Investigation of policies for arterial street operations

Traffic signal operations are often the determining factor in the functioning of urban street systems. Efficient signal control policies help in improving mobility and in reducing congestion in an urban area. They allow effective utilization of existing resources, which is particularly relevant in today's funding climate. Proper design and evaluation procedures are essential in optimizing the performance of signal systems. In this dissertation we outline deficiencies in existing procedures and recommend alternative ways for performance estimation of signalized intersections. We further extend the performance estimation functions for optimal coordination of arterial signal systems using Dynamic Programming.;The Highway Capacity Manual (HCM 2000), which is prepared and published by the Transportation Research Board (TRB), provides the most widely used procedures for Level-of-Service (LOS) analysis. LOS at a signalized intersection is determined by the delay incurred by the approaching vehicles. The HCM prescribes the use of Progression Adjustment Factors that modulate control delay at a signalized intersection approach in accordance with the quality of coordination. These factors correspond to six discrete arrival types that need to be estimated from given or observed flow data. This procedure is inaccurate and may lead to incorrect determination of Level of Service. This thesis introduces an improved procedure for the determination of quality of progression. The procedure is sensitive to different traffic conditions, can be easily calibrated and produces more accurate results. We develop a periodic, continuously variable Link Performance Function (LPF) termed Cyclic Coordination Function (CCF). This function measures delay or travel time as a function of offsets along an urban street segment and depends on a variety of factors, including: traffic flow characteristics, link physical characteristics, and traffic signal controls. Being periodic with the cycle time it can be modeled as a Fourier Series consisting of a sum of harmonics. A few harmonics provide good approximations to the original functions. We show how to obtain the principal harmonics from the underlying traffic, link and signal data resulting in a simple and more accurate procedure for quality of progression estimation.;Performance estimation at a signalized intersection is then extended to urban streets. The LOS for urban streets is based on the average travel speed on the arterial and the urban street classification. The average travel speed is computed from the length of the segments and the running times, which includes the control delay of through movements at signalized intersections. In this dissertation we discuss several approaches for urban street performance estimation and effects of progression schemes on the arterial. We utilize the previously developed Cyclic Coordination Function to quantify the quality of progression in conjunction with the HCM procedure and to analyze the LOS on an urban street. Both one-way and two-way scenarios are considered. Additionally, we introduce the Progression Potential Frontier concept to quantify tradeoffs in performance of traffic control schemes on an arterial street considering one-way and two-way progressions.;An analysis is also presented of the impact of offsets on successive links on delay at the downstream intersection. When analyzing an arterial street, the delay on any link is dependent on the offset on that link and, in addition, may also be dependent on offsets on previous links. We studied the impact of offsets on the previous link (i.e., a feeder link) on the delay at an intersection, thus obtaining a three-dimensional Link Performance Function (LPF). The characteristic of the 3D LPF depends in a significant way on v/c ratio: for high v/c ratios (>0.90) delay depends only on a single offset. For lower v/c ratios it also depends on the previous offset. Criteria were developed for determining when dependence on prior-link offsets should be considered and its effect on the intersection delay estimation.;The LPFs developed in this thesis were also extended for signal coordination and synchronization. The Cyclic Coordination Functions are used in conjunction with an innovative Dynamic Programming (DP) model for optimizing signal settings along an arterial street. This model provides a rigorous procedure for signal coordination and synchronization.