Adaptive Signal Control Research Articles

Timing is Everything

When bad weather hits, traffic slows down. But Paul Willis uncovers a research project in Canada that’s developing new SPaT programming that keeps traffic moving, without the need for adaptive signal control

Virtual Controller Interface Device for Hardware-in-the-Loop Simulation of Traffic Signals

Microscopic traffic simulation has been widely used for the design and/or evaluation of a new signal control strategy prior to field implementation. One of the most popular simulation approaches is the Hardware-in-the-Loop Simulation (HILS), which employs a Controller Interface Device (CID) to connect a physical signal controller to the simulation software to improve the accuracy of simulation. This paper documents the development and performance evaluation of a Virtual Controller Interface Device (VCID) based HILS. Different from the traditional physical CIDs, VCID provides wireless communication between signal controllers and simulation software through the National Transportation Communications for ITS Protocol (NTCIP). Controlled simulation experiments were conducted to compare the performance of VCID-based HILS with the Emulator-in-the-Loop Simulation (EILS), Software-in-the-Loop Simulation (SILS), and the traditional HILS. Queue length, delay, and vehicle trajectories were employed as performance measures. Simulation results show that VCID is an accurate and reliable tool for traffic signal simulation. In addition, since VCID-based HILS enables real-time updating of signal controller parameters during the simulation through the Simple Network Management Protocol (SNMP), it can efficiently assess the adaptive signal control algorithms that are developed based on the real-time traffic flow data collected by Connected Vehicles (CV) or Intelligent Transportation Systems (ITS).

Feasibility of using a single traffic signal controller to accommodate adjacent intersections

Abstract The operation cost of signalized intersections is usually higher than unsignalized intersections, not only because of the expenses on hardware devices but also due to the device maintenance as well as software upgrading. Moreover, at a signalized corridor, the coordination between signals is also considered to be necessary, which requires additional labor and maintenance costs. With these considerations, this paper aimed to investigate the feasibility of using a single signal controller to control two or multiple adjacent intersections. This study developed a procedure to decide when to use one-controller strategy and evaluated the operation of two real-world cases in Reno, Nevada, where the previous two-controller strategy has just replaced by a one-controller strategy. Based on microsimulation study, it was concluded that the Level-of-Service (LOS), delay, and the average number of stops under one-controller strategy maintained a similar condition in comparison with the previous two-controller strategy, indicating that the proposed one-controller strategy would be a feasible alternative to reduce the operation costs of adjacent intersections. It is expected that reducing the number of signal controllers will not only reduce the infrastructure cost, but also lead to notable operation benefits, such as it facilitates the development of signal coordination plans, and the implementation of future adaptive signal control in a connected vehicle environment, since it reduced the number of communication nodes within the arterial system.

Coordinated control method of intersection traffic light in one-way road based on V2X

One-way roads have potential for improving vehicle speed and reducing traffic delay. Suffering from dense road network, most of adjacent intersections' distance on one-way roads becomes relatively close, which makes isolated control of intersections inefficient in this scene. Thus, it is significant to develop coordinated control of multiple intersection signals on the one-way roads. This paper proposes a signal coordination control method that is suitable for one-way arterial roads. This method uses the cooperation technology of the vehicle infrastructure to collect intersection traffic information and share information among the intersections. Adaptive signal control system is adopted for each intersection in the coordination system, and the green light time is adjusted in real time based on the number of vehicles in queue. The offset and clearance time can be calculated according to the real-time traffic volume. The proposed method was verified with simulation results by VISSiM traffic simulation software. The results compared with other methods show that the coordinated control method proposed in this paper can effectively reduce the average delay of vehicles on the arterial roads and improve the traffic efficiency.

Dynamic platoon dispersion model based on real‐time link travel time

Adaptive signal control system aims to improve the operation efficiency of intersections, in which the platoon dispersion model is one of the key component modules. The problem of the traditional Robertson's model is that they are based on historical data, which cannot effectively capture the dynamic characteristics of traffic flow and be further applied to adaptive signal control systems. Thanks to the unprecedented availability of real-time travel time data provided by the modern intelligent transportation systems, there is an opportunity to calibrate Robertson's model online using real-time travel time data. This study proposes a dynamic Robertson's platoon dispersion model based on real-time travel times. Then, based on field observations, the prediction error and computational efficiency of the traditional Robertson's model, the Gaussian model, and the proposed model are compared and analysed. The results show that the proposed model can better capture the law of dynamic dispersion for traffic flow. The reduced average mean square error of prediction can be as much as 16.86 and 7.07% compared with the traditional Robertson's model and the Gaussian model, respectively. In addition, the average computational time of the proposed model is 0.46 s, which can be applied in the dynamic predictive signal control systems.

