Hybrid ConvLSTM-BiLSTM Model for Urban Traffic Flow Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic flow prediction methods based on vehicle-passage flow data are limited by insufficient data feature analysis and are only applicable to single road condition scenarios, failing to accurately predict traffic congestion in urban regions.
Innovation Solution
An urban-region road network vehicle-passage flow prediction method and system utilizing a ConvLSTM and BILSTM hybrid deep learning model, which performs spatial and temporal distribution feature analysis, constructs a predictive model, and identifies traffic states through real-time data processing and Prophet linear time series prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing traffic flow prediction methods based on vehicle-passage flow data are used, then the method is simple to implement, but the data feature analysis is insufficient and only applicable to single road condition scenario
Solution Approach 1:
The patent combines ConvLSTM and BILSTM into a hybrid deep learning model to integrate spatial-temporal and bidirectional temporal features. This merging of multiple neural network components enables the model to handle diverse road condition scenarios (congestion, smooth flow, peak hours) while maintaining a unified framework, thereby improving adaptability without proportionally increasing complexity
Solution Approach 2:
The patent segments the traffic flow prediction task into multiple feature analysis components: spatial distribution features, temporal distribution features, and their interactions. By dividing the complex prediction problem into manageable feature extraction and analysis modules, the system achieves comprehensive feature analysis across different road conditions while keeping each module relatively simple
2Measurement precision
If comprehensive spatial and temporal feature analysis is performed, then the prediction accuracy is improved, but the calculation time and processing complexity increase
Solution Approach 1:
The patent performs preliminary feature extraction and preprocessing on spatial and temporal distributions before feeding data into the hybrid deep learning model. By pre-computing statistical features, correlation matrices, and normalized data, the system reduces the computational burden during the actual prediction phase, thereby maintaining high prediction accuracy while minimizing real-time processing time
Solution Approach 2:
The patent implements continuous feature extraction and real-time model inference that operates seamlessly during normal traffic monitoring. The system maintains continuous prediction capability without interrupting traffic flow monitoring operations, ensuring that the useful action of prediction occurs continuously with minimal time loss
Data Source
AI summary
Disclosed are urban-region road network vehicle-passage flow prediction method and system based on a hybrid deep learning model. The method comprises: compiling statistics on traffic flow on the basis of vehicle-passage data of a checkpoint; performing spatial and temporal distribution feature analysis on vehicle-passage flow data of the checkpoint, and performing feature extraction according to an analysis result, so as to acquire a spatial and temporal influence factor; constructing and training a ConvLSTM and BILSTM hybrid deep learning model according to the spatial and temporal influence factor; performing synchronous prediction on traffic flow of an urban-region road network, selecting prediction loss functions and evaluation indicators, and performing visual representation on a result; and calculating a traffic flow variation degree by means of a linear time series prediction model Prophet, and performing traffic state identification, so as to realize traffic state pre-determination.