In order to balance the temporal-spatial distribution of urban traffic flow, this paper establishes a model for combined urban traffic signal control and traffic flow guidance. With consideration of the widely used of fixed control at intersections, traffic assignment under traffic flow guidance, dynamic characteristics of urban traffic management, a tri-level programming model is present. To reflect the impact of intersection delay on traffic assignment, some virtual roads are drawn into the lower level model, which is a modified User Equilibrium model. Aimed at the minimization of average vehicle delay and generalized saturation, the middle level model, contains some definitional constraints for different phase modes, is built for the traffic signal control optimization. Based on nonlinear 0-1 integer programming, the upper level model is put forward for the tide lane management. A Heuristic Iterative Optimization Algorithm (HIOA) is set up to solve the tri-level programming model. The lower model is solved by Method of Successive Averages (MSA), the middle level model is solved by Non-dominated Sorting Genetic Algorithm II (NSGA II), and the upper model is solved by Genetic Algorithm (GA). The case study shows the efficiency and applicability of the proposed modelling and computing method.

The performance of an adaptive signal control system depends on the embedded traffic flow prediction and control algorithm. This paper proposed an improved adaptive control method, comprised of a vehicle arrival estimation model and a signal optimization algorithm. A microscopic model was developed to capture vehicle arrival dynamics both in red and green durations. Based on the upstream detection information, vehicle arrival time is predicted taking account of signal timing, vehicle trajectory and variable queue length. The real-time signal control algorithm was formulated based on a Dynamic Programming (DP) procedure, which can support NEMA phase configuration without difficulty. Three objective functions were considered, including minimization of delay, queue length and maximization of throughput. The proposed adaptive control method was implemented in VISSIM on a real signalized intersection. It is shown that the DP with NEMA generally outperforms DP with 4-phase and the signal plan generated by SYNCHRO, under varying demand levels and planning horizons. Different objective functions lead to different control performance in terms of various demands and planning horizons. It is reasonably believed that the model has potential applicability in real time traffic adaptive signal control systems.

In recent years considerable attention has been given to Green Light Optimal Speed Advisory (GLOSA) systems in the literature. These systems typically compute a single speed for a vehicle in the range between minimum and maximum values. The goal is to optimise for a particular objective, whether it be minimisation of emissions (for environmental reasons), fuel usage or delay. This paper generalises the advice given to vehicles, by optimising primarily for delay over the entire trajectory instead of suggesting an individual speed, regardless of initial conditions. The relationship between delay and fuel usage and emissions is investigated, and the evidence of the importance of taking a platoon-based approach is presented.

As will be shown, the effectiveness of GLOSA systems is heavily reliant on the leading vehicle that is given trajectory advice, and a "time loop" technique is proposed which allows accurate identification to occur even when there are complex interactions between preceding vehicles.

A number of studies had worked on Green Light Optimal Speed Advisory (GLOSA). Most of them focused on the macro level where the process of speed changes was ignored. Rakha et al. considered the acceleration process, but there was no simulation and field data to validate the method. As the vehicle approaches the intersection with different speed, the fuel consumption will be different. This study introduces a green light speed advisory strategy in a connected vehicle environment, which helps the vehicle to avoid unnecessary stop before intersections and improve the fuel consumption efficiency. First, this paper formulates explicit fuel-consumption objective functions based on VT micro model to measure the fuel consumption of vehicle’s different speed profiles. Then, green light speed advisory strategies are developed to reduce fuel consumption. Over 100,000 rounds of simulation experiment are conducted in MATLAB. The simulation results show that the average fuel saving is 56% for green light signal, and is 21% for red signal scenarios. In addition, an android app is developed to implement the proposed strategy in a connected vehicle environment. The results from the field test show that fuel consumption is reduced by 13% after implementing GLOSA.

