Grid-Based Arrival Time Prediction for Public Transit
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current public transportation systems face inaccuracies in predicting the arrival time of vehicles due to unforeseen circumstances like traffic delays, leading to unreliable schedules and long waiting times for commuters.
Innovation Solution
A grid-based framework that models location data to predict the arrival time of moving objects by determining cell transition times and averaging these across multiple vehicles, providing real-time and accurate estimates through GPS and GLONASS data collection and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional scheduling systems are used, then system simplicity is maintained, but prediction accuracy deteriorates due to inability to account for unforeseen circumstances
Solution Approach 1:
The system segments the service area into a grid of cells and divides the vehicle route into multiple segments corresponding to these cells. This segmentation allows the system to track vehicle position at multiple discrete points along the route, improving prediction accuracy by enabling more granular monitoring without requiring complete system redesign.
Solution Approach 2:
The patent introduces a prediction server as an intermediary component that receives location data from vehicles and schedule data from transportation operators, then processes this information to generate arrival time predictions. This intermediary layer handles the complex data processing and algorithm execution, isolating the complexity from both the vehicle tracking system and the user interface.
2Loss of information
If real-time location tracking is implemented, then information availability is improved, but data processing complexity increases
Solution Approach 1:
The service area is divided into a grid of cells, and the vehicle route is segmented into multiple segments. This segmentation allows the system to process location data at discrete intervals (cell boundaries) rather than continuously, reducing the overall data processing complexity while maintaining real-time tracking capabilities.
Solution Approach 2:
The system focuses processing efforts on critical events - specifically when vehicles enter or exit grid cells - rather than processing every continuous location update. This partial action approach processes only the essential data points needed for prediction while ignoring redundant intermediate positions.
3Reliability
If scheduled timetables are published, then commuter planning capability is improved, but reliability deteriorates when vehicles deviate from schedule
Solution Approach 1:
The system performs preliminary calculations of average travel times between consecutive grid cells based on historical data before vehicles actually travel these routes. These pre-computed averages are stored and readily available when vehicles are tracked, enabling immediate prediction updates without real-time computation delays.
Solution Approach 2:
The system continuously monitors actual vehicle positions and compares them against predicted positions based on the schedule and historical averages. This feedback mechanism allows the system to detect deviations from the schedule in real-time and update predictions accordingly, maintaining reliability even when vehicles deviate from their planned routes.
Data Source
AI summary
Described herein is a framework for predicting arrival time of moving objects based on grid modeling of location data. In accordance with one aspect of the framework, location data of a moving object is received and positions of the moving object may be identified based on a grid model by mapping the location data to the grid model. The grid model may be segmented into cells where a cell transition time of the moving object from a current cell to another cell may be determined. The arrival time of the moving object may then be predicted based on the cell transition time of the moving object. The arrival time of the moving object may further be presented to one or more user devices.