Trip Duration Estimation Using Real-Time and Historical Data Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated methods for generating trip duration estimates are not always accurate, as they rely on incomplete or non-existent graphs, assume specific routes, and do not account for extra travel time at intersections or transitions between road classes, leading to inaccurate predictions.
Innovation Solution
A computer-implemented method that combines real-time and historical trip data using machine learning to estimate trip durations by training models iteratively based on recent trip data, adjusting initial predictions, and determining weighting factors to improve accuracy, incorporating features like road segments, traffic conditions, and provider characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional graph-based methods are used for trip duration estimation, then the estimation process is simple and fast, but the accuracy is poor due to incomplete graphs and failure to account for real-time conditions
Solution Approach 1:
The system segments trip duration estimation into multiple independent models: a route prediction model that determines the path, a trip duration prediction model that estimates time, and a selector model that chooses between real-time and historical data sources. Each model specializes in one aspect, improving overall accuracy while maintaining manageable complexity through modular design.
Solution Approach 2:
The patent introduces a selector model as an intermediary that mediates between real-time trip data and historical trip data. This intermediary evaluates current conditions and determines which data source should inform the prediction, enabling the system to adapt to varying conditions without requiring complete retraining of all models.
2Reliability
If real-time data is used for trip duration estimation, then current conditions are captured, but the system may be affected by anomalies and incomplete data
Solution Approach 1:
The system merges real-time trip data with historical trip data through a weighted combination approach. The selector model determines the appropriate weighting based on current conditions, allowing the system to leverage both the currency of real-time data and the completeness of historical data, thereby improving reliability while compensating for data limitations.
Solution Approach 2:
The system prepares historical trip data in advance as a backup and reference source. When real-time data is incomplete or anomalous, the pre-prepared historical data provides a cushion that prevents prediction failures, ensuring continuous reliable operation even when real-time data quality is poor.
3Loss of information
If historical data is used for trip duration estimation, then data completeness is improved, but the estimates do not reflect current traffic conditions
Solution Approach 1:
The system dynamically adjusts the balance between historical and real-time data based on current conditions. The selector model continuously evaluates whether to prioritize historical patterns or real-time observations, making the system adaptable to changing traffic conditions while maintaining the benefits of historical data completeness.
Solution Approach 2:
The patent changes the parameter of data source selection based on trip characteristics and current conditions. By adjusting which data source (historical or real-time) is weighted more heavily according to specific trip parameters and environmental conditions, the system achieves both data completeness and adaptability to current situations.
4Measurement precision
If multiple models are used to combine real-time and historical data, then estimation accuracy is improved, but computational complexity increases
Solution Approach 1:
The complex estimation task is segmented into specialized models: one for route prediction, one for trip duration prediction, and one for selecting between data sources. This segmentation improves accuracy by allowing each model to focus on its specific function while keeping individual model complexity manageable through clear separation of concerns.
Data Source
AI summary
A system uses machine models to estimate trip durations or distance. The system trains a historical model to estimate trip duration using characteristics of past trips. The system trains a real-time model to estimate trip duration using characteristics of recently completed trips. The historical and real-time models may use different time windows of training data to predict estimates, and may be trained to predict an adjustment to an initial trip estimate. A selector model is trained to predict whether the historical model, the real-time model, or a combination of the historical and real-time models will more accurately estimate a trip duration, given features associated with a trip duration request, and the system accordingly uses the models to estimate a trip duration. In some embodiments, the real-time model and the selector may be trained using batch machine learning techniques which allow the models to incorporate new trip data as trips complete.