Travel Time Database Model Construction Using Vehicle Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Providers of traffic information often lack a statistically adequate database of current data for generating real-time traffic information, relying on historical time variation curves that are influenced by traffic characteristics such as time of day, weather, and events, which limits the accuracy of travel time predictions.
Innovation Solution
A method is developed to create a travel time database model using vehicle-collected data, where relative travel time losses on leg sections are calculated and weighted based on traffic characteristics, enabling the estimation of probability distributions for travel times, and updating the knowledge base using Bayesian approaches and learning algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If historical time variation curves are used for generating traffic information, then the system can operate without a large database, but the accuracy of travel time predictions deteriorates
Solution Approach 1:
The system performs preliminary data collection and model training by storing travel time observations from multiple vehicles in a database, preparing probability distribution models in advance for different road sections and traffic characteristics. This preliminary action enables accurate real-time predictions without requiring continuous large-scale data collection during operation.
Solution Approach 2:
The system creates simplified probability distribution models that copy the essential patterns from historical data, representing complex traffic behavior through statistical distributions (e.g., gamma distributions) for each road section. These copied models enable accurate predictions while requiring minimal real-time data.
2Measurement precision
If a statistically adequate database is created by recording data from multiple vehicles, then the accuracy of travel time predictions improves, but the complexity of data collection and processing increases
Solution Approach 1:
Vehicles automatically collect and transmit their own travel time data without requiring external intervention. The system uses the vehicles' existing sensors and communication capabilities to self-generate the database, eliminating the need for complex external data collection infrastructure.
Solution Approach 2:
The system uses a universal data collection approach where any vehicle can contribute data for any road section it traverses. The same data collection mechanism serves multiple purposes: building the database, validating models, and providing real-time predictions across the entire road network.
3Measurement precision
If multiple possible routes are analyzed with stochastic weighting, then the accuracy of travel time estimation improves, but the computational complexity increases
Solution Approach 1:
The system segments the road network into discrete road sections with pre-defined probability distribution models. Each section is analyzed independently, and results are combined along routes. This segmentation reduces computational complexity by avoiding full-route stochastic simulations while maintaining accuracy through localized model precision.
Solution Approach 2:
The system changes the computational approach from direct stochastic simulation of multiple routes to using pre-computed probability distribution parameters (mean, variance) for each road section. This parameter-based approach dramatically reduces computational power requirements while preserving the benefits of analyzing multiple routes through the statistical properties embedded in the parameters.
Data Source
AI summary
According to a method for creating a model for a travel time database, a leg network comprising leg sections between a starting point and a destination point are analyzed. Multiple routes are ascertained between the starting and destination points. Each leg section of a route is associated with a relative travel time loss and is weighted. The ascertained travel time losses and the associated weightings are used as input data for a learning method by way of which an existing knowledge base is iteratively expanded.