Methods of pre-processing probe data
AssigneeTomTom North America, Inc.
InventorsClayton Richard