S. Ilgin Guler

Relationship between mean and day-to-day variation in travel time in urban networks

The day-to-day reliability of transportation facilities significantly affects travel behavior. To better understand how travelers use these facilities, it is critical to understand and characterize this reliability for different facilities. Early work in this area assumed that the variance of day-to-day travel times (a measure of the inverse of reliability) increases proportionally with the mean travel time; i.e., as the mean travel time increases, travel time reliability decreases. However, recent empirical data for a single bottleneck facility and a small urban network suggest a more complex relationship that exhibits hysteresis. When this phenomenon is present, the variance in travel time is larger as the mean travel time decreases (congestion recovery) than as the mean travel time increases (congestion onset). This paper presents an elegant theoretical model to describe the variance of travel times across many days in an urban network. This formulation shows that the hysteresis behavior observed in empirical floating car data on urban networks should not be unexpected, and that it is linked to the hysteresis loops that often exist in the Macroscopic Fundamental Diagram of urban traffic. To verify the validity of this formulation, data from a micro-simulation of the City of Orlando, Florida, are used to derive an observed relationship with which to compare to theory. The simulated data are shown to match the theoretical predictions very well, and confirm the existence of hysteresis in the relationship between the mean and variance of travel times that is suggested by theory. These results can be used as a first step to more accurately represent travel time reliability in future models of traveler decision-making.

On the impact of obstructions on the capacity of nearby signalised intersections

Various obstructions exist that can impede maximum vehicle flow through signalised intersections. Examples include buses or freight vehicles dwelling at loading areas near the intersection, stalled vehicles, pre-signals that temporarily block car traffic to provide bus priority, on-street parking manoeuvres and permanent road fixtures. If the effects of these obstructions are not recognised or accounted for, vehicle discharge capacities at these critical locations can be overestimated, leading to ineffective traffic management strategies. This paper examines the capacity of an isolated signalised intersection when a nearby roadway obstruction is present in either the upstream or downstream direction. To quantify the loss of capacity caused by an obstruction, the paper applies the variational theory of kinematic waves in a moving-time coordinate system, which simplifies the traditional variational theory by reducing the number of local path costs that must be considered. The result is a simple recipe that requires few calculations and can be used to gain insights into signal operations when obstructions are present. Capacity formulae for general cases are also developed from the recipe. The results, recipe and formulae can be used to guide policies on the location of obstructions that can be controlled, like bus stops, pre-signals or permanent road fixtures and to develop strategies to mitigate the effects of obstructions that can be identified in real time. As an example, a simple adaptive signal control scheme is created using this methodology to more efficiently allocate green time between competing directions when an obstruction is present.

Pre-signals for bus priority: basic guidelines for implementation

When compared to cars, public transportation (e.g., buses) can carry more people using less space. Hence, by increasing the share of people traveling by bus within an urban network, we can improve the efficiency of the urban transportation system, ultimately making it more sustainable. Unfortunately, buses operating mixed with cars can often get stuck in car congestion. One commonly used solution is to dedicate a lane for bus-use only. However, when bus flows are low, dedicated lanes running through intersections can reduce the discharge flows from these locations and lead to increased car delays, car queues, and all the negative externalities associated with congestion. This, in turn, can reduce the overall efficiency of the transportation network. Therefore, a solution is to discontinue the dedicated lane upstream of the main signal, removing bus priority at intersections. In this paper, we advocate the use of pre-signals upstream of signalized intersections to continue providing bus priority while minimizing the disruptions to car traffic. Pre-signals can allow buses to jump the car queues upstream of signalized intersections, while allowing cars to utilize the full capacity of the main signal when buses are not present. In this paper we provide practical guidelines on how to implement pre-signals at signalized intersections. Ideas on how to operate pre-signals are provided by using recent analytical and empirical findings from previous research on pre-signals. The reduction of system-wide (buses and cars) person hours of delay by using pre-signals, as compared to mixed-use lanes or dedicated bus lanes is also quantified. By doing so, the domains of application of pre-signals are also defined. This information can then be used to determine where and when pre-signals should be implemented in real urban networks and to quantify their benefits to the system.

Axle Load Power for Pavement Fatigue Cracking: Empirical Estimation and Policy Implications

Deterioration equivalence factors are an essential part of deterioration models used in pavement management systems. For accurate cost estimation, it is important to determine the correct axle load power. The data set from the AASHO Road Test was used to estimate a hazard rate function. The results showed that the axle load power for initiation of fatigue cracking was significantly higher than the power typically assumed. This result has implications for pavement deterioration charging schemes that are based on marginal cost pricing.

Effects of Multimodal Operations on Urban Roadways

Public transportation (e.g., buses) can provide more efficient urban transportation systems by carrying more people in the same space. A commonly used solution to prioritize this mode is to dedicate a lane for bus use only. However, the changes in system capacity are not clear for the use of a dedicated bus lane instead of completely mixed-use lanes. Even if the capacity of the dedicated bus lane were not fully used, this strategy could still increase car capacity in the remaining lanes in two ways: (a) buses traveling on a separate lane would eliminate conflicting bus maneuvers and (b) the reduced number of lanes available for cars could reduce the number of lane changes and could smooth traffic. This paper empirically analyzes differences in car capacity between (a) a mixed-use scenario without the influence of buses and a dedicated-lane scenario and (b) a mixed-use scenario with and without buses. Results show that with mixed-use lanes the car capacity per lane remains the same as compared with a dedicated-lane scenario. However, in mixed-use conditions, the presence of a bus in traffic flow can reduce capacity by 20%. Based on these findings, a simple analysis is carried out to compare passenger delay at urban signalized intersections with a mixed use versus dedicated bus lane. It is shown that passenger delay can be reduced with use of dedicated bus lanes if bus occupancies are relatively high or if car demand is low.

