Loading...
Publication
DTIP: a scalable pipeline for traffic congestion detection using floating car data
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| Dissertação de mestrado_41805 | 1.15 MB | Adobe PDF |
Advisor(s)
Abstract(s)
O congestionamento do tráfego urbano continua sendo um obstáculo crítico para a mobilidade, segurança e sustentabilidade nas cidades modernas. Apresentamos o Distributed Traffic Intelligence Pipeline (DTIP), um sistema modular e interpretável projetado para estimar níveis de congestionamento a partir de Floating Car Data (FCD), ou seja, trajetórias de veículos baseadas em Global Positioning System (GPS), e para apoiar a validação de relatórios de
perigos relacionados ao tráfego.
O framework proposto integra ferramentas de código aberto para processamento de dados, “map-matching” e extração de características, culminando num modelo de aprendizagem supervisionado baseado em Extreme Gradient Boosting (XGBoost). O modelo foi treinado com dados de Vila Nova de Gaia, Portugal, e alcançou um F1-score ponderado acima de 97%, distinguindo com sucesso quatro classes de severidade de congestionamento.
Para avaliar ainda mais a plausibilidade das suas previsões, uma camada de simulação qualitativa utilizando Simulation of Urban Mobility (SUMO) foi incorporada. Os resultados da simulação alinharam-se bem com as saídas do modelo na maioria dos cenários de teste, reforçando a validade comportamental das estimativas de congestionamento.
Desenvolvido com escalabilidade e implantação de baixa latência em mente, o DTIP oferece uma contribuição prática para o desenvolvimento de sistemas de monitoramento de tráfego urbano transparentes e eficientes. A sua natureza aberta e modular o torna adequado para adaptação a outras cidades ou para integração futura em infraestruturas de suporte a decisões em tempo real.
Urban traffic congestion remains a critical obstacle to mobility, safety, and sustainability in modern cities. We present Distributed Traffic Intelligence Pipeline (DTIP), a modular and interpretable system designed to estimate congestion levels from Floating Car Data (FCD), i.e., Global Positioning System (GPS) based vehicle trajectories, and to support the validation of trafficrelated hazard reports. The proposed framework integrates open-source tools for data processing, map matching and feature extraction, culminating in a supervised learning model based on XGBoost. The model was trained on data from Vila Nova de Gaia, Portugal, and the system achieved a weighted F1-score above 97%, successfully distinguishing between four classes of congestion severity. To further assess the plausibility of its predictions, a qualitative simulation layer using Simulation of Urban Mobility (SUMO) was incorporated. The simulation results aligned well with the model’s output in most test scenarios, reinforcing the behavioral validity of the congestion estimates. Designed with scalability and low-latency deployment in mind, DTIP offers a practical contribution to the development of transparent and efficient urban traffic monitoring systems. Its open and modular nature makes it suitable for adaptation to other cities or future integration into real-time decision-support infrastructures.
Urban traffic congestion remains a critical obstacle to mobility, safety, and sustainability in modern cities. We present Distributed Traffic Intelligence Pipeline (DTIP), a modular and interpretable system designed to estimate congestion levels from Floating Car Data (FCD), i.e., Global Positioning System (GPS) based vehicle trajectories, and to support the validation of trafficrelated hazard reports. The proposed framework integrates open-source tools for data processing, map matching and feature extraction, culminating in a supervised learning model based on XGBoost. The model was trained on data from Vila Nova de Gaia, Portugal, and the system achieved a weighted F1-score above 97%, successfully distinguishing between four classes of congestion severity. To further assess the plausibility of its predictions, a qualitative simulation layer using Simulation of Urban Mobility (SUMO) was incorporated. The simulation results aligned well with the model’s output in most test scenarios, reinforcing the behavioral validity of the congestion estimates. Designed with scalability and low-latency deployment in mind, DTIP offers a practical contribution to the development of transparent and efficient urban traffic monitoring systems. Its open and modular nature makes it suitable for adaptation to other cities or future integration into real-time decision-support infrastructures.
Description
Keywords
Detecção de congestionamento de tráfego Dados de Veículos Flutuantes (FCD) Análise de trajetórias GPS Correspondência de mapas (Valhalla) Engenharia de características Índice de Redução de Velocidade (SRI) XGBoost (Extreme Gradient Boosting) Sistemas de Transporte Inteligente (ITS) Processamento de dados escalável (Dask Parquet) SUMO (Simulação de Mobilidade Urbana) Aprendizado de máquina interpretável Monitoramento de mobilidade urbana Validação de perigos Traffic congestion detection Floating Car Data (FCD) GPS trajectory analysis Map-matching (Valhalla) Feature engineering Speed Reduction Index (SRI) XGBoost (Extreme Gradient Boosting) Intelligent Transportation Systems (ITS) Scalable data processing (Dask Parquet) SUMO (Simulation of Urban Mobility) Interpretable machine learning Urban mobility monitoring Hazard validation