Recently, automated traffic signal performance measures (ATSPMs) have been deployed with increasing frequency. With increased use, new agency needs are being identified with respect to the technology. In particular, methods of data aggregation to the system level and context for evaluating system metrics are needed. This paper proposes a method of evaluating corridor performance at the system level using high-resolution data. The method is demonstrated for eight signalized corridors in Indiana, including 87 intersections. This method develops five subscores for the areas of communication, detection, safety, capacity allocation, and progression; these five interrelated aspects of performance are each given a letter grade based on quantitative performance measures, with scales appropriate to the context of the operation. An aggregate score for each corridor is determined based on the lowest subscore of each of the five areas. This approach simplifies the analysis process, as opposed to examining several hundred individual movements as currently would be required using ATSPM tools that are commonly available at present. The methodology is presented as a prototype for further development and adaptation to individual agency objectives and data sources.

There has been considerable implementation of commercial probe data average segment speeds, where segments are on the order of 1 mile in length.  This paper presents a method for computing control delay using high fidelity commercial probe vehicle trajectory data.  In collaboration with a commercial probe data provider, approximately 38 million records starting in, ending in, or going through Salt Lake City were obtained for a two-month period in 2017.  To identify relevant data for this study, virtual detection boxes were defined between eight signalized intersections along a corridor in Cottonwood Heights, UT.  Those detection boxes were used to capture spatial waypoints from anonymous probe vehicle trips over the two-month period with reporting intervals ranging from 1 to 5 seconds.   The control delay was computed for each intersection using the difference between probe vehicle travel time and ideal travel time for vehicles travelling at the posted speed limit with no delay.  Corridor-level control delay statistics were prepared and identified two intersections in the northern part of the corridor that had the highest control delay.  High-resolution traffic signal controller data was analyzed for those intersections to confirm split failures as a source of delay.    Although the penetration rate was found to be only 0.3% during the study period, this paper demonstrates the viability for using high fidelity commercial probe vehicle trajectory data to generate control delay measures.  The virtual detection box methodology described in this paper provides a framework for government agencies to specify to commercial probe data vendors a method of providing true intersection control delay, without encountering privacy issues or large data management issues, because no actual trajectory data would need to be transferred to agency computers. 

There is extensive literature on the use of traffic signal system performance measures for tuning state-of-the-art signal systems. However, little has been written on performance measures that should be used during re-construction and infrastructure modernization. For agencies to be successful in their infrastructure modernization initiatives, there is a need to focus on a small set of performance measures, to ensure fundamental infrastructure and systems are operating before contract pay items are accepted. 

Using a case study for illustration, this paper presents a series of communication, detection, and signal operation performance measures that should be incorporated into pay items for contractors. Although a few are operational in nature, including them as pay items will ensure that contractors devote reasonable attention to maintenance of signal equipment during construction and that the system has functional performance measures when construction is completed. 

Contraflow left-turn lane (CLL) design has been increasingly used in China to help relieve traffic congestions associated with the left-turning vehicles at signalized intersections. With the CLL design, left-turn lanes are setup in the opposing exit lanes that are adjacent to conventional left-turn lanes. The basic idea of the CLL design is to provide additional capacity to left-turning vehicles by making use of the opposing exit lanes. The primary objective of the present study was to propose an actuated signal control strategy to improve the operations of the CLL design at signalized intersections. The proposed control strategy introduced an overlap phase between the pre-signal and the main signal at the signalized intersections with the CLL design. Two detectors were installed: one at the upstream median opening and the other at the contraflow lanes to cooperate with the proposed actuated control strategy. We also proposed a procedure for optimizing the distance between the upstream median opening and the main signal, such that the discharge rate of the left-turning vehicles and the utilization rate of the contraflow lanes can be maximized. Simulation experiments were conducted to compare the operational performance associated with the proposed actuated control strategy and those with the conventional fixed-time control strategy. The results show that the proposed design concept outperforms the conventional fixed-time control strategy in increasing capacity and reducing delay for the left-turning vehicles.

Probe vehicle technology is an effective way of collecting traffic trajectory data needed for system evaluation and optimization. This paper focus on arterial traffic signal coordination performance evaluation using probe vehicle data under the situation of low probe market penetration and low sampling frequency. The Arterial Traffic Signal Coordination Index proposed in this paper is established with a synthetic consideration of stops and running speed using vehicle trajectory data. Various numerical results are documented to show how evaluation errors behave by different market penetration and sampling frequency, which reflect the rationality and adaptation of this evaluation method. On this basis, a two-level method is proposed to meet the challenge of low probe market penetration and sampling frequency. By incorporating historical data, the evaluation error caused by low probe market penetration can be gradually reduced. Besides, Ordinal Regression Model is introduced to predict the number of stops of probe vehicles with low sampling frequency, which can also significantly improve the accuracy of evaluation result. The effectiveness of enhenced method is validated based on simulation. According to the analysis results, the improved method has satisfactory evaluation results under the situation of low probe market penetration and low sampling frequency. Even in the worst case that probe market penetration is only 5% and sampling interval reaches 20 seconds, the evaluation error of improved method will not exceed 10%.