In this study, we develop a stage-based max-pressure signal control policy for the adaptive optimization of signal timings. The objective is to test the performance of the proposed algorithm compared to other existing signal control algorithms, at isolated signalized junctions using a microscopic simulation. The proposed algorithm consists of three parts; the real-time estimation method of lane-based queue lengths, the global optimization of signal timings, and the local adjustment of green phases. The challenges are to estimate lane-based queue lengths and integrate the global optimization scheme with the local adjustment scheme. Information from upstream and downstream detectors at isolated signalized junctions is used to estimate real-time lane-based queue lengths and traffic intensities, which are essential input data for the global optimization and the local adjustment. The queue lengths on individual lanes are calculated using discriminant models and downstream arrivals. Allsop’s stage-based method is applied as the global optimization scheme at the end of the cycle to optimize cycle time and green time for stages. Stage-green time in the current second is adjusted to the greatest stage-pressure in the previous second, using Varaiya’s method as the local adjustment scheme in real time. The performance of the proposed algorithm and existing algorithms are simulated at an isolated intersection. The computer simulation will show that the proposed method outperforms existing signal control algorithms, particularly on asymmetric traffic conditions, as evidenced by the high accuracy of the estimation over time of lane-based queue lengths and the dynamic optimization of signal timings.

Increasing traffic volumes result in congestion especially at intersections in urban areas. Effective regulation of vehicle flows at intersections has always been an important issue in the traffic control system. This paper presents a multi-objective optimization method at intersections with cellular automata. Vehicle conflicts and pedestrian interference are considered. Three categories of the traffic performance are studied including transportation efficiency, energy consumption and road safety. The signal timing and lane assignment are optimized for different traffic flows. The multi-objective optimization problem is solved with cell mapping method. The optimization results show the conflicting nature of different traffic performance. The influence of different traffic variables on the intersection performance is investigated in the perspective of optimization. It is observed that the proposed optimization method is effective in controlling the traffic at the intersection to meet multiple objectives. Transportation efficiency can be usually improved by the permissive left-turn signal, which sacrifices safety. Right-turn traffic suffers significantly when the right-turn lanes are shared with the through vehicles. The effect of vehicle flow on the optimization results is significant. Pedestrians have strong interference with the traffic system.

Signal control systems have been continuously developed all over the world, either in the setting methods of signal parameters or in control strategies. It is of importance to provide a convenient and cheap approach to improve the existing signal control based on the current road infrastructure. Fuzzy Intelligent Traffic Signal Control (FITS) is an intermediate hardware device capable of receiving message from signal controller hardware as well as real-time overriding traffic light indications. Different signal control strategies and optimization toolbox can be integrated as the software in FITS hardware device. In order to evaluate the effects of FITS system, this attempts to develop a simulation-based evaluation framework for FITS system. A case study is carried out by comparing different commonly used signal control strategies, including stage-based vehicle actuated control and dual-ring vehicle actuated control as well as group-based actuated control, to fuzzy control strategy, the default control strategy in FITS. The simulation results reveal that FITS outperforms all of the tested signal control strategies due to its ability in generating flexible phase structures and flexible cycles. In addition, the negative effects of detection malfunction have been investigated in this study. The experiment results point out that the influence on FITS is acceptable when a small amount of detectors are fault.

It is widely-accepted that travelers value the reliability of travel time in addition to its mean or expected value.  Most strategies for traffic signal control typically seek to optimize average travel times, however, whereas reliability is in general not explicitly taken into account.

In this paper, we propose a new framework for evaluating the consequences of signal-control tactics on both reliability and expected values of travel time, based on a new analytic model of travel time distribution. A Genetic-algorithm-based approach is then employed to identify optimal multi-criterion signal control strategies, including sensitivity analysis to the relative weighting between reliability and expected value.

We expose the properties of the proposed framework via an empirical case study of three alternative optimization approaches (the signal setting optimized with TRANSYT, the widely-used SCATS adaptive-control strategy, and the newly-proposed model) under various traffic conditions.  Results indicate that the newly-proposed framework outperforms the two other signal control strategies in terms of travel time variability and even expected travel time.

