Traffic flow modelling of long and short trucks using a hybrid artificial neural network optimized by particle swarm optimization

Oyeyemi Olayode, Isaac; Du, Bo; Kwanda Tartibu, Lagouge; Justice Alex, Frimpong · 2024 · Crossref

DOI: 10.1016/j.ijtst.2023.04.004

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the critical need for accurate traffic flow modeling of long and short trucks, which are major contributors to congestion and accidents on freeways. Despite the growing volume of truck traffic globally, limited research has focused on predicting their specific flow patterns using advanced machine learning techniques. The authors aim to fill this gap by developing a hybrid model that combines Artificial Neural Networks (ANN) with Particle Swarm Optimization (PSO) to analyze truck traffic dynamics. The research is motivated by the increasing prevalence of trucks in supply chains and the resulting safety and efficiency challenges on road networks, particularly in developing countries with sophisticated data collection infrastructure like South Africa. The methodology involves collecting real-world traffic data from the N1 freeway in South Africa, one of the busiest routes in the region. Data were gathered in 2019 using inductive loop detectors and video cameras, focusing on peak hours (6 AM to 11 AM) across weekdays. The dataset comprised 920 records of traffic variables, including speed, time, density, and volume, which were normalized and split into 70% for training, 15% for testing, and 15% for validation. The authors formulated a mathematical model based on traffic speed and density equations to describe truck movement. They then developed the ANN-PSO hybrid model in MATLAB, where PSO was used to optimize the ANN’s parameters, such as the number of neurons and accelerating factors, leveraging PSO’s ability to converge quickly to optimal solutions with minimal adjustment parameters. The results demonstrate that the ANN-PSO model effectively models the traffic flow of long and short trucks. The model achieved high accuracy, with an R² value of 0.9990 for training and 0.9930 for testing, indicating a strong fit between predicted and actual traffic conditions. The analysis revealed distinct traffic patterns, identifying periods of high traffic volume and density (on-peak) versus low volume and free-flowing traffic (off-peak). The study also provided a longitudinal analysis of truck traffic variables, offering insights into how truck density and speed interact on the freeway. This is noted as the first study to apply a metaheuristic algorithm like ANN-PSO for the longitudinal analysis of long and short truck traffic flow on a freeway. The significance of this research lies in its contribution to intelligent transportation systems and traffic management. By providing a robust model for predicting truck traffic flow, the study offers valuable insights for transportation planners and engineers aiming to minimize truck-related accidents and congestion. The findings support the use of hybrid machine learning models in transportation engineering, demonstrating their superiority over traditional parametric models in handling complex, non-linear traffic data. The study underscores the importance of integrating advanced AI techniques with real-time data to enhance the safety and efficiency of freight transportation networks.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• traffic density