Destination Estimation via Area and Link History Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems face accuracy issues when estimating a vehicle's destination, especially when detours are taken or when there is limited history of travel on the current path, leading to reduced estimation accuracy.
Innovation Solution
A destination estimating system that uses a processor and memory to obtain traveling area and link histories, calculating area and link estimation values based on past destinations to accurately estimate the current destination, even when detours are used or stops are made.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If destination estimation is based on frequency of arrival via current traveling path, then estimation can be performed using travel history, but accuracy is reduced when detours are used or when there is limited history
Solution Approach 1:
The patent segments the traveling path into multiple discrete links (road segments) and uses link-level history data to estimate destination probability. This segmentation allows the system to handle detours and unfamiliar paths by evaluating individual link histories rather than requiring complete path matches, thereby improving accuracy when taking detours or using roads not normally traveled.
Solution Approach 2:
The patent introduces a new dimension of analysis by considering not just the overall traveling path but also the specific links (road segments) within the path. By adding link-level detail to the estimation process, the system can accurately estimate destinations even when the overall path differs from historical patterns, resolving the contradiction between using travel history and accommodating detours.
2Measurement precision
If only frequency-based estimation is used, then simple processing is achieved, but estimation accuracy deteriorates when there is no or little history
Solution Approach 1:
The patent merges multiple estimation approaches by combining link-based probability calculation with area-based estimation. The link estimation value is calculated based on historical link data, and the area estimation value is calculated based on the area where the vehicle is currently located. These two values are then combined to determine the final destination, providing accurate estimation even when history is limited while maintaining reasonable processing complexity.
Solution Approach 2:
The patent changes the parameters used for estimation from simple frequency counts to a combination of link-level probability values and area-based probabilities. By transforming the estimation parameters to include both link history and current area information, the system achieves high accuracy even with limited travel history without requiring overly complex processing.
3Measurement precision
If destination estimation relies on past traveling history, then estimation can be performed, but accuracy is reduced when using detours or unfamiliar roads
Solution Approach 1:
The patent makes the estimation system dynamic by continuously updating link estimation values based on current traveling links and area estimation values based on current location. This dynamic approach allows the system to adapt to detours and unfamiliar roads in real-time, improving both accuracy and adaptability to scenarios where the vehicle takes paths different from historical patterns.
Data Source
AI summary
It is an object of the present invention to provide a destination estimating system and a destination estimating method capable of estimating a destination with a high degree of accuracy even when a user uses a detour or stops off somewhere. The destination estimating system of the present invention includes: obtaining a traveling area and a traveling link of a moving body; obtaining, as a traveling history, an area history for storing, for each area, information of a past destination to which the moving body has traveled via the area previously and a link history for storing, for each link, information of a past destination to which the moving body has traveled via the link previously; estimating, based on the traveling area and the area history, a value regarding a probability of traveling to the past destination as an area estimation value; estimating, based on the traveling link and the link history, a value regarding a probability of traveling to the past destination as a link estimation value; and estimating the destination from the past destination based on the area estimation value and the link estimation value.