Boosting Max-Pressure Signal Control Into Practical Implementation: Methodologies And Simulation Studies In City Networks

Abstract

This dissertation presents innovative modifications to the Max-Pressure (MP) control policy, an adaptive traffic signal control strategy tailored to various urban traffic conditions. The max-pressure control offers two pivotal advantages that underscore its significance for in-depth research and future implementation: Firstly, MP operates on a decentralized basis, enabling real-time solutions. Secondly, MP control guarantees maximum stability, implying it can accommodate as much given demand as any alternative signal timing strategy. Initially, the MP control policy was adapted to transit signal priority (MP-TSP). It delivered enhanced bus travel times, outperforming both fixed-time signal controls with TSP and other adaptive signal controls in efficiency. Subsequently, the pedestrian-friendly max-pressure signal controller (Ped-MP) was developed. This marked a pioneering effort in crafting an MP control to boost pedestrian access without compromising vehicle throughput. The Ped-MP, backed by analytical proof for maximum stability, illustrated an inverse relation between pedestrian delay and tolerance time during simulations on the Sioux Falls network. This suggests the potential for urban spaces that are more pedestrian-oriented, even in areas of elevated pedestrian traffic. The third innovation addressed the practical feasibility of the position-weighted back-pressure (PWBP) controller. Although the initial PWBP controller was effective in simulations, it was found to be impractical due to its need for density information from everywhere of the road link. This observation paved the way for the approximate position-weighted back-pressure (APWBP) control, which significantly reduces sensor requirements by utilizing only two loop detectors per link (one downstream and one upstream). A comparative analysis revealed that the APWBP's efficacy closely paralleled the original PWBP, validating its practicality. Finally, recognizing the MP controller’s deficit in coordinated phase selection, the Smoothing-MP approach was conceptualized. Incorporating signal coordination, this novel strategy not only maintained its maximum stability properties but also amplified traffic flow efficiency, as confirmed by mathematical proofs and numerical studies in both the Grid Network and the Downtown Austin Network.