of CAVs are modelled as multi-agent Markov decision processes.•A decentralized coordination multi-agent learning approach (DCL-AIM) is proposed.•DCL-AIM explicitly identifies ...and dynamically adapts agent coordination needs.•Effectiveness of DCL-AIM is demonstrated through extensive simulation scenarios.•DCL-AIM outperforms the benchmarks: FCFS, LQF and fixed signal control policies.
Conventional intersection managements, such as signalized intersections, may not necessarily be the optimal strategies when it comes to connected and automated vehicles (CAVs) environment. Autonomous intersection management (AIM) is tailored for CAVs aiming at replacing the conventional traffic control strategies. In this work, using the communication and computation technologies of CAVs, the sequential movements of vehicles through intersections are modelled as multi-agent Markov decision processes (MAMDPs) in which vehicle agents cooperate to minimize intersection delay with collision-free constraints. To handle the huge dimension scale incurred by the nature of multi-agent decision making problems, the state space of CAVs are decomposed into independent part and coordinated part by exploiting the structural properties of the AIM problem, and a decentralized coordination multi-agent learning approach (DCL-AIM) is proposed to solve the problem efficiently by exploiting both global and localized agent coordination needs in AIM. The main feature of the proposed approach is to explicitly identify and dynamically adapt agent coordination needs during the learning process so that the curse of dimensionality and environment nonstationarity problems in multi-agent learning can be alleviated.
The effectiveness of the proposed method is demonstrated under a variety of traffic conditions. The comparison analysis is performed between DCL-AIM and the First-Come-First-Serve based AIM (FCFS-AIM), with Longest-Queue-First (LQF-AIM) policy and the signal control based on the Webster’s method (Signal) as benchmarks. Experimental results show that the sequential decisions from DCL-AIM outperform the other control policies.
•A backpressure approach built on macroscopic traffic theory.•A decentralized approach suitable for implementation in large urban networks.•Theoretical proof of network-wide traffic stability ...properties.•Very low computational costs.•Extensive experiments and comparisons that highlight the superiority of the proposed approach.
Decentralized intersection control techniques have received recent attention in the literature as means to overcome scalability issues associated with network-wide intersection control. Chief among these techniques are backpressure (BP) control algorithms, which were originally developed of for large wireless networks. In addition to being light-weight computationally, they come with guarantees of performance at the network level, specifically in terms of network-wide stability. The dynamics in backpressure control are represented using networks of point queues and this also applies to all of the applications to traffic control. As such, BP in traffic fail to capture the spatial distribution of vehicles along the intersection links and, consequently, spill-back dynamics.
This paper derives a position weighted backpressure (PWBP) control policy for network traffic applying continuum modeling principles of traffic dynamics and thus capture the spatial distribution of vehicles along network roads and spill-back dynamics. PWBP inherits the computational advantages of traditional BP. To prove stability of PWBP, (i) a Lyapunov functional that captures the spatial distribution of vehicles is developed; (ii) the capacity region of the network is formally defined in the context of macroscopic network traffic; and (iii) it is proved, when exogenous arrival rates are within the capacity region, that PWBP control is network stabilizing. We conduct comparisons against a real-world adaptive control implementation for an isolated intersection. Comparisons are also performed against other BP approaches in addition to optimized fixed timing control at the network level. These experiments demonstrate the superiority of PWBP over the other control policies in terms of capacity region, network-wide delay, congestion propagation speed, recoverability from heavy congestion (outside of the capacity region), and response to incidents.
Intersection management is one of the most challenging problems within the transport system. Traffic light-based methods have been efficient but are not able to deal with the growing mobility and ...social challenges. On the other hand, the advancements of automation and communications have enabled cooperative intersection management, where road users, infrastructure, and traffic control centers are able to communicate and coordinate the traffic safely and efficiently. Major techniques and solutions for cooperative intersections are surveyed in this paper for both signalized and nonsignalized intersections, whereas focuses are put on the latter. Cooperative methods, including time slots and space reservation, trajectory planning, and virtual traffic lights, are discussed in detail. Vehicle collision warning and avoidance methods are discussed to deal with uncertainties. Concerning vulnerable road users, pedestrian collision avoidance methods are discussed. In addition, an introduction to major projects related to cooperative intersection management is presented. A further discussion of the presented works is given with highlights of future research topics. This paper serves as a comprehensive survey of the field, aiming at stimulating new methods and accelerating the advancement of automated and cooperative intersections.
Fluid flow tests were conducted on four kinds of fracture intersections, namely straight fracture intersection, buckling fracture intersection, crossing fracture intersection and furcating fracture ...intersection. Corresponding numerical simulations were performed by solving the Navier-Stokes equations. The resultant flow force perpendicular to fracture walls were derived. The results show that at high hydraulic gradients the increased normal resultant force perpendicular to fracture walls will change the original flow direction and consequently cause the formation of eddies which triggers flow nonlinearity. Greater buckling angle and smaller crossing angle would result in the onset of nonlinear flow at a higher critical hydraulic gradient for the cases of buckling and crossing fracture intersections. The critical hydraulic gradient for the case of furcating fracture intersection depends on the absolute angle between furcating flow and the original flow direction. Based on the analyses, three nonlinear models for coefficient B in the Forchheimer law corresponding to three cases of fracture intersections are proposed, which are related to the fracture intersection angle and hydraulic aperture. These models are further validated using flow experimental data from a real fracture network created in a rock specimen with the corresponding fracture model constructed using the computed tomography (CT) method.
Cooperation of connected vehicles is a promising approach for autonomous intersection control. This article presents a systematic approach to the cooperation of connected vehicles at unsignalized ...intersections without global coordination. A task-area partition framework is proposed to decompose the mission of cooperative passing into three main tasks, i.e., vehicle state observation, arriving time optimization, and trajectory tracking control. To accomplish these tasks, a distributed observation algorithm is introduced to achieve fixed-time observation of other vehicles' states for passing sequence determination, a distributed optimization algorithm is introduced to schedule conflict-free arriving times for trajectory planning, and a distributed control algorithm is proposed to address parameter mismatches and acceleration saturation for fixed-time trajectory tracking control. Numerical simulations demonstrate that the proposed method can achieve cooperative passing of vehicles without global coordination at the cost of a growth of 8.8-18.1% average travel times in low and medium traffic volumes.