Vehicle Position Detection Using Error Covariance Map-Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems face inaccuracies in vehicle position detection due to sensor errors, particularly those related to GPS multipath, wheel slip, and inertial navigation accumulation, which are not effectively addressed by current map-matching technologies that do not consider time-series sensor errors and do not integrate radio navigation methods.
Innovation Solution
A position detection apparatus and program that utilize a GPS receiver, angular speed sensor, and vehicle speed sensor to calculate the vehicle's position and error variance, then perform highly accurate map-matching by selecting the candidate point with the maximum likelihood based on road data and error covariance matrices using a Kalman filter and error ellipses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional map-matching is performed using Euclidean distance from calculated own-vehicle position to road links, then the process is simple and fast, but the map-matched position is largely displaced from actual position due to sensor errors
Solution Approach 1:
The patent changes the parameter used for map-matching from simple Euclidean distance to Mahalanobis distance, which incorporates the error covariance matrix. This parameter transformation allows the system to account for sensor errors and their correlations, significantly improving map-matching accuracy while maintaining computational feasibility through efficient matrix operations.
Solution Approach 2:
The patent introduces the error covariance matrix as an intermediary element that mediates between sensor data and map-matching results. This matrix captures the statistical properties of sensor errors and is used to weight the distance calculation, effectively bridging the gap between raw sensor measurements and accurate position estimation on road networks.
2Measurement precision
If sensor errors are considered in map-matching, then accuracy improves, but computational complexity increases due to error covariance matrix operations
Solution Approach 1:
The patent applies partial action by selectively considering only the most relevant error components in the covariance matrix for map-matching calculations. Rather than processing all possible error sources equally, the system focuses on the dominant error terms that most significantly impact position accuracy, reducing computational overhead while maintaining improvement in accuracy.
3Reliability
If only inertial navigation is used for position calculation, then sensor error accumulation is avoided initially, but long-term accuracy degrades and radio navigation integration is lost
Solution Approach 1:
The patent merges inertial navigation and radio navigation (GPS) methods into a unified position calculation system. By combining the complementary strengths of both methods—the responsiveness and short-term accuracy of inertial navigation with the absolute positioning and long-term stability of GPS—the system achieves reliable short-term performance while maintaining long-term accuracy through sensor fusion.
Solution Approach 2:
The patent implements feedback by continuously updating the error covariance matrix based on incoming sensor measurements and comparing calculated positions with map-matched positions. This feedback loop allows the system to adapt to changing error conditions, correct drift accumulation, and maintain optimal performance over extended periods by adjusting its confidence in different sensor inputs.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly improves the accuracy of vehicle position detection by considering time-series sensor errors and integrating radio and inertial navigation methods, leading to more precise map-matching and reduced displacement from the actual vehicle position.
Implementation Method 1
a GPS receiver (21)
Implementation Method 2
an angular speed sensor (22)
Implementation Method 3
a vehicle speed sensor (23)
Data Source
Figure 1
Figure 2A~2C
Figure 3~4
AI summary
A position detection apparatus 1 includes a storage unit 60 for storing map information including road information; a detection unit 20 for detecting information for calculating an absolute position Xc(t) of a moving object; a position calculation unit 110 for calculating the position Xc(t) and an error variance relating to an error of the object from the detected information; a read-out unit 121 for reading out from the storage unit 60 the road information of a road relating to the position Xc(t); an existence probability calculation unit for calculating an existence probability of the object on the road from the position Xc(t), the error variance, and the read-out road information; a selection unit 123 for selecting a position whose existence probability is maximum out of the existence probability; and a map-matching processing unit 122 for making the selected position a position XI(t) of the object on the road.