Search for: All records

Perimeter metering control has long been an active research topic since well-defined relationships between network productivity and usage, that is, network macroscopic fundamental diagrams (MFDs), were shown to be capable of describing regional traffic dynamics. Numerous methods have been proposed to solve perimeter metering control problems, but these generally require knowledge of the MFDs or detailed equations that govern traffic dynamics. Recently, a study applied model-free deep reinforcement learning (Deep-RL) methods to two-region perimeter control and found comparable performances to the model predictive control scheme, particularly when uncertainty exists. However, the proposed methods therein provide very low initial performances during the learning process, which limits their applicability to real life scenarios. Furthermore, the methods may not be scalable to more complicated networks with larger state and action spaces. To combat these issues, this paper proposes to integrate the domain control knowledge (DCK) of congestion dynamics into the agent designs for improved learning and control performances. A novel agent is also developed that builds on the Bang-Bang control policy. Two types of DCK are then presented to provide knowledge-guided exploration strategies for the agents such that they can explore around the most rewarding part of the action spaces. The results from extensive numerical experiments on two- and three-region urban networks show that integrating DCK can (a) effectively improve learning and control performances for Deep-RL agents, (b) enhance the agents’ resilience against various types of environment uncertainties, and (c) mitigate the scalability issue for the agents. 

This paper proposes a novel decentralized signal control algorithm that seeks to improve traffic delay equity, measured as the variation of delay experienced by individual vehicles. The proposed method extends the recently developed delay-based max pressure (MP) algorithm by using the sum of cumulative delay experienced by all vehicles that joined a given link as the metric for weight calculation. Doing so ensures the movements with lower traffic loads have a higher chance of being served as their delay increases. Three existing MP models are used as baseline models with which to compare the proposed algorithm in microscopic simulations of both a single intersection and a grid network. The results indicate that the proposed algorithm can improve the delay equity for various traffic conditions, especially for highly unbalanced traffic flows. Moreover, this improvement in delay equity does not come with a significant increase to average delay experienced by all vehicles. In fact, the average delay from the proposed algorithm is close to—and sometimes even lower than—the baseline models. Therefore, the proposed algorithm can maintain both objectives at the same time. In addition, the performance of the proposed control strategy was tested in a connected vehicle environment. The results show that the proposed algorithm outperforms the other baseline models in both reducing traffic delay and increasing delay equity when the penetration rate is less or equal to 60%, which would not be exceeded in reality in the near future. 

Two key aggregated traffic models are the relationship between average network flow and density (known as the network or flow macroscopic fundamental diagram [flow-MFD]) and the relationship between trip completion and density (known as network exit function or the outflow-MFD [o-FMD]). The flow- and o-MFDs have been shown to be related by average network length and average trip distance under steady-state conditions. However, recent studies have demonstrated that these two relationships might have different patterns when traffic conditions are allowed to vary: the flow-MFD exhibits a clockwise hysteresis loop, while the o-MFD exhibits a counter-clockwise loop. One recent study attributes this behavior to the presence of bottlenecks within the network. The present paper demonstrates that this phenomenon may arise even without bottlenecks present and offers an alternative, but more general, explanation for these findings: a vehicle’s entire trip contributes to a network’s average flow, while only its end contributes to the trip completion rate. This lag can also be exaggerated by trips with different lengths, and it can lead to other patterns in the o-MFD such as figure-eight patterns. A simple arterial example is used to demonstrate this explanation and reveal the expected patterns, and they are also identified in real networks using empirical data. Then, simulations of a congestible ring network are used to unveil features that might increase or diminish the differences between the flow- and o-MFDs. Finally, more realistic simulations are used to confirm that these behaviors arise in real networks. 

Recent studies have leveraged the existence of network macroscopic fundamental diagrams (MFD) to develop regional control strategies for urban traffic networks. Existing MFD-based control strategies focus on vehicle movement within and across regions of an urban network and do not consider how freeway traffic can be controlled to improve overall traffic operations in mixed freeway and urban networks. The purpose of this study is to develop a coordinated traffic management scheme that simultaneously implements perimeter flow control on an urban network and variable speed limits (VSL) on a freeway to reduce total travel time in such a mixed network. By slowing down vehicles traveling along the freeway, VSL can effectively meter traffic exiting the freeway into the urban network. This can be particularly useful since freeways often have large storage capacities and vehicles accumulating on freeways might be less disruptive to overall system operations than on urban streets. VSL can also be used to change where freeway vehicles enter the urban network to benefit the entire system. The combined control strategy is implemented in a model predictive control framework with several realistic constraints, such as gradual reductions in freeway speed limit. Numerical tests suggest that the combined implementation of VSL and perimeter metering control can improve traffic operations compared with perimeter metering alone. 

Restricting left turns throughout a network improves overall flow capacity by eliminating conflicts between left-turning and through-moving vehicles. However, doing so requires vehicles to travel longer distances. Implementing left-turn restrictions at only a subset of locations can help balance this tradeoff between increased capacity and longer trips. Unfortunately, identifying exactly where these restrictions should be implemented is a complex problem because of the many configurations that must be considered and interdependencies between left-turn restriction decisions at adjacent intersections. This paper compares three heuristic solution algorithms to identify optimal location of left-turn restrictions at individual intersections in perfect and imperfect grid networks. Scenarios are tested in which restriction decisions are the same for all intersection approaches and only the same for approaches in the same direction. The latter case is particularly complex as it increases the number of potential configurations exponentially. The results suggest all methods tested can be effectively used to solve this problem, although the hybrid method proposed in this paper appears to perform the best under scenarios with larger solution spaces. The proposed framework and procedures can be applied to realistic city networks to identify where left-turn restrictions should be implemented to improve overall network operations. Application of these methods to square grid networks under uniform demand patterns reveal a general pattern in which left turns should be restricted at central intersections that carry larger vehicle flows but allowed otherwise. Such findings can be used as a starting point for where to restrict left turns in more realistic networks. 

The development of traffic models based on macroscopic fundamental diagrams (MFD) enables many real-time control strategies for urban networks, including cordon-based pricing schemes. However, most existing MFD-based pricing strategies are designed only to optimize the traffic-related performance, without considering the revenue collected by operators. In this study, we investigate cordon-based pricing schemes for mixed networks with urban networks and freeways. In this system, heterogeneous commuters choose their routes based on the user equilibrium principle. There are two types of operational objective for operating urban networks: (1) to optimize the urban network’s performance, that is, to maximize the outflux; and (2) to maximize the revenue for operators. To compare those two objectives, we first apply feedback control to design pricing schemes to optimize the urban network’s performance. Then, we formulate an optimal control problem to obtain the revenue-maximization pricing scheme. With numerical examples, we illustrate the difference between those pricing schemes.