Road Closure Prediction Using Probe Data and Slowdown Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems fail to provide timely and accurate predictions of road closures caused by vehicle incidents, leading to potential safety risks and economic losses due to inadequate warning signals for upstream vehicles.
Innovation Solution
A system and method for predicting road closures using real-time and historical data from multiple sources, including probe data from vehicles, to identify dangerous slowdown events and predict road closures with confidence adjustments, providing real-time updates and alternative routes for autonomous and semi-autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time probe data from multiple vehicles is collected and analyzed to predict road closures, then prediction accuracy and timeliness improve, but system complexity and data processing requirements increase
Solution Approach 1:
The system segments the road network into multiple road segments and processes probe data for each segment independently. This allows parallel processing of data from multiple vehicles while maintaining manageable complexity for each individual segment analysis.
Solution Approach 2:
The patent introduces an intermediary processing layer that aggregates probe data from multiple vehicles, applies machine learning models, and generates predictions. This intermediary layer simplifies the overall system architecture by centralizing the complex data processing logic.
2Reliability
If machine learning models are trained on historical probe data to identify dangerous slowdown events, then prediction reliability improves, but data processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-training machine learning models on historical probe data offline. This allows the models to be ready for real-time prediction without requiring extensive processing time during actual road closure events.
Solution Approach 2:
The patent applies partial action by using simplified features and selective data sampling for real-time predictions, rather than processing complete historical datasets. This reduces computational time while maintaining acceptable prediction reliability.
3Loss of time
If the system monitors multiple road segments and provides real-time alerts for upstream vehicles, then safety warning timeliness improves, but information processing load and communication requirements increase
Solution Approach 1:
The system applies local quality by providing targeted warnings only to vehicles that are actually at risk of encountering road closures. Instead of broadcasting to all vehicles, alerts are sent selectively based on each vehicle's location, route, and proximity to predicted closure zones.
Solution Approach 2:
The patent segments the monitoring task by dividing the road network into multiple segments and assigning different processing priorities. This allows the system to focus computational resources on critical segments while reducing overall information processing load.
Data Source
Figure 1~2
Figure 3A
Figure 3B
AI summary
A method, a system, and a computer program product are provided for predicting road closure in a region. The method comprises obtaining, probe data, such as sensor data, and map data, for the region. The method may include detecting a change in speed of a one or more vehicle on a road, based on the obtained probe data and the obtained map data for the region, wherein the change in speed is associated with a slowdown event associated with the one or more vehicles. The method may include identifying a vehicle event on the road based on the detected change in speed of the one or more vehicles, wherein the vehicle event is associated with a location corresponding to a matching trajectory of the one or more vehicles and predicting the road closure based on the identified vehicle event.