Key words: traffic control, travel time reliability, Genetic Algorithm, signal optimization

The two-stage stochastic mixed-integer linear program with recourse formulation of the existing lane-based traffic signal setting and lane marking design model at an isolated signalized intersection is proposed with two alternative objective functions: capacity maximization and cycle-length minimization.  The traffic volumes and lane saturation flows are explicitly considered as random variables with known probability distributions.  The Monte Carlo bounding techniques with independent-random-number (IRN) and common-random-number (CRN) sampling strategies are employed to determine the optimality gap of the stochastic solution obtained from the external sampling procedure.  From illustrative examples, for capacity-maximization problems, the optimal cycle lengths in deterministic and stochastic solutions are at the maximum value.  For cycle-length-minimization problems, the minimal cycle lengths in stochastic solutions are higher than those in deterministic solutions.  For both deterministic and stochastic solutions, the computational times for cycle-length-minimization problems are less than those for capacity-maximization problems on the same intersection size.  The CRN sampling strategy outperforms the IRN sampling strategy for capacity-maximization problems because the CRN strategy yields tighter 95% confidence intervals (with greater-than-one variance reduction value) and less total computational time than the IRN strategy.  The CRN and IRN sampling strategies appear non-dominated for cycle-length-minimization problems because the upper-bound and lower-bound estimators are not positively correlated.  It is valuable to solve the stochastic model for capacity-maximization problem as the 95% confidence interval of the value of stochastic solution (VSS) in terms of reserve capacity is up to 44.93%±1.23%.  The VSS cannot be estimated because the deterministic solution is infeasible in the stochastic environment. 

Optimizing intersection signal timing plans over an urban network is essential to effectively utilize the capacity of the existing transportation system and mitigate urban traffic congestion. The proposed research develops a new optimization model to address intersection signal timing optimization over a network, integrating the intersection control delay captured by the formulation given in HCM 2010 and the traffic flow response captured by traffic user equilibrium. More exactly, both the time delay formulation in HCM 2010 and traffic user equilibrium are formulated as complementarity constraints. Consequently, the proposed problem is formulated as mathematical programming with equilibrium constraints (MPEC). One advantage of the MPEC formulation is that it can be efficiently solved by GAMS/NLPEC solver without involving significant computational efforts of simulation-based network models for intersection signal timing optimization over a large urban street network. Computational experiments built upon a subnetwork in the Chicago central district indicate that the proposed approach outperforms the extent of reductions in network-wide intersection control delay and total travel cost that can be achieved by Synchro tool, a widely used and accepted signal timing software. Additionally, the computational experiments also indicate that the proposed MPEC formulation can be solved fairly quickly. The proposed model is expected to help traffic engineers refine intersection signal timing plans in urban areas.

Traffic signal cycle lengths are typically computed to minimize the intersection vehicle delay using the Webster formula. The objectives of this study are two-fold. First, it validates the Webster formula against simulated data. Second, it develops new formulations to compute the optimum cycle length considering other measures of effectiveness including vehicle fuel consumption levels and tailpipe emissions. The microscopic simulation software, INTEGRATION, is used to simulate a two-phase intersection over a range of cycle lengths, traffic demand levels, and signal timing lost times. Intersection delay, fuel consumption levels, and hydrocarbon (HC), carbon monoxide (CO), oxides of nitrogen (NO_x), and carbon dioxide (CO₂) emissions were derived from the simulation model. The cycle lengths that minimized the various measures of effectiveness were then used to develop the proposed models. The first effort entailed re-calibrating the Webster model to the simulated data. The second effort entailed enhancing the Webster model by incorporating an additional intercept term. The proposed model is demonstrated to produce better traffic signal timings and is calibrated to minimize delay, fuel consumption and CO ₂ emission levels. The model estimates produce shorter cycle lengths when compared to the Webster model and also considers fuel consumption and Green House Gas (GHG) emissions in the optimization procedure.

