Lane Closure Detection Using Probe Data Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current GPS systems have insufficient accuracy to determine the specific lane of a road segment, making it difficult to detect lane closures effectively, as the 95% confidence interval of GPS location is twice the width of a typical lane, hindering the determination of lane-level traffic information.
Innovation Solution
A method that analyzes probe data using a distance parameter to identify spatial clusters along a road segment, comparing current cluster numbers to historical data to detect lane closures, and provides notifications for route calculations or alerts when a reduction in clusters indicates a lane closure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS location data is used to determine lane-level traffic information, then real-time location tracking is provided, but the accuracy is insufficient to determine the specific lane due to GPS confidence interval being twice the lane width
Solution Approach 1:
The road segment is divided into multiple lanes based on historical probe data clustering. By segmenting the spatial data into distinct lane clusters from historical information, the system can identify which specific lanes are closed even though individual GPS points are imprecise. The segmentation of spatial data into discrete lane entities enables lane-level detection without requiring high GPS precision.
Solution Approach 2:
Historical probe data is pre-analyzed to establish the expected number of lanes and their spatial distributions before detecting current lane closures. This preliminary clustering of historical data creates a baseline model of normal traffic patterns, which is then compared against current probe data to identify anomalies such as lane closures, enabling detection without relying on high GPS accuracy.
2Loss of information
If probe data analysis is performed to detect lane closures, then lane-level traffic information can be determined, but the system complexity increases due to clustering algorithms and historical data comparison
Solution Approach 1:
The system uses probe data from vehicles themselves to detect lane closures without requiring external infrastructure such as cameras or radar. The vehicles' own GPS and communication systems are leveraged to provide the detection function, reducing the need for additional complex detection hardware. The clustering algorithm processes data autonomously using only the probe information already being collected.
3Loss of time
If real-time lane closure detection is implemented using probe data, then timely alerts and route adjustments can be provided, but the processing time and computational resources are required for cluster analysis and historical comparison
Solution Approach 1:
The system performs clustering analysis only on the distance parameters extracted from probe data rather than processing all raw GPS coordinates and vehicle information. By extracting and analyzing only the relevant distance parameter (lateral distance from road centerline), the computational burden is reduced while still enabling effective lane closure detection. This selective processing maintains real-time capability with lower computational resource requirements.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A plurality of instances of probe data are received. Each instance is received from a probe apparatus of a plurality of probe apparatuses each comprising a plurality of sensors and being onboard a vehicle. An instance comprises location information indicating a location of the corresponding probe apparatus. For each of one or more instances, a distance parameter is determined based on the location information and a road segment corresponding to the location. A set of distance parameters is defined based on the distance parameter determined for each of the one or more instances. The set of distance parameters is analyzed to identify clusters of probe data. The number of clusters identified is determined and compared to a historical number of clusters. If the number of clusters identified is less than the historical number of clusters, it is determined that there is a lane closure corresponding to the road segment.