Traffic Prediction Using Directed Graph Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current traffic condition prediction systems are unreliable due to inconsistent traffic signal timings and manual control, leading to inaccurate travel planning and increased traffic delays.
Innovation Solution
A method and system that generate a directed graph of a road network based on static data, using position information from vehicles to determine filtered time differences and predict traffic conditions, incorporating current, historical, and cluster-level traffic data to optimize routing and travel times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If current traffic conditions are predicted based on consistent traffic signal time, then travel planning is simplified, but prediction accuracy deteriorates due to inconsistent actual traffic signal timings
Solution Approach 1:
The system continuously collects actual vehicle position data and traffic condition information from multiple vehicles traveling through the road network. This feedback mechanism allows the system to update traffic signal timing predictions based on real observed behavior, gradually converging toward accurate predictions despite initial inconsistencies in traffic signal timing.
Solution Approach 2:
The system performs preliminary data collection and analysis of traffic patterns, vehicle positions, and signal timings before generating final predictions. By pre-processing historical data and establishing baseline models, the system prepares the foundation for accurate real-time predictions without requiring perfect initial signal timing information.
2Adaptability or versatility
If traffic signal timings are manually controlled or vary by hour, then local needs are met, but prediction reliability deteriorates
Solution Approach 1:
The system dynamically adapts to varying traffic signal timings by continuously monitoring actual signal behavior and adjusting predictions accordingly. Rather than assuming fixed timing patterns, the system learns the actual dynamic behavior of signals including manual controls and hourly variations, maintaining prediction reliability despite these adaptations.
Solution Approach 2:
The system changes its prediction parameters based on observed traffic conditions, time of day, and signal behavior patterns. By adjusting model parameters to match actual traffic patterns rather than relying on fixed assumptions, the system maintains accuracy despite manual controls or hourly variations in signal timing.
3Productivity
If traffic conditions are predicted without filtering vehicle-related service time, then calculation is simplified, but prediction accuracy deteriorates
Solution Approach 1:
The system extracts and removes vehicle-related service time periods (such as stops for loading/unloading passengers or vehicle maintenance) from the raw traffic condition data. By separating these non-traffic-related time periods, the system calculates more accurate traffic flow predictions while maintaining computational efficiency through targeted filtering rather than complete data exclusion.
Data Source
AI summary
A method and a system for predicting traffic conditions of a geographical area are provided. Position information, received from devices of corresponding vehicles that are traversing between road segments of a road network including at least first and second road segments, is located on a directed graph of the road network. Further, first and second times are determined based on the located position information on the directed graph. A traffic time of the first road segment is determined based on an average of time differences of the first and second times of each of the one or more first vehicles after filtering out a time period between the first and second times. The traffic conditions of the road network are predicted based on the determined traffic time of each of the road segments.