PTV Epics, supported by PTV Group, is a real time adaptive signal controller method and  integrated with Balance can be used to optimize and control single isolated intersection or entire corridor.  Epics is used to perform real time optimization on single intersection where, according to PTV Group, would offer the best results. PTV VISSIM microsimulation software was used as the evaluation and optimization tool for Epics. The purpose of PTV Balance is to provide corridor/network coordination and optimization. Balance calibration and assessment is done through PTV Visum. Epics-Balance integration is encouraged by PTV Group, since it is expected to offer the greatest improvements in performance metrics, where traffic operations would be optimized on the intersection and corridor level simultaneously. This study proposes to evaluate free running Epics and Epics coordinated operation, while Balance central coupled with Epics will be evaluated in later studies. State Route421, 12-intersection corridor in Volusia County, FL is used as a case study.  VISSIM model was carefully calibrated and validated to replicate field conditions as much as possible. Epics model is calibrated as instructed by PTV Group user manual. Results show that Epics (free running and coordinated) reduces vehicle delays, stop delays and queue lengths, while number of stops is increased compared to RBC. A major Epics benefit proved to be the reduction of side street delay (greater than 50%), mainly due to fact that it doesn’t allow preferential treatment for any of the approaches.

We propose an algorithm for real-time optimization of traffic lights in urban road networks. The algorithm is based on maximizing controller flow subject to a macroscopic traffic dynamics model (the cell transmission model). We make simplifications to this formulation that preserve its effectiveness at optimizing traffic lights while decreasing the computational cost by an order of magnitude. The resulting algorithm is based on alternately solving a number of continuous knapsack problems (which are computationally cheap) and simulating traffic dynamics (also cheap). Our algorithm is centralized, allowing coordination between traffic lights, however computationally cheap enough for use in real-time on road networks of moderate size. Numerical results are presented which support this claim and demonstrate the effectiveness of our algorithm at maximizing flow and minimizing delay.

In urban transportation system, the goal of traffic signal control is to reduce individual delay and improve safety for all travelers in different modes. To achieve signal control optimization from all user’s perspective, this paper proposes a platoon-based adaptive control (PASC) strategy to provide multi-modal signal control based on the on-line Connected Vehicle (CV) data. By introducing unified phase precedence constraints, PASC strategy is not restricted by fixed cycle length and offsets. A Mixed Integer Linear Programming (MILP) model was developed to optimize signal timings in a real-time manner, with platoon arrival and discharge dynamics at stop line modeled as constraints. Based on the passenger occupancy, the objective function aims to minimize total passenger delay for bus and automobiles. With the communication between signals, PASC achieves to provide implicit coordination for the signalized arterial, usually in both directions. The simulation results indicate that the PASC model successfully reduces 40% and 10% total passenger delay at high and medium demand scenario respectively compared with SYNCHRO. In addition, PASC generates more than 20% individual delay reduction for transit buses. Results from sensitivity analysis demonstrate that the requested penetration rate range is from 40% to 60% for the implementation, indicating the efficiency and the capability of the model.

This paper describes a scalable approach to simulation of decentralized adaptive signal control systems (Surtrac in our case). The approach centers around a simulation controller interface called VISCO, which links the VISSIM microscopic traffic simulator to a set of externally hosted local intersection control processes. Local control processes are free to communicate with each other and exchange control information in the same manner that they would in a field implementation. VISCO coordinates all interaction with the simulator process to create a distributed software-in-the-loop simulation architecture. To illustrate and analyze the efficacy of the approach, we described simulation analysis of the downtown triangle area of Pittsburgh PA. A 63-intersection VISSIM model of this site was constructed and analyses were presented to characterize both the efficiency of the distributed architecture and the potential utility of Surtrac adaptive control. With respect to the former, the distributed simulation of Surtrac control processes was found to run in roughly 3 times faster than real-time, in comparison to the 11 times faster than real-time speed that a conventional VISSIM simulation of this model with fixed timing plans performed. However, it was also shown that over 90% of this slowdown is due to the COM communication interface that VISSIM provides for external controller integration, suggesting the potential for significant performance improvement. With respect to adaptive signal system improvement in the downtown triangle area of Pittsburgh, the simulation analysis showed strong benefit over both the existing timing plans in use in this area and Synchro optimized plans generated with perfect knowledge of traffic volumes and turning counts.