A dynamic optimization method for adaptive signal control in a connected vehicle environment

In a connected vehicle environment, vehicle location, speed, and other traffic information are readily available; hence, such environments provide new data sources for traffic signal control optimization. Existing adaptive signal control systems based on fixed detectors cannot directly obtain vehicle location and speed information, and thus, cannot provide accurate information about real-time traffic flow changes. This study presents a dynamic optimization method for adaptive signal control in a connected vehicle environment. First, the proposed method developed a dynamic platoon dispersion model to predict vehicle arrivals by using connected vehicle data. Then, a signal timing optimization model is constructed by regarding the minimization of average vehicle delay as the optimization objective, and setting the green time duration of each phase as a constraint. To achieve real-time adaptive signal control, a genetic algorithm is adopted to solve the optimization model through rolling optimization. Finally, a real-world road network was modeled in Vissim to validate the proposed method. Simulation results show that compared with the classical adaptive signal control algorithm, the proposed method is able to reduce vehicle delays and queue lengths at least 50% penetration rates. At 100% penetration rate, the proposed method improved the average vehicle delay and the average queue length by 22.7% and 24.8%, respectively. Moreover, it catered to all directions in a balanced manner.

On the estimation of connected vehicle penetration rate based on single-source connected vehicle data

With more connected vehicles (CVs) in the networks, the big data era leads to the availability of abundant data from CVs. CV penetration rate is the fundamental building block of tremendous applications, such as traffic data estimation, CV-based adaptive signal control and origin-destination estimation. While CV penetration rate is a random variable unknown in nature, the current estimation method of penetration rate mainly relies on two sources of data —detector and CV data. Penetration rate across the link is computed as CV flow divided by all traffic flow over a certain period of time. However, the current method is constrained by availability and quality of detector data. This paper proposes a simple, analytical, non-parametric, and most importantly, unbiased single-source data penetration rate (SSDPR) estimation method for estimating penetration rate solely based on CV data. It subtly and simultaneously fuses two estimation mechanisms—(1) the measurement of the probability of the first vehicle in a queue being a CV and (2) the direct estimation of the penetration rate of a sample queue—to constitute a single estimator to handle all the possible sample queue patterns. Applicability of the proposed method is not confined to a specific arrival pattern. It solely utilizes the number of the observed CVs and the number of vehicles before the last observed CV in a sample queue. Combining with bridging the queue algorithm, the proposed SSDPR estimation method is extended to overflow or oversaturated conditions. Simulation results show that the proposed method is able to accurately estimate penetration rate as low as 0.1% for all the situations considered. To illustrate the applicability of the proposed method, a case study of fundamental diagram estimation of a link without being installed with any detector is presented.

Decentralized network level adaptive signal control by multi-agent deep reinforcement learning

Adaptive traffic signal control systems are deployed to accommodate real-time traffic conditions. Yet travel demand and behavior of the individual vehicles might be overseen by their model-based control algorithms and aggregated input data. Recent development of artificial intelligence, especially the success of deep learning, makes it possible to utilize information of individual vehicles to control the traffic signals. Several pioneering studies developed model-free control algorithms using deep reinforcement learning. However, those studies are limited to isolated intersections and their effectiveness was only evaluated in ideal simulated traffic conditions by hypothetical benchmarks. To fill the gap, this study proposes a network-level decentralized adaptive signal control algorithm using one of the famous deep reinforcement methods, double dueling deep Q network in the multi-agent reinforcement learning framework. The proposed algorithm was evaluated by the real-world coordinated actuated signals in a simulated suburban traffic corridor which emulates the real-field traffic condition. The evaluation results showed that the proposed deep-reinforcement-learning-based algorithm outperforms the benchmark. It is able to reduce 10.27% of the travel time and 46.46% of the total delay.

