Traffic Signal Prediction Correction Using Real-Time Probe Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting traffic signal state changes are inaccurate due to unpredictable anomalies such as clock drifts and special events, which can cause deviations from scheduled timing plans, affecting the reliability of predictions for drivers and autonomous vehicles.
Innovation Solution
The system generates preliminary predictions based on scheduled timing plans and adjusts them using real-time data from traffic signal controllers to correct for variations, and further validates these predictions using real-time probe data from GPS sources to ensure accuracy before dissemination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traffic signal state changes are predicted based on scheduled timing plans, then predictions can be generated in advance, but accuracy deteriorates due to clock drifts and special control events causing deviations from planned schedules
Solution Approach 1:
The system continuously monitors actual traffic signal state changes and compares them with predicted changes from timing plans. When deviations are detected (such as from clock drift or preemptions), the system uses this feedback to identify and correct inaccurate predictions, ensuring ongoing accuracy despite schedule variations
Solution Approach 2:
The system generates preliminary predictions based on scheduled timing plans in advance, then prepares correction mechanisms by establishing baseline expected states. This allows the system to have predictions ready while maintaining the capability to adjust them when anomalies occur
2Measurement precision
If real-time probe data collection and validation is implemented, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The system uses probe data from GPS sources as an intermediary validation mechanism. Rather than directly monitoring all traffic signal controllers, it leverages independent third-party data to verify predictions, simplifying the validation architecture while maintaining accuracy
Solution Approach 2:
The system validates its own predictions using externally sourced probe data, performing self-checks on prediction accuracy. This autonomous validation reduces the need for complex external verification systems while maintaining high accuracy standards
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems to improve the accuracy of traffic signal state change predictions (142) are disclosed. Predictions for fixed-time signals are generated based on their scheduled timing plan and the current clock/time (142), but these predictions are subject variations, for example, due to traffic signal controller clock drift (Fig. 2). Real-time actual, not predicted, data is collected and utilized to correct for these variations (Fig. 3). Further, real-time probe data is collected (122) and used to validate correctness (146) of the corrected predictions in real time. In one embodiment, GPS data from travelers' devices (100) is utilized to assess validity of the generated predictions, looking particularly at signal stop line crossings relative to predicted green time window (128), (Fig. 4).