Estimating the Impacts of Bus Stops and Transit Signal Priority on Intersection Operations

Transit signal priority (TSP) can be used to improve bus operations at signalized intersections, often to the detriment of general car traffic. However, the impacts of TSP treatments applied to intersections with nearby bus stop locations are currently unknown. This paper quantifies changes in intersection capacity, car delay, and bus delay when priority is provided to buses that dwell at near- or farside bus stop locations through green extension or red truncation. Variational and kinematic wave theories are used to estimate car capacity and bus delay for oversaturated traffic conditions; queuing theory is used to estimate car and bus delays for undersaturated conditions. Numerical analyses are conducted to explore the impacts on various bus stop locations and bus dwell time durations. These results illustrate clear trade-offs between reduced bus delays and increased car delays or reduced intersection capacities that can be quantified with the proposed method. The results also reveal that the effects of TSP vary dramatically with bus dwell times for a given bus stop location. The proposed method and associated results can be used to implement TSP strategies to meet the specific needs of local agencies.

Implementing transit signal priority in a connected vehicle environment with and without bus stops

ABSTRACT Connected vehicles give more precise and detailed information on vehicle movements, thus can be beneficial to provide priority to public transportation. This paper proposes a transit signal priority algorithm using connected vehicle information for multimodal traffic control. The algorithm can also be adapted to scenarios with near-side or far-side bus stops. Moreover, it can minimize either signal delay or schedule delay for buses while minimizing additional car delays. Simulation is conducted for different volume to capacity ratios, bus arrivals, bus occupancies, and penetration rates. Results show that this algorithm successfully reduces the total passenger delay. It is also shown that this algorithm is not sensitive to the assumed bus passenger occupancy, nor the estimation of bus dwell time, hence does not require accurate information on these parameters. Overall, this algorithm seems rather promising as it significantly reduces the delay of buses with minimal increase to the delay of cars in the conflicting approach.

Optimizing Transit Signal Priority Implementation along an Arterial

Transit signal priority (TSP) is a common method of providing priority to buses at signalized intersections. The implementation of TSP can affect travel time of cars traveling in the same, opposite, and cross directions. The bus delay savings and car travel-time impacts are not expected to increase linearly when considering multiple intersections along an arterial. This paper quantifies the influence of TSP on arterials with dedicated bus lanes considering an arterial-wide approach utilizing variational theory. Existing tools were modified to quantify the change in capacity along an arterial where TSP was implemented and it was shown that this effect was negligible. In addition, the bus delay savings and cross-street capacity losses were determined. Case studies provided insights into the influence of TSP among different network homogeneities and bus frequencies. Using these tools, an optimization framework was developed to determine where to implement TSP along an arterial to maximize the marginal benefits, or minimize marginal costs. In addition, a comparison of evaluating an arterial as a sum of isolated intersections as opposed to evaluating an arterial as a whole is presented. This analysis indicates the necessity of the arterial-based method in considering TSP impacts along corridors.

Signal Timing Optimization with Connected Vehicle Technology: Platooning to Improve Computational Efficiency:

This paper develops a real-time traffic signal optimization algorithm in the presence of connected and autonomous vehicles (CAVs). The proposed algorithm leverages information from connected vehicles (CVs) arriving at an intersection to identify naturally occurring platoons that consist of both CVs and non-CVs. Signal timings are then selected to optimize the sequence at which these platoons are allowed to discharge through the intersection to minimize total vehicle delay. Longitudinal trajectory guidance that explicitly accounts for vehicle acceleration and deceleration behavior is provided to the lead autonomous vehicle (AV) in any platoon to minimize the total number of stopping maneuvers performed by all vehicles. Simulation tests reveal that the proposed platoon-based algorithm provides superior computational savings (over 95%) compared with a previously developed algorithm that focuses on optimizing departure sequences of individual vehicles, with negligible changes in operational performance. The computational savings allow the platoon-based algorithm to accommodate intersections with four multi-lane approaches and left turns, whereas large computational costs limited the previous vehicle-based algorithm to only two single-lane approaches without conflicting left turns. Additional simulation tests of the platoon-based algorithm on these more realistic intersection configurations show that intersection performance increases as the penetration rate of CAVs in the vehicle fleet increases. However, the marginal benefits decrease rapidly after the fleet is composed of 40% CAVs.

Pavement Friction Degradation Based on Pennsylvania Field Test Data

Pavement surface–tire friction is a critical safety element associated with roadway design, construction, and maintenance practices. The skid resistance of pavements generally declines over time and increases the risk of skidding-related crashes. On horizontal curves, lateral friction may be associated with lane-departure incidents, particularly as the pavement ages and drivers demand more lateral friction than the pavement surface–tire interaction can supply. On tangent roadway sections, longitudinal friction affects braking distances. As the skid-resistance properties of a pavement surface decline over time, braking distances increase, and may increase risks to driver safety. A comprehensive understanding of the process of pavement friction degradation could help highway agencies identify roadway segments that need maintenance to reduce the probability of skid-related incidents. This paper presents a survival analysis of friction degradation for asphalt pavement surfaces. Duration models were estimated with data collected annually along an Interstate highway in Pennsylvania to investigate the degradation of friction over time. These models consider traffic volume and roadway features to determine the probability that friction levels will remain above various friction thresholds. The resulting statistical models can help transportation agencies make better decisions about pavement maintenance to reduce safety risk.