When the traffic state becomes oversaturated, intersections are not immune to congestion.  It is necessary to understand the traffic dynamic of the signalized arterial network to determine proper signal control plans. A number of elaborate arterial traffic flow models, which were deductively derived to describe the complex interactions between traffic state evolution and key control parameters, have been applied to provide relatively accurate predictions. Model-based proactive control systems can adjust, in real time, signal timing plans to handle fluctuations and congestion-based predicted traffic states. A particular simplified control design pursued in various past studies is based on store-and-forward modeling (SFM). This paper presents a new methodology for optimizing the signal timing controls of an oversaturated corridor based on an enhanced SFM model and a multi-objective technique. The multi-objective model seeks to find trade-offs between maximizing throughput and minimizing the average queue ratio. An adaptive control simulation platform is developed using a full-scale signal simulator, ASC/3, in VISSIM. A case study of a 7.4-kilometer arterial corridor reveals that (1) the Pareto-frontier can provide more approriate altermatives under particular situations, and (2) the proposed model is  promising for use in the design of arterial signals, especially under congested, high demand traffic conditions, as compared to traditional actuated signal control.

Simultaneous Perturbation Stochastic Approximation (SPSA) has gained favor as an efficient method for optimizing computationally expensive, “black box” traffic simulations. However, few recent studies have investigated the efficiency of SPSA for traffic signal timing optimization. It is important for this to be investigated, because significant room for improvement exists in the area of signal optimization. Some signal timing methods and products perform optimization very quickly, but deliver mediocre solutions. Other methods and products deliver high-quality solutions, but deliver those solutions very slowly. When using commercialized desktop signal timing products, engineers are often forced to choose between speed and solution quality. Real-time adaptive control products, which must optimize timings within seconds on a cycle-by-cycle basis, do not have enough time to reach a high-quality solution. Based on research results in the literature, SPSA holds the possibility of upgrading both desktop and adaptive solutions alike, by delivering high-quality solutions within seconds. This paper describes an extensive set of optimization tests involving SPSA. The final results suggest that today’s signal timing solutions could be improved significantly by incorporating SPSA, genetic algorithms, and ‘playbooks’ of pre-optimized starting points.

Basic signal operation strategies allocate green time to different traffic movements to control the traffic flow at an intersection. Signal control applications consider different aspects such as coordination with multiple intersections, multi-modal priority and safety issues. Currently, real-time signal control applications mainly rely on infrastructure-based detection data. With the emergence of connected vehicle technology, high resolution data from connected vehicles will become available for signal control. This paper presents a framework that utilizes connected vehicle data for adaptive signal control including consideration of dilemma zone protection, multi modal signal priority and coordination. Initially the market penetration rate of connected vehicles will be low, so infrastructure-based detector actuation logic is integrated into the framework to improve performance. Results show the good results for all applications when the penetration rate is medium to high and the actuation logic is necessary when the penetrate rate is low.

Connected vehicle (CV) data has the potential to transform traffic signal operations, but the success of control methods based on CV data will depend on the level of market penetration. Recent studies of real-time operational strategies in CV environments suggest that penetrations exceeding 20% will be required. This study explores the feasibility of using CV data to generate arrival profiles for optimizing arterial progression. Applications to offline (3-hour analysis period) and online (15-minute analysis period) offset optimization are considered. Vehicle arrival profiles obtained from real world measurement are used as a basis for comparison. Subsampled distributions are used to estimate the potential distributions that might be obtained from CVs, and these are statistically analyzed to explore the impacts of penetration rate, analysis period, and traffic volume. For selected penetration rates ranging from 0.1% to 50%, the subsampled distributions are used to optimize the corridor, and the results are evaluated in the complete-data model. The results show that over a 3-hour window, successful offline optimization can be achieved with a CV penetration rate as low as 1%.  Layering multiple days of data could potentially allow offline optimization with penetration rates as low as 0.1%. Online optimization with 15-minute windows require somewhat higher penetration rates of at least 5%. The results suggest that early applications of CV data may be possible at very low levels of market penetration. In corridors with high penetration of connected mobile devices, some private sector probe data services may be at the cusp of providing the necessary data to facilitate detector-free optimization.