Date of Award
Master of Science in Civil Engineering (MSCE)
Civil and Environmental Engineering
Yaser E. Hawas
Dr. Younes Hamdouch
Setting the traffic controller parameters to perform effectively in real-time is a challenging task, and it entails setting several parameters to best suit some predicted traffic conditions. This study presents the framework and method that entail the application of the Response Surface Methodology (RSM) to calibrate the parameters of any control system incorporating advanced traffic management strategies (e.g., the complex integrated traffic control system developed by Ahmed and Hawas). The integrated system is a rule-based heuristic controller that reacts to specific triggering conditions, such as identification of priority transit vehicle, downstream signal congestion, and incidents by penalizing the predefined objective function with a set of parameters corresponding to these conditions. The integrated system provides real time control of actuated signalized intersections with different phase arrangements (split, protected and dual).
The premise of the RSM is its ability to handle either single or multiple objective functions; some of which may be contradicting to each other. For instance, maximizing transit trips in a typical transit priority system may affect the overall network travel time. The challenging task is to satisfy the requirements of transit and non-transit vehicles simultaneously.
The RSM calibrates the parameters of the integrated system by selecting the values that can produce optimal measures of effectiveness. The control system was calibrated using extensive simulation-based analyses under high and very high traffic demand scenario for the split, protected, and dual control types.
A simulation-based approach that entailed the use of the popular TSIS software with code scripts representing the logic of the integrated control system was used. The simulation environment was utilized to generate the data needed to carry on the RSM analysis and calibrate the models.
The RSM was used to identify the optimal parameter settings for each control type and traffic demand level. It was also used to determine the most influential parameters on the objective function(s) and to develop models of the significant parameters as well as their interactions on the overall network performance measures.
RSM uses the so-called composite desirability value as well as the simultaneous multi-objective desirabilities (e.g., the desirability of maximizing the transit vehicles throughput and minimizing the average vehicular travel time) estimates of the responses to identify the best parameters. This study also demonstrated how to develop “mathematical” models for rough estimation of the performance measures vis-à-vis the various parameter values, including how to validate the optimal settings. The calibrated models are proven to be significant.
The optimal parameters of each control type and demand level were also checked for robustness, and whether a universal set of relative parameter values can be used for each control type. For the high traffic demand level, the optimal set of parameters is more robust than those of the very high traffic demand. Besides, the dual actuated controller optimal setting under the very high traffic demand scenario is more robust (than other control types settings) and shows the best performance.
Alam, MD Didarul, "Calibrating and Evaluating Dynamic Rule-Based Transit-Signal-Priority Control Systems in Urban Traffic Networks" (2017). Civil and Environmental Engineering Dissertations. 1.