Abstract
With increasing congestion and associated challenges to manage the transportation network, intelligent transportation systems (ITS) have gained popularity due to their data-driven approach and application of advanced technologies. A variable speed limit (VSL) is a popular ITS-based solution which uses dynamic speed limit to promote harmonization along a corridor. However, not much was done in identifying road links and influencing variables for their applicability. Therefore, this paper focuses on examining road link-level data to identify road links and variables influencing the applicability of VSL signs. A multivariate cluster analysis was first used to identify potential road links susceptible to speed variation for the implementation of VSL. A supervised machine learning algorithm, forest-based classification and regression, was then used to model and examine the influence of average annual daily traffic (AADT), historical speed of the road link, and the speeds of upstream and downstream road links on the average speed of the corresponding road link. Modelling and validation were performed using data for Mecklenburg County, North Carolina, USA, for all the road links as well as for road links with low- and high-speed variation.
Keywords:
- Intelligent transportation systems, Variable speed limit, Supervised machine learning.
How to Cite:
Duvvuri, S., Mathew, S., Gouribhatla, R., Pulugurtha, S., Duvvuri, S., Mathew, S., Gouribhatla, R., Pulugurtha, S., Duvvuri, S., Mathew, S., Gouribhatla, R. & Pulugurtha, S.,
(2020) “Identifying Road Links and Variables Influencing the Applicability of Variable Speed Limits Using Supervised Machine Learning and Travel Time Data”,
Journal of Modern Mobility Systems 1,
125-130.
doi: https://doi.org/10.13021/jmms.2020.2931