Probe Data Preprocessing for GPS Error Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic map systems face inaccuracies due to non-Gaussian errors in GPS positional measurements, particularly in urban areas where systematic errors from obstructions and multi-path phenomena occur, affecting the precision of road centerline inference from probe trace data.
Innovation Solution
A method to identify and remove non-random noise from probe trace data, replacing it with pseudo data points that conform to the data curve, allowing for more accurate street centerline calculations using Gaussian statistical averaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS positional measurements are used in urban areas with obstructions and multi-path phenomena, then probe trace data can be collected for road centerline inference, but systematic non-Gaussian errors reduce measurement precision
Solution Approach 1:
The patent identifies and extracts anomalous data points that exhibit non-Gaussian error patterns from the probe trace dataset. By detecting these outliers using statistical methods and removing them before road centerline calculation, the systematic errors caused by urban obstructions and multi-path phenomena are eliminated, improving the precision of positional measurements used for mapping
Solution Approach 2:
The patent implements a feedback mechanism where the statistical properties of the probe trace data are continuously analyzed to identify deviations from Gaussian distribution. This feedback loop allows the system to detect and correct for systematic errors in real-time, adjusting the data processing to compensate for urban environment effects and maintain measurement precision
2Measurement precision
If statistical averaging of multiple probe traces is used to improve road centerline accuracy, then precision improves, but non-Gaussian errors violate the assumptions of Gaussian statistics
Solution Approach 1:
The patent applies preliminary preprocessing actions to the probe trace data before statistical averaging. By detecting and removing anomalous data points that violate Gaussian assumptions through statistical analysis and outlier detection, the data is prepared in advance to meet the requirements of Gaussian statistics, ensuring that subsequent averaging operations are reliable and valid
Solution Approach 2:
The patent changes the statistical parameters of the dataset by removing outliers and adjusting the distribution characteristics. This transformation modifies the data to better conform to Gaussian distribution assumptions, making the statistical averaging method applicable and reliable for road centerline inference even in environments where systematic errors were present
Data Source
Figure 1~2
Figure 3
Figure 4a~4b
AI summary
Methods of pre-processing probe trace data include examining a sequential dataset for a route of travel between two points, identifying at least one anomaly in the sequential dataset, identifying at least one datum of the sequential dataset occupying a sequential position adjacent to the identified at least one anomaly, and inserting at least one non-anomalous datum to occupy a sequential position of the identified at least one anomaly in the sequential dataset, the at least one non-anomalous datum being determined based on one of extrapolation, smoothing and interpolation of the at least one datum.