Road traffic modelling and management : using statistical monitoring and deep learning /

Detalles Bibliográficos
Clasificación:Libro Electrónico
Otros Autores: Harrou, Fouzi (Editor )
Formato: eBook
Idioma:Inglés
Publicado: Amsterdam : Elsevier, 2022.
Temas:
Traffic engineering > Mathematical models.
Traffic engineering.
Technique de la circulation > Mod�eles math�ematiques.
Technique de la circulation.
traffic engineering.
Traffic engineering
Traffic engineering > Mathematical models
Acceso en línea:Texto completo
Tabla de Contenidos:
  • Front Cover
  • Road Traffic Modeling and Management
  • Copyright
  • Contents
  • 1 Introduction
  • 1.1 Introduction
  • 1.1.1 Types road traffic sensors
  • 1.1.2 Key traffic features
  • 1.1.3 Traffic flow modeling
  • 1.2 Objectives and structure of the book
  • References
  • 2 Road traffic modeling
  • 2.1 Introduction
  • 2.2 Traffic detector types
  • 2.2.1 Embedded traffic detectors
  • 2.2.2 Intrusive detectors
  • 2.2.3 Non-intrusive detectors
  • 2.3 Traffic modeling model-based approaches
  • 2.3.1 Microscopic approach
  • 2.3.1.1 The microscopic variables
  • 2.3.1.2 Microscopic models
  • 2.3.2 Mesoscopic approach
  • 2.3.2.1 Mesoscopic models
  • 2.3.3 Macroscopic approach
  • 2.3.3.1 The macroscopic variables
  • 2.3.4 The macroscopic models
  • 2.3.4.1 First-order macroscopic traffic models
  • 2.3.4.2 The cell transmission model
  • 2.3.5 PWSL traffic flow model
  • 2.3.5.1 PWSL discrete dynamics
  • 2.3.5.2 PWSL continuous dynamics
  • 2.3.5.3 Boundary conditions
  • 2.3.5.4 Insider road section traffic density dynamics
  • 2.3.5.4.1 Simple link density dynamics
  • 2.3.5.4.2 Merge and diverge links dynamics
  • Merge links dynamics (the ON ramp configuration)
  • Diverge links dynamics (the OFF ramp configuration)
  • 2.3.5.5 PWSL parameter calibration
  • Geometric and data traffic description
  • Calibration methodology
  • 2.3.6 Model validation and simulation results
  • 2.3.6.1 Measurements of effectiveness
  • 2.4 Conclusion
  • References
  • 3 Road traffic density estimation
  • 3.1 Introduction
  • 3.2 Observability of PWSL model
  • 3.2.1 Generalities of observability study
  • 3.2.2 PWSL observability study
  • 3.3 PWSL-based observer
  • 3.3.1 Principle and description of hybrid observer
  • 3.3.2 Synthesis of PWSL-based hybrid observer
  • 3.3.2.1 Discrete mode identification
  • 3.3.2.2 Continuous state estimation.
  • 3.3.3 The PWSL-Kalman filter design for traffic state estimation
  • 3.4 Application
  • 3.5 Discussion
  • References
  • 4 Model-based techniques for traffic congestion detection
  • 4.1 Introduction
  • 4.2 The PWSL-KF design for traffic state estimation
  • 4.2.1 PWSL approach
  • 4.2.2 Kalman filter for traffic density estimation
  • 4.3 kNN-based monitoring charts
  • 4.3.1 Mixed kNN-Shewhart approach
  • 4.3.2 Coupled kNN-EWMA approach
  • 4.3.3 kNN-based schemes with nonparametric thresholds
  • 4.3.3.1 KDE
  • Kernel function selection
  • Bandwidth selection
  • 4.3.3.2 Nonparametric kNN-based detectors
  • 4.4 Combining PWSL model with monitoring schemes
  • 4.5 Results and case study
  • 4.5.1 Model performance evaluation
  • 4.5.2 Detection results
  • 4.5.2.1 Performance evaluation
  • 4.5.2.2 Detection of abrupt congestions
  • 4.5.2.3 Scenarios with intermittent congestion
  • 4.5.2.4 Detection of gradual congestion
  • 4.6 Discussion
  • References
  • 5 Traffic congestion detection: data-based techniques
  • 5.1 Introduction
  • 5.2 Data preprocessing and denoising
  • 5.2.1 Exponentially weighted moving-average (EWMA) filter
  • 5.2.2 Wavelet-based multiscale filtering
  • 5.3 Univariate statistical monitoring schemes
  • 5.3.1 EWMA (exponentially weighted moving average)
  • 5.3.2 Generalized likelihood ratio charts
  • 5.4 Features extraction with principle component analysis (PCA)
  • 5.4.1 Features extraction with PCA
  • 5.4.2 Determination of model order
  • optimal dimension of PC
  • 5.4.2.1 Cumulative percent variance
  • 5.4.2.2 Cross-validation
  • 5.4.2.3 Scree test
  • 5.4.2.4 Parallel analysis
  • 5.4.2.5 Eigenvalue-one rule
  • 5.5 PCA-based anomaly detection charts
  • 5.5.1 SPE chart
  • 5.5.2 Hotelling's T2
  • 5.5.3 The basic procedure for the PCA with Matlab�
  • 5.5.4 A combined T2 and Q index.
