Neural Network Graph Completion for Traffic Flow Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current traffic systems face challenges in providing accurate and reliable real-time traffic flow monitoring due to erroneous data from vehicle sensors and fixed roadside units, limiting their ability to scale effectively in large-scale road networks.
Innovation Solution
A traffic system that uses a trained neural network model to aggregate and complete a graph structure of traffic flows from sensor-rich vehicle data, incorporating vehicle-to-everything communication and a hierarchical vehicular platform, to generate a reliable and accurate graph model of traffic flows, thereby overcoming data gaps and noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If fixed roadside units and vehicle counters are used for traffic flow monitoring, then real-time traffic data can be collected, but the data quality deteriorates due to high noise, missing data, and high-error rates
Solution Approach 1:
The patent introduces a neural network model as an intermediary between the raw sensor data from fixed RSUs and vehicle counters and the final traffic flow information. This neural network processor acts as a mediator that filters, completes, and validates the erroneous data, transforming unreliable inputs into reliable traffic flow outputs through learned patterns and relationships.
Solution Approach 2:
The patent replaces the mechanical/data-processing approach of directly using raw sensor counts with a neural network-based processing system. Instead of relying on the physical limitations and error-prone nature of fixed RSUs and vehicle counters, the system substitutes a learned computational model that can infer accurate traffic flows even from incomplete or noisy inputs.
2Ease of manufacture
If fixed roadside units are deployed in limited geographic areas, then traffic monitoring can be implemented, but the system cannot scale to large-scale road networks
Solution Approach 1:
The neural network model serves as a universal processor that can handle traffic data from any geographic location uniformly. Once trained, the same model architecture and processing logic can be applied across diverse road networks of varying sizes and characteristics, enabling the system to scale from local to regional implementations without requiring location-specific customization of the core processing mechanism.
3Productivity
If complete and accurate traffic flow information is provided to vehicles, then navigation efficiency and congestion management improve, but the system complexity increases due to the need for graph completion and neural network processing
Solution Approach 1:
The neural network model is trained in advance on historical and simulated traffic data to learn the complex relationships between sensor inputs and actual traffic flows. This preliminary training phase allows the model to be deployed as a pre-configured processing system that automatically performs graph completion and data validation without requiring complex real-time computations during operational use, thereby reducing online system complexity.
Data Source
AI summary
System, methods, and other embodiments described herein relate to improving monitoring of traffic flows. In one embodiment, a method includes aggregating perception data associated with a road network from information sources to a server over a network. The method also includes generating a graph structure from the perception data in association with a neural network model. The graph structure is an incomplete representation of the road network in view of missing data. The method also includes completing the graph structure using the neural network model that forms a graph model of the traffic flows to de-noise the graph structure according to road constraints between two points in the road network. The method also includes communicating the graph model of the traffic flows to a vehicle to navigate traffic in the road network.