Operational performance evaluation of adaptive traffic control systems: A Bayesian modeling approach using real-world GPS and private sector PROBE data

This evaluation ascertained the operational impacts of the SUTRAC (Scalable Urban Traffic Control) Adaptive Signal Control Technology (ASCT) in an urban corridor consisting of 23 intersections in Pittsburgh, Pennsylvania. A combination of real-world GPS floating car runs and private sector probe data from INRIX was used to assess the impact of the ASCT. Data were collected with the ASCT active and inactive to determine the operational impacts on the mainline and cross streets. The ASCT was found to produce significant improvements in the number of stops made along the corridor. On Baum and Center, travel times improved during the AM and PM peak in the WB direction. Speeds were also observed to improve significantly during the Midday period on Baum EB and during the AM and PM peak periods on Baum WB. Similarly, statistically significant improvements in speed were observed on Center WB during the AM and PM periods, while a statistically significant decrease in speed was observed during the Midday period. Six months of private sector probe data was used to examine travel time reliability along the corridor, and reliability was also found to have improved. Further, Bayesian models were calibrated to account for variations in speeds and acceleration/deceleration. The Bayesian models revealed that driving was less volatile with the ASCT system in operation over instantaneous periods, which also points towards improved operations. The findings of this study are generally consistent with past evaluations of other ASCTs, indicating that the SURTRAC system is another potential tool for managing congestion on signalized urban arterial networks.

A Novel Decentralized Game-Theoretic Adaptive Traffic Signal Controller: Large-Scale Testing.

This paper presents a novel de-centralized flexible phasing scheme, cycle-free, adaptive traffic signal controller using a Nash bargaining game-theoretic framework. The Nash bargaining algorithm optimizes the traffic signal timings at each signalized intersection by modeling each phase as a player in a game, where players cooperate to reach a mutually agreeable outcome. The controller is implemented and tested in the INTEGRATION microscopic traffic assignment and simulation software, comparing its performance to that of a traditional decentralized adaptive cycle length and phase split traffic signal controller and a centralized fully-coordinated adaptive phase split, cycle length, and offset optimization controller. The comparisons are conducted in the town of Blacksburg, Virginia (38 traffic signalized intersections) and in downtown Los Angeles, California (457 signalized intersections). The results for the downtown Blacksburg evaluation show significant network-wide efficiency improvements. Specifically, there is a reduction in travel time, a reduction in queue lengths, and a reduction in emissions relative to traditional adaptive traffic signal controllers. In addition, the testing on the downtown Los Angeles network produces a reduction in travel time on the intersection approaches, a reduction in queue lengths, and a reduction in emissions compared to traditional adaptive traffic signal controllers. The results demonstrate significant potential benefits of using the proposed controller over other state-of-the-art centralized and de-centralized adaptive traffic signal controllers on large-scale networks both during uncongested and congested conditions.

An adaptive signal control using connected-vehicle data

Traffic signal control at signalised intersections in urban cities has significant impacts on traffic performance. Connected-vehicle data, which is wirelessly transmitted by individual vehicles to roadside units, produces a clearer picture of traffic. Using this source of new data, an adaptive signal-control algorithm was developed with the aim of reducing traffic delays at intersections. A certain type of traffic-state information (elapsed time, vehicle stops or queue length) was first integrated for each traffic phase to reveal current traffic conditions. Phase sequence and duration were then adjusted, with the maximum and minimum restrictions of green time. The proposed algorithm was tested in an intersection simulated in the Vissim software package and the algorithm based on elapsed time was found to perform best. A comparison of the elapsed-time-based control scheme and the optimal fixed control scheme indicated that the former scheme outperformed the latter, with lower average vehicle delays in each traffic demand case even under low connected-vehicle penetration rates and with less variance in average delay for each traffic movement in most scenarios. In summary, the proposed signal-control algorithm could improve the operation of isolated intersections efficiently and serve all traffic movements in a more impartial way.

A Two-Level Rolling Optimization Model for Real-time Adaptive Signal Control

Recently, dynamic traffic flow prediction models have increasingly been developed in a connected vehicle environment, which will be conducive to the development of more advanced traffic signal control systems. This paper proposes a rolling optimization model for real-time adaptive signal control based on a dynamic traffic flow model. The proposed method consists of two levels, i.e., barrier group and phase. The upper layer optimizes the length of the barrier group based on dynamic programming. The lower level optimizes the signal phase lengths with the objective of minimizing vehicle delay. Then, to capture the dynamic traffic flow, a rolling strategy was developed based on a real-time traffic flow prediction model. Finally, the proposed method was compared to the Controlled Optimization of Phases (COP) algorithm in a simulation experiment. The results showed that the average vehicle delay was significantly reduced, by as much as 17.95%, using the proposed method.

An Adaptive Signal Control Method with Optimal Detector Locations

The adaptive traffic signal control system is a key component of intelligent transportation systems and has a primary role in effectively reducing traffic congestion. The high costs of implementation and maintenance limit the applicability of the adaptive traffic signal control system, especially in developing countries. This paper proposes a low-cost adaptive signal control method that is easy to implement. Two detectors are installed in each vehicle lane at an optimal location determined by the proposed method to detect green and red redundancy time, based on which the original signal timing is adjusted through a signal controller. The proposed method is evaluated through case studies with low and high volume-to-capacity ratio intersections. The results show that the proposed adaptive signal control method can significantly reduce total traffic delay at intersections.