  • 5.5.5 Anomaly detection using coupled EWMA-based SPE and T2 methods
  • 5.5.5.1 The amalgamated Q-EWMA monitoring scheme
  • 5.5.5.2 The amalgamated T2-EWMA monitoring scheme
  • 5.5.6 Parametric and nonparametric PCA-based monitoring charts
  • 5.6 Combining PCA with monitoring schemes for traffic congestion detection
  • 5.7 Results and discussion
  • 5.7.1 Data description
  • 5.7.2 Data analysis
  • 5.7.3 PCA modeling
  • 5.7.4 Traffic anomalies detection
  • 5.7.4.1 Detection of abrupt congestions
  • 5.7.4.2 Scenarios with intermittent congestion
  • 5.7.4.3 Detection of gradual congestion
  • 5.8 Discussion
  • References
  • 6 Recurrent and convolutional neural networks for traffic management
  • 6.1 Introduction
  • 6.2 Recurrent neural network for traffic time series data
  • 6.2.1 Data collection
  • 6.2.2 Preprocessing the data before building a model
  • 6.2.3 Data smoothing
  • 6.2.3.1 Exponentially weighted moving average (EWMA) filter
  • 6.2.3.2 Wavelet-based multiscale filter
  • 6.2.4 Recurrent neural network (RNN)-based models
  • 6.2.4.1 RNN model
  • 6.2.4.2 LSTM model
  • 6.2.4.3 LSTM implementation steps
  • 6.2.4.4 Enhanced performance of LSTM models
  • 6.2.5 Evaluating the model
  • 6.2.6 GRU models
  • 6.3 Road traffic forecasting using LSTM- and RNN-driven models
  • 6.3.1 Proposed traffic forecasting methods
  • 6.3.1.1 Forecasting results and discussion
  • 6.3.1.2 Data description
  • 6.3.2 Forecasting results
  • 6.4 Detecting traffic congestion using deep learning-based monitoring charts
  • 6.4.1 EWMA and DEWMA monitoring schemes
  • 6.4.1.1 EWMA monitoring scheme
  • 6.4.1.2 Double EWMA scheme
  • 6.4.1.3 KDE-based DEWMA monitoring scheme
  • 6.4.2 Traffic congestion detection using deep learning-based DEWMA schemes
  • 6.4.3 Traffic congestion detection: case study
  • 6.4.3.1 LSTM- and GRU-based traffic flow modeling
  • 6.4.3.2 Detection results.
  • 6.4.3.2.1 Abrupt traffic congestion
  • 6.4.3.2.2 Detection of gradual congestion
  • 6.5 Vision-based traffic monitoring using CNN
  • 6.5.1 Convolutional neural networks
  • 6.5.1.1 Converting network traffic to images
  • 6.5.1.2 CNN for network traffic prediction
  • 6.5.2 TensorFlow integration with Python
  • 6.5.2.1 Quickstart of TensorFlow Lite with Python
  • 6.5.2.2 Installation of the TensorFlow Lite for Python
  • 6.6 Discussion
  • References
  • 7 Conclusion and further research directions
  • 7.1 Introduction
  • References
  • Index
  • Back Cover.

Ejemplares similares