Passenger Flow Prediction Using Historical Travel Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional passenger flow prediction systems in transportation rely solely on current sensor data, leading to inaccurate assessments of transport capacity and potential overcrowding or underutilization, which can result in suboptimal service quality and increased costs.
Innovation Solution
A method and system that utilize historical travel data to estimate the number of waiting passengers, delay times, and transport congestion, while also determining a switching probability for passengers to alternate transport options, thereby predicting passenger flow and dropout risks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional sensor-based techniques are used to compute passenger flow and transport delays, then current state monitoring is achieved, but future state prediction accuracy deteriorates leading to misleading capacity assessments
Solution Approach 1:
The system performs preliminary actions by collecting and analyzing historical travel data before making predictions. It determines acceptable delay times and congestion thresholds in advance, then uses this pre-established information to predict future passenger flow and dropout risks, rather than merely reacting to current sensor readings
Solution Approach 2:
The system introduces historical travel data as an intermediary element between current sensor data and future state predictions. This historical information acts as a mediator that bridges the gap between present observations and future outcomes, enabling more accurate forecasting of passenger behavior and transport conditions
2Reliability
If sensor data from cameras and weight sensors are used to monitor current transport state, then real-time monitoring is achieved, but cost efficiency deteriorates due to expensive sensor deployment
Solution Approach 1:
Instead of deploying expensive physical sensors in every transport, the system creates a virtual copy of the monitoring function by collecting and analyzing existing historical travel data. This data copy enables prediction of passenger flow and transport conditions without requiring costly hardware installation, thereby maintaining reliability while reducing implementation costs
Solution Approach 2:
The system replaces expensive, long-term sensor infrastructure with inexpensive, easily deployable data collection methods. By using readily available historical travel data from existing sources, the system achieves monitoring capabilities at minimal cost, sacrificing the need for expensive permanent sensor installations
3Ease of operation
If dispatch plans are created based on current sensor data assuming all waiting passengers will board, then operational simplicity is maintained, but service quality deteriorates due to overcrowding or underutilization
Solution Approach 1:
The system changes the key parameter from assuming 100% boarding to calculating a dropout risk probability based on historical data. By transforming the boarding assumption into a probabilistic model that considers acceptable delay times and congestion thresholds, the system maintains operational simplicity while significantly improving passenger flow estimation accuracy
Solution Approach 2:
The system incorporates feedback by using historical travel data to continuously refine predictions of passenger dropout risk. This feedback loop allows dispatch planners to adjust their expectations and planning based on actual historical passenger behavior patterns, improving accuracy without complicating the dispatch process
Data Source
Figure 1~2
Figure 3
Figure 4A~4B
AI summary
The present disclosure relates to a method and system for predicting passenger flow in a transportation system. The system receives historical travel data associated with a plurality of passengers of a transport to determine an acceptable delay time and an acceptable transport congestion. The system estimates total number of waiting passengers at a station, based on the historical travel data. The system estimates a delay time and congestion within the transport, based on the historical travel data. The system determines a switching probability to switch the transport, for each of one or more waiting passengers. Further, the system determines a risk of dropping out of the one or more passengers, from boarding the transport, for the station. Thereafter, the system predicts a passenger flow at the station, based on the estimated total number of one or more waiting passengers and the risk associated with the one or more waiting passengers.