Crash severity effects of adaptive signal control technology: An empirical assessment with insights from Pennsylvania and Virginia

Adaptive signal control technology (ASCT) is an intelligent transportation systems (ITS) technology that optimizes signal timings in real time to improve corridor flow. While a few past studies have examined the impact of ASCT on crash frequency, little is known about its effect on injury severity outcomes. Similarly, the impact of different types of ASCTs deployed across different states is also uncertain. This paper therefore, used ordered probit models with random parameters to estimate the injury severity outcomes resulting from ASCT deployment across Pennsylvania and Virginia. Two disparate systems deployed across the two different states were analyzed to assess whether they had similar impacts on injury severity, although signal timings are optimized using different algorithms by both systems. The estimation results revealed that both ASCT systems were associated with reductions in injury severity levels. Marginal effects showed that Type A ASCT systems reduced the propensity of severe plus moderate and minor injury crashes by 11.70% and 10.36% while type B ASCT reduced the propensity of severe plus moderate and minor injury crashes by 4.39% and 6.92%. Similarly, the ASCTs deployed across the two states were also observed to reduce injury severities. The combined best fit model also revealed a similar trend towards reductions in severe plus moderate and minor injury crashes by 5.24% and 9.91%. This model performed well on validation data with a low forecast error of 0.301 and was also observed to be spatially transferable. These results encourage the consideration of ASCT deployments at intersections with high crash severities and have practical implications for aiding agencies in making future deployment decisions about ASCT.

Adaptive traffic signal control for real-world scenarios in agent-based transport simulations

This study provides an open-source implementation of a decentralized, adaptive signal control algorithm in the agent-based transport simulation MATSim, which is applicable for large-scale real-world scenarios. The implementation is based on the algorithm proposed by Lämmer and Helbing (2008), which had promising results, but was not applicable to real-world scenarios in its published form. The algorithm is extended in this paper to cope with realistic situations like different lanes per signal, small periods of overload, phase combination of non-conflicting traffic, and minimum green times. Impacts and limitations of the adaptive signal control are analyzed for a real-world scenario and compared to a fixed-time and traffic-actuated signal control. It can be shown that delays significantly reduce and queue lengths are lower and more stable than with fixed-time signals. Another finding is that the adaptive signal control behaves like a fixed-time control in overload situations and, therefore, ensures system-wide stability.

Cooperative Deep Q-Learning With Q-Value Transfer for Multi-Intersection Signal Control

The problem of adaptive traffic signal control in the multi-intersection system has attracted the attention of researchers. Among the existing methods, reinforcement learning has shown to be effective. However, the complex intersection features, heterogeneous intersection structures, and dynamic coordination for multiple intersections pose challenges for reinforcement learning-based algorithms. This paper proposes a cooperative deep Q-network with Q-value transfer (QT-CDQN) for adaptive multi-intersection signal control. In QT-CDQN, a multi-intersection traffic network in a region is modeled as a multi-agent reinforcement learning system. Each agent searches the optimal strategy to control an intersection by a deep Q-network that takes the discrete state encoding of traffic information as the network inputs. To work cooperatively, the agent considers the influence of the latest actions of its adjacencies in the process of policy learning. Especially, the optimal Q-values of the neighbor agents at the latest time step are transferred to the loss function of the Q-network. Moreover, the strategy of the target network and the mechanism of experience replay are used to improve the stability of the algorithm. The advantages of QT-CDQN lie not only in the effectiveness and scalability for the multi-intersection system but also in the versatility to deal with the heterogeneous intersection structures. The experimental studies under different road structures show that the QT-CDQN is competitive in terms of average queue length, average speed, and average waiting time when compared with the state-of-the-art algorithms. Furthermore, the experiments of recurring congestion and occasional congestion validate the adaptability of the QT-CDQN to dynamic traffic environments.

