Predictive Modeling of Traffic Density for Adaptive Urban Road Lighting Applications

Accurate prediction of traffic density is a key requirement for adaptive road lighting systems that aim to improve energy efficiency without compromising road safety. Conventional lighting infrastructure operates on fixed schedules and maintains constant illumination throughout the night, resulting in unnecessary energy consumption during low-traffic hours. This study proposes a machine learning-based predictive modeling approach that uses artificial neural networks and extreme gradient boosting to forecast hourly traffic volumes and to inform dynamic lighting control. Traffic data recorded at two-minute intervals on Cendere Avenue in Istanbul between 2015 and 2017 were cleaned, aggregated to hourly totals, and used to train both models, with the dataset randomly divided into training and test subsets containing 70% and 30% of the observations, respectively. Candidate hyperparameter configurations were compared using the coefficient of determination and the mean absolute error. The best neural network, a four-layer structure with twenty-five neurons in each layer, reached a coefficient of determination of 0.766 and a mean absolute error of 261.3 vehicles per hour on the test set, whereas the best boosting model, with a maximum tree depth of five, one hundred trees, and a learning rate of 0.2, reached 0.833 and 221.3, respectively. Gradient boosting therefore provides higher accuracy than the neural network, making it the more suitable option for real-time deployment. The resulting forecasts provide a basis for lighting control strategies that reduce illumination in advance of periods of low expected demand. The findings indicate that traffic-aware lighting can lower operating costs and support sustainable urban development. Future work will incorporate environmental variables, such as weather and visibility, along with real-time data acquisition from sensor networks.

Authors

Institutions

Publication Details

Journal
Black Sea Journal of Engineering and Science
Published
2026-11-15
DOI
https://doi.org/10.34248/bsengineering.2013917
Primary Topic
Traffic Prediction and Management Techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Predictive Modeling of Traffic Density for Adaptive Urban Road Lighting Applications

M. Berker Yurtseven
Black Sea Journal of Engineering and Science
Traffic Prediction and Management Techniques
article

Predictive Modeling of Traffic Density for Adaptive Urban Road Lighting Applications

M. Berker Yurtseven
article en

Abstract

Accurate prediction of traffic density is a key requirement for adaptive road lighting systems that aim to improve energy efficiency without compromising road safety. Conventional lighting infrastructure operates on fixed schedules and maintains constant illumination throughout the night, resulting in unnecessary energy consumption during low-traffic hours. This study proposes a machine learning-based predictive modeling approach that uses artificial neural networks and extreme gradient boosting to forecast hourly traffic volumes and to inform dynamic lighting control. Traffic data recorded at two-minute intervals on Cendere Avenue in Istanbul between 2015 and 2017 were cleaned, aggregated to hourly totals, and used to train both models, with the dataset randomly divided into training and test subsets containing 70% and 30% of the observations, respectively. Candidate hyperparameter configurations were compared using the coefficient of determination and the mean absolute error. The best neural network, a four-layer structure with twenty-five neurons in each layer, reached a coefficient of determination of 0.766 and a mean absolute error of 261.3 vehicles per hour on the test set, whereas the best boosting model, with a maximum tree depth of five, one hundred trees, and a learning rate of 0.2, reached 0.833 and 221.3, respectively. Gradient boosting therefore provides higher accuracy than the neural network, making it the more suitable option for real-time deployment. The resulting forecasts provide a basis for lighting control strategies that reduce illumination in advance of periods of low expected demand. The findings indicate that traffic-aware lighting can lower operating costs and support sustainable urban development. Future work will incorporate environmental variables, such as weather and visibility, along with real-time data acquisition from sensor networks.

Black Sea Journal of Engineering and ScienceVol. 9(6)
Istanbul Technical University (TR)
Traffic Prediction and Management Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon