Map Matching Position Calculation Using Azimuth Error Covariance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing position calculation methods for mobile objects fail to accurately handle errors in estimated position and azimuth due to GPS signal quality and map variations, leading to mismatching issues at narrow angles or near parallel roads.
Innovation Solution
A method that calculates the current position of a mobile object by evaluating the probability of being on a specific link based on error covariances, distance, and azimuthal differences between the object's position and link candidate points, using a combination of data items at predetermined intervals to determine the most likely road and adjust for errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If map matching is implemented using only estimated position error, then the position calculation is simple, but mismatching occurs at narrow angle forks and parallel roads due to unaccounted azimuth errors and link data variations
Solution Approach 1:
The invention changes the parameters considered in map matching from only position error to include both position error and azimuth error. By incorporating azimuth error covariance and link azimuth error variance into the evaluation value calculation, the system accurately distinguishes between true road deviations and measurement errors, particularly at narrow angle forks and parallel roads where azimuth differences are critical.
2Measurement precision
If error variances are calculated and used in evaluation, then matching accuracy improves, but calculation complexity increases
Solution Approach 1:
The invention performs preliminary calculation of error covariance matrices and error variances before the map matching evaluation. By pre-computing the position error covariance, azimuth error covariance, and link error variances, the system prepares all necessary statistical parameters in advance, making the actual matching process more efficient despite the sophisticated error modeling.
Solution Approach 2:
The invention replaces simple geometric distance comparison with a statistical probability model based on error covariances and variances. Instead of mechanically checking if the vehicle is within a fixed distance threshold of a road link, the system uses probabilistic evaluation values that account for uncertainties in both vehicle position/azimuth and link position/azimuth measurements.
3Adaptability or versatility
If GPS signal quality varies and traveling route affects position accuracy, then position error increases, but the system should adapt to maintain matching accuracy
Solution Approach 1:
The invention makes the error covariance matrices dynamic rather than static. The position error covariance and azimuth error covariance are updated based on current GPS signal quality and traveling route conditions. This allows the system to adapt to varying measurement uncertainties in real-time, maintaining accurate map matching even when GPS quality degrades or the vehicle traverses areas with poor satellite visibility.
Data Source
AI summary
A position calculation method includes: by using a current position of a mobile object, a link candidate point position, an error variance of positions of a plurality of links included in a region around the current position, and an error variance of azimuths of the plurality of links, calculating, an evaluation value that corresponds to a probability that the mobile object is traveling upon a road corresponding to each link candidate point, for each link including the each link candidate point; and calculating the current position based upon the evaluation value, by taking the mobile object as being positioned at a link candidate point, among a plurality of link candidate points, for which the probability is highest.