Estimation of Turning Ratio at Intersections Based on Detector Data using Kalman Filter

Abstract Adaptive signal control system shows strong flexibility and high performance in traffic signal control compared with other control systems. Even though it still has weaknesses such as in turning ratio estimation on so called mixed lanes, on which traffic streams turn into different directions. This may lead to wrong estimation of platoon intensities and as a consequence to non-optimal and unreliable signal control. Therefore, this study proposes an extended Kalman filter (EKF) based on extended cell transmission model (ECTM), to estimate those turning ratios using local detector data of the related intersectons and its neighbors. The ECTM models the traffic situation in signalized urban area macroscopically. The ECTM was developed by Zeng (2012) and Papapanagiotou (2013) based on the orignal cell transimission model (CTM) but overcomes some disadvantages of the original CTM according to urban signalized intersections. In this contribution the ECTM along with the measurement model is organized in a compact state-space form, based on which a turning ratio estimator is designed using EKF. While creating the turning ratio estimator, attention has been especially paid to the linearization of ECTM, in order to organize it into the state-space model and make sure that the EKF is able to interpret the traffic information properly. For estimating the turning ratios, the cell occupancies of the ECTM are also estimated and corrected. Therefor the ECTM based EKF provides a traffic state estimation besides turning ratio estimation. In this study, traffic situation is simulated with Vissim and it is taken as a substitution of reality. The estimated turning ratio using EKF based on ECTM is compared with the turning ratios configured in Vissim and evaluated from two aspects: the mechanism analysis and the sensitivity analysis. In summary, the estimation of turning ratio by ECTM based EKF provides a reliable traffic state estimation. Therefore, it promises a better optimization of signals control, for increasing efficiency in traffic management.

Joint optimization of vehicle trajectories and intersection controllers with connected automated vehicles: Combined dynamic programming and shooting heuristic approach

Connected and automated vehicle (CAV) technologies offer promising solutions to challenges that face today’s transportation systems. Vehicular trajectory control and intersection controller optimization based on CAV technologies are two approaches that have significant potential to mitigate congestion, lessen the risk of crashes, reduce fuel consumption, and decrease emissions at intersections. These two approaches should be integrated into a single process such that both aspects can be optimized simultaneously to achieve maximum benefits. This paper proposes an efficient DP-SH (dynamic programming with shooting heuristic as a subroutine) algorithm for the integrated optimization problem that can simultaneously optimize the trajectories of CAVs and intersection controllers (i.e., signal timing and phasing of traffic signals), and develops a two-step approach (DP-SH and trajectory optimization) to effectively obtain near-optimal intersection and trajectory control plans. Also, the proposed DP-SH algorithm can also consider mixed traffic stream scenarios with different levels of CAV market penetration. Numerical experiments are conducted, and the results prove the efficiency and sound performance of the proposed optimization framework. The proposed DP-SH algorithm, compared to the adaptive signal control, can reduce the average travel time by up to 35.72% and save the consumption by up to 31.5%. In mixed traffic scenarios, system performance improves with increasing market penetration rates. Even with low levels of penetration, there are significant benefits in fuel consumption savings. The computational efficiency, as evidenced in the case studies, indicates the applicability of DP-SH for real-time implementation.

Adaptive multi-objective traffic signal control using NLRMNSGA-II algorithm

In this paper, we consider an adaptive system for controlling green times at junction. For this adaptive system, we present a multi-objective optimization model, which is much easier to solve than some existing models. Furthermore, to solve the new model, we suggest an algorithm, called NLRMNSGA-II, which is based on the nonlinear least regression and a modified non-dominated sorting genetic algorithm. Our numerical experiments indicate that the NLRMNSGA-II is an efficient algorithm for the considered adaptive system.

Utilizing bluetooth and adaptive signal control data for real-time safety analysis on urban arterials

Real-time safety analysis has been widely adopted to reveal the relationship between real-time traffic characteristics and crash occurrence, and these results could be applied to improve active traffic management systems and enhance safety performance. Most of the previous studies have been applied to freeways and seldom to arterials. This study attempts to examine the relationships between crash occurrence and real-time traffic and signal timing characteristics based on four urban arterials in Central Florida. Bayesian conditional logistic models (BCL) were developed by incorporating the Bluetooth, adaptive signal control, and weather data, which were extracted for a period of 20 min (four 5-minute intervals) before the time of crash occurrence. Model comparison results indicated that the model based on 5–10 min interval dataset performed the best. It revealed that the average speed, upstream left-turn volume, downstream green ratio, and rainy indicator were found to have significant effects on crash occurrence. Furthermore, Bayesian random parameters conditional logistic model (BRPCL) outperformed Bayesian random parameters logistic (BRPL) and Bayesian conditional logistic models (BCL) in terms of the area under the receiver operating characteristics curve (AUC) and Deviance Information Criterion (DIC) values. These results are important in real-time safety applications in the context of Integrated Active Traffic Management (IATM).