Traffic Prediction Using Historical Deviation Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional traffic prediction methods are ineffective in providing accurate and reliable predictions, especially in the absence of real-time data, and struggle to handle missing data, which leads to invalidation or faulty predictions, and are computationally burdensome, limiting their ability to provide state-dependent internet mapping and route guidance for large areas.
Innovation Solution
A method that uses historical traffic patterns to estimate missing real-time data and calculate traffic predictions by employing a deviation-based approach, incorporating a calibrated model that accounts for correlations across links and updates data periodically to provide accurate and fast traffic predictions, even in the presence of missing data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional traffic prediction methods are used, then predictions can be made, but they become invalid or faulty when real-time data is missing
Solution Approach 1:
The system performs preliminary actions by pre-calculating and storing historical traffic patterns, average travel times, and deviation values in databases before actual prediction is needed. When real-time data is missing, these pre-computed historical values are retrieved and used to fill gaps, enabling reliable predictions without requiring complete real-time data availability.
Solution Approach 2:
The patent introduces intermediary components including a deviation database that stores historical deviations from average travel times, and a processing system that acts as a mediator between available real-time data and missing data requirements. This intermediary structure allows the system to infer missing real-time values using historical deviation patterns, maintaining prediction reliability despite data gaps.
2Reliability
If conventional traffic prediction methods are used, then predictions can be made, but they are computationally intensive and cannot handle large areas
Solution Approach 1:
The patent segments the traffic network into discrete links and nodes, and divides the prediction process into modular components: retrieving average travel times for individual links, calculating deviations from historical patterns, and aggregating results. This segmentation allows the system to handle large areas by processing predictions link-by-link using simplified formulas rather than complex global models.
Solution Approach 2:
The system changes parameters by transforming complex traffic flow simulations into simpler statistical calculations based on historical deviations. Instead of running computationally intensive traffic assignment models, the system uses pre-computed average travel times and stores deviation values that can be quickly retrieved and applied, reducing computational complexity while maintaining accuracy for large geographic areas.
3Productivity
If average travel times are used for route guidance, then simple computations can be made, but they fail to account for time-dependent congestion variations
Solution Approach 1:
The system performs preliminary computation of historical traffic patterns and stores deviation values from average travel times in a database. During actual route guidance, these pre-computed deviations are quickly retrieved and applied to current conditions, allowing the system to maintain high computation speed while accurately reflecting time-dependent congestion variations without requiring complex real-time simulations.
Solution Approach 2:
The patent creates a simplified copy of historical traffic behavior in the form of stored deviation values that represent typical variations from average travel times. Instead of re-simulating complex traffic flow patterns in real-time, the system uses these pre-computed historical copies adjusted for current conditions, maintaining both computational efficiency and accuracy in capturing congestion variations.
Data Source
AI summary
A method and apparatus for predicting traffic on a transportation network where real time data points are missing. In one embodiment, the missing data is estimated using a calibration model comprised of historical data that can be periodically updated, from select links constituting a relationship vector. The missing data can be estimated off-line whereafter it can be used to predict traffic for at least a part of the network, the traffic prediction being calculated by using a deviation from a historical traffic on the network. The invention further discloses a method for in-vehicle navigation; and a method for traffic prediction for a single lane.