Abstract:Traffic flow is chaotic due to nonstationary realistic factors, and revealing the internal nonlinear dynamics of chaotic data and making high-accuracy predictions is the key to traffic control and inducement. Given that high-quality phase space reconstruction is the foundation of predictive modeling. Firstly, an improved C-C method based on the fused norm search domain is proposed to address the issue that the C-C method in the phase space reconstruction algorithm does not meet the Euclidean metric accuracy and reduces the reconstruction quality when the infinite norm metric is used. Secondly, to address the problem of insufficient learning ability of traditional convolutional combinatorial modeling for complex phase space laws of chaotic traffic flow, the high-dimensional phase space features are extracted using the layer-by-layer pretraining mechanism of convolutional deep belief networks (CDBNs), and the temporal features are extracted by combining with long short-term memory (LSTM). Finally, an improved probabilistic dynamic reproduction-based genetic algorithm (PDRGA) is proposed to address the problem of the hybrid model falling into a local optimum when learning the phase space law. Experiments are conducted in three aspects: phase space reconstruction quality analysis, comparison of optimization algorithm convergence, and prediction model performance comparison. The experimentation with two data sets demonstrates that the improved C-C method combines the advantages of the high accuracy metric of the L2 norm with the low operational complexity of the infinite norm, achieving a balance between reconstruction quality and algorithm efficiency. The proposed PDRGA optimization algorithm is a lightweight improvement of the traditional genetic algorithm (GA) and solves the problem that the model tends to fall into a local optimum by optimizing the initial weights of CDBN. Meanwhile, the five error evaluation indexes of the proposed PDRGA-CDBN-LSTM hybrid model are lower than those of the baseline model, providing a new modeling idea for chaotic traffic flow prediction.

Chaotic Features for Traffic Video Classification

Chaos Real-Time Recognition of Traffic Flow by Using SVM Rough Neural Network

&Lt;title>automatic Traffic Real-Time Analysis System Based on Video</title>

Exploiting Chaotic Feature Vector for Dynamic Textures Recognition.

Real-Time Intelligent Recognition of Chaos in Traffic Flow Using Reduced Support Vector Machine

A FuzzyHiddenMarkovModelforTraffic Status Classification Basedon VideoFeatures

A Fuzzy Hidden Markov Model for Traffic Status Classification Based on Video Features

Chaotic property analysis and prediction model study for traffic flow time series

Research on Recognition and Classification of Moving Objects in Mixed Traffic Based on Video Detection

Research on Rapid Recognition of Chaos in Traffic Flow Based on Support Vector Machine

Spatiotemporal Analysis of Static and Dynamic Traffic Elements from Road Scenes.

A Combined Algorithm for Real-time On-line Rapid Identification of the Chaos in Traffic Flow

A method of on-line fast identification for the chaos in traffic flow

Chaotic Features for Dynamic Textures Recognition with Group Sparsity Representation

Video scene understanding in complicated conditions

A Method of Fast Detecting the Chaotic Characteristics of Video Sequence

Traffic Flow Predicting Of Chaos Time Series Using Support Vector Learning Mechanism For Fuzzy Rule-Based Modeling

Chaotic analysis of traffic time series

An Improved Phase Space Reconstruction Method-Based Hybrid Model for Chaotic Traffic Flow Prediction

On the Chaotic Dynamics Analysis of Internet Traffic

Vehicle Feature Point Trajectory Clustering and Vehicle Behavior Analysis in Complex Traffic Scenes.