Traffic Prediction Using Link Interaction Deviations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic prediction methods are inadequate for providing accurate, real-time predictions in large networks, especially when high-quality weather and event data are unavailable, and they fail to account for link interactions and spatial correlations, leading to reduced accuracy and computational inefficiency.
Innovation Solution
A method that calculates traffic predictions using deviations from historical data, incorporating time-dependent traffic state data and correlation techniques across links, with multiple prediction schemes for short and medium-term forecasts, and adapts to recent traffic information for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional traffic prediction methods use average travel time values for route guidance, then the method is simple to implement, but the prediction accuracy deteriorates considerably during peak congestion periods when travel times vary widely
Solution Approach 1:
The patent applies dynamics by transitioning from static average travel time values to dynamic, time-dependent travel time predictions. The system uses historical traffic data and temporal patterns to generate predictions that adapt to varying congestion conditions throughout the day, allowing route guidance to reflect current and future traffic states rather than relying on fixed averages that become inaccurate during peak periods.
2Measurement precision
If detailed real-time traffic data collection methods are implemented to improve prediction accuracy, then prediction quality improves, but the system complexity and computational requirements increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-processing and storing historical traffic data in structured formats that capture temporal patterns and deviations. This preprocessing work is done in advance, allowing the prediction system to efficiently query and analyze pre-organized data rather than processing raw data in real-time, thus improving prediction accuracy without proportionally increasing real-time computational complexity.
Solution Approach 2:
The patent uses copying by creating simplified representations of traffic patterns through historical data analysis. Instead of collecting and processing all possible detailed real-time data, the system creates predictive models that copy essential traffic behavior patterns from historical records, enabling accurate predictions with reduced data processing requirements.
3Measurement precision
If comprehensive link interaction models are incorporated into traffic prediction to improve accuracy, then prediction quality improves, but the computational time increases making real-time prediction infeasible
Solution Approach 1:
The patent applies segmentation by breaking down the traffic network into individual links and analyzing their interactions in a structured, modular manner. Rather than computing all possible link interactions simultaneously across the entire network, the system segments the problem into manageable components that can be processed efficiently, allowing comprehensive interaction modeling without prohibitive computational costs.
Data Source
AI summary
A method and structure for predicting traffic on a network, includes a receiver which receives data related to traffic on at least a portion of a network. A calculator calculates a traffic prediction for at least a part of the network, the traffic prediction being calculated by using a deviation from a historical traffic on the network.