The fitting of roundabout model with gradient-based minimization

• Main Authors: Cheah, Yuat Hoong, Yeak, Su Hoe
• Format: Conference or Workshop Item
• Language: English
• Published: 2023
• Subjects: QA Mathematics
• Online Access: http://eprints.utm.my/107859/1/YeakSuHoe2023_TheFittingofRoundaboutModelwithGradient.pdf; http://eprints.utm.my/107859/; http://dx.doi.org/10.1088/1742-6596/2609/1/012003
• Institution: Universiti Teknologi Malaysia

The objective of this research is to create an appropriate roundabout model for all countries. To date, the four-arm roundabout macroscopic model has been created. In order to ensure the feasibility and suitability of the model for all countries, the fitting process must be implemented since the speed of vehicles varies in each country. Thus, the parameter estimation on the rate of exiting roundabout is to be determined because the mean speed of vehicles is related to the rate of exiting the roundabout. In the minimization process, we have proposed an efficient and reliable framework as it includes the calculation of gradients used in minimization so called the user supplied-gradient minimization, as compared to non-user supplied-gradient minimization. The including of the calculation of gradient is to produce more accurate results by the built-in MATLAB minimization routine for parameter fitting. In this research, five pseudo experiments with numerous parameters are carried out. The rate of exiting the roundabout is set initially in order to compute the Total Travel Time and Total Waiting Time. The simulation showed a highly converged and accurate solution by the user supplied-gradient minimization. Lastly, this parameter estimation can be implemented that will enable the roundabout model to be applied worldwide if there is actual data for Total Travel Time and Total Waiting Time.