Queue length is vitally important for signal control optimization and congestion management of urban arterials. Due to the existence of mutual influence of neighboring intersections, cycle-by-cycle estimation of queue length under spillover conditions remains a challenging problem given low-resolution detector data. On most urban arterials in China, point detectors are commonly installed at the intersection approaches and the upstream segments, providing volume, occupancy and speed data at a time interval of 1 min, i.e., a double-section low-resolution detection environment. Thus, the objective of this study is to propose a method to estimate the cycle-based queue lengths at signalized intersections considering spillover conditions, in the context of the above detection environment. Considering dense traffic conditions, uniform distribution is used to transform the low-resolution volume data into individual vehicle arrivals on a basis of second. To deal with spillover conditions, mutual effects of neighboring intersections are classified into four different cases, based on the offsets between the neighboring intersections. Detector data at the upstream intersection approach are used to modify the demand of the downstream intersection when long queue occurs, and the effect of spillover on the queue length of the upstream intersection can thus be formulated analytically using the shock-wave theory. Finally, the queue length is calculated by accumulating queue lengths of all offset conditions within a cycle. Empirical results show that the accuracy of the proposed method can reach 90% and the method is more applicable given low-resolution data input as compared with existing methods.

Development of signal timing plans and definition of their operating days is one of the ways to deal with traffic fluctuations. While the analysis of time of day (TOD) break points have been widely documented, we do not know how many peak-period signal plans are needed to cope with day-to-day traffic fluctuations. Once number of plans is defined how do we know when to deploy them? If the traffic varies wildly from day to day is it still good to develop new plans or should we look into adaptive traffic control solutions? This paper investigates these questions, which have been unaddressed so far. Furnished with a large data set from a 5-mile corridor in Fort Lauderdale, FL we analyze traffic profiles using 15-minute volume and travel time data over a two-year period. $K$-means clustering algorithm is applied to extract distinctive traffic profiles for a morning peak-hour. Traffic signal plans are developed for each of representative profile days by using Vistro. Results show that clustering of volume and travel time lead to the different sets of representative days. In addition, generated signal plans for these days differ from each other. More importantly, the compactness of revealed traffic patterns suggests that three additional signal timing plans may be appropriate solutions for 80% of the days in a year. Future research should consider further development of this idea and its integration in a decision-support system. Also, a cost-benefit analysis of developing signal plans versus deploying adaptive control may be in order.

The purpose of this paper is to discuss the results of an investigation into validity of the three-year rule for retiming traffic signals.  The goal of the research was to determine if the three-year rule is reasonable or if a different interval is more appropriate.  

This investigation leveraged real-world traffic data archived by the City of Durham, NC dating back to 1991 in simulation models to evaluate the performance of traffic signal timing in ever-changing, long term conditions.  Using eight coordinated networks and considering a wide range of costs per intersection for retiming, the research determined preferred retiming intervals. 

It was determined that preferred retiming intervals are sensitive to both retiming cost and network type, though those factors alone were insufficient to predict preferred intervals.  The three-year rule, or any other single, universal retiming interval is unlikely to serve as a reasonable guide for when to retime signals.

The main performance measures used in the assessment and optimization of traffic signal timing have traditionally been restricted to mobility, with limited consideration of safety and emissions. The goal of this study is to develop a signal timing optimization method that can consider mobility, safety, and environmental measures simultaneously. The objectives of the research are to: a) identify a set of equations that can be used to predict crashes as a function of intersection and signal control characteristics; b) develop a methodology for optimizing signal control in terms of safety (crashes), environmental impacts (emissions) and  mobility (delay); c) ; and d) conduct a sensitivity analysis to gain an understanding of the interactions among the influencing variables on performance and the tradeoffs between safety, mobility, and emissions.

Optimization results and statistical analysis of the sensitivity scenarios showed that the effect of each variable on the overall performance of the model is highly dependent on the combination of other variables. For instance, the use of specific control decisions could improve mobility, safety and emissions to a degree which is a function of intersection size, traffic volumes, and turning percentage levels. The size of the intersection, defined by the number of lanes on the arterial, was found to be the most significant variable, largely affecting all performance measures.

Keywords: Urban arterials, signalized intersections, mobility, safety, emissions, optimization

Performance measures are essential for managing transportation systems, including signalized corridors. Coordination is an essential element of signal timing, enabling reliable progression of traffic along corridors. Improved progression leads to less user delay, which leads to user cost savings and lower vehicle emissions. This paper presents a comparative study of signal coordination assessment using four different technologies. These technologies include detector-based high-resolution controller data, Bluetooth/Wi-Fi sensors, segment-based probe vehicle data, and automated vehicle location data consisting of GPS-based vehicle trajectories, representing the data anticipated from emerging connected vehicle technologies. The data were compiled for a 4.2-mile corridor in Holland, Michigan. The results show that all of the data sources were able to identify, at some level, where coordination issues existed. Detector-based controller data and GPS-based vehicle trajectory data were capable of showing greater detail, and could be used to make offset adjustments. The paper concludes by demonstrating the identification of signal coordination issues with the use of visual performance metrics incorporating connected vehicle trajectory data.

This study evaluates two groups of methods to model traffic signal operations in microscopic simulation: Hardware-In-the-Loop Simulation (HILS) and Software-In-the-Loop Simulation (SILS). These methods have become standards for accurate modeling of traffic signal operations but in spite of large number of available options there were not studies that have conducted relevant comparative evaluations. This study bridges this gap by investigating signal timing and operational differences of these two methods in basic actuated operations of a single signalized intersection. The emphasis is given to broad examination of various platforms as opposed to more complex experiments be done with individual platforms. A representative number of 65-minute simulation runs were executed for each experimental scenario. The results showed that differences between various HILS and SILS platforms are large enough that one cannot confidently switch between the platforms without impacting the final outcomes. The study confirmed previous findings about impact that the initialization process has on the final outcome of the signal and traffic performance. These findings put a significant restriction on how various HILS and SILS platforms, either alone or in conjunction with other higher forms of traffic control strategies, can be used in joint fashion. Future research should address higher-complexity experiments with SILS and should attempt to control initialization process to reduce its impact on the stochasticity of the results.

Automated Traffic Signal Performance Measures (ATSPMs) are an effort to equip traffic signal controllers with high-resolution data logging capabilities and utilize this data to generate performance measures. These measures allow practitioners to improve operations, maintaining and operating their systems in a safe and efficient manner.  While these measures have really changed the way operators manage their systems, several shortcomings of the tool, identified by talking with signal operators, are a lack of data quality control and that properly using the tool for system-wide management is resource intensive.  To address these shortcomings, this paper presents Intelligent Traffic Signal Performance Measurements (ITSPM), using the concepts of machine learning, traffic flow theory and data visualization to reduce the operator resources needed for data-driven traffic signal management systems.  By applying these concepts, ITSPM presents graphical tools to identify and remove logging errors and data from bad sensors, intelligently determine trends in demand, and address the question of whether or not coordination may be needed at an intersection.  The focus of ATSPM and ITSPM on performance measures for multimodal users is identified as a pressing need for future research.

Eco-driving control systems aim to reduce the fuel consumptions by optimizing the vehicle trajectories near the signalized intersection using the signal phase and timing (SPaT) information.  This paper presents a framework for developing an eco-driving application for semi-actuated signals. The proposed algorithm takes into consideration the queue effects due to traditional and connected/automated vehicles. The algorithm can easily be implemented with the installment of additional vehicle detector 300 m upstream of each minor approach. Results showed that the estimated fuel consumption for vehicles guided by the developed model was about 29.2% less than that of the case with no guidance. Moreover, a sensitivity analysis was performed to assess the impact of market penetration (MP), and results indicated that for MP the saving in fuel consumption increases with the increase in MP. Furthermore, when MP is greater than 50%, the model provides appreciable savings in the travel time. Furthermore, the estimated acceleration noise for the guided case was less than that of the case with no guidance by about 21.9%. This reduction in fuel consumption and acceleration noise is an indicator for reducing vehicle emissions and improving traffic safety. This finding is considered promising for potential applications of the developed model and further enhancement of its features.

Since the 2000’s, researchers have been designing and examining the potential of visualizing high-resolution signal and vehicle trajectory data for the purpose of gaining better insight into traffic control system performance, corridor progression along with the other system measures of effectiveness. Along with technological advancements, obtaining finer granularity data became achievable. Visualizing relevant signal performance data in an easy-to-understand format is crucial when identifying problem areas and/or causes of problems. Furthermore, visualizing of now-obtainable information needs to be accompanied by a tangible and robust performance metric. Traditional measures of effectiveness may not suffice this need. Therefore, to analyse the additional information provided by the high-resolution data alternative performance measures are required. The proposed rate of utilization of available green time and space incorporates multiple aspects of signal performance assessment: total delay, progression speed, quality of progression, as well as the impact of oversaturation. It can be applied at an individual vehicle level, measuring how well the vehicle is utilizing time and space under specific signal control operational conditions, or any other more aggregate level. The concept utilizes vehicle-trajectory and signal phase status and duration data to ascertain the responsiveness of trajectory-based measures when different control strategies are applied. To develop and demonstrate the concept, the study uses simulation data in a format that corresponds to high resolution data - signal status and vehicle positions at a one-second frequency.