Date of Award

11-2015

Document Type

Thesis

Degree Name

Master of Science in Civil Engineering (MSCE)

Department

Civil Engineering

First Advisor

Prole sor Yaser Ha> as

Second Advisor

Dr. Moza Tahnoon AI Nahyan

Third Advisor

Dr. Kamran Ahmed.

Abstract

Different traffic signal control systems are applied nowadays. They vary in complexity and approach but mostly aimed at improving the intersection level of service and the efficiency of the traffic network. Fuzzy Logic Models (FLM) have been widely regarded as quite appealing for real-time applications, in addition to their explicit considerations for stochasticities or uncertainties of traffic measures. The majority of the FLMs for signal control were developed to handle isolated intersections in specific. Very few of these FLM systems explicitly account for neighbor intersections’ traffic impact on the decisions to be made at the underlying intersections.

The aim of this research is to develop a fuzzy logic control system for real time signal control that accounts for various traffic conditions. The system is envisaged to demonstrate signal control settings close to those developed by optimization methodologies.

The literature review highlighted the structure of the existing FLMs for traffic signal control, input and output parameters and testing. The simulation environment (SYNCHRO) was utilized to model and design the traffic signals of an isolated intersection using 289 different traffic configurations and traffic volumes. The cases were carefully selected to cover the domain of different levels of service along the competing approaches. A FLM was developed and calibrated to emulate the Highway Capacity Manual (HCM) method for optimal signal control. The general outcome was that the proposed FLM is effective in replicating “optimization” signal control procedures (SYNCHRO).

The devised FLM was thoroughly evaluated against the benchmark solutions of optimal cycle times, green times of the various phases. Further analysis was carried out to validate the devised FLM and to assess its effectiveness if deployed in conditions other than those conditions used model calibration.

The validation results indicated that the FLM is mostly effective for cases of medium or high downstream congestion and medium to high traffic flows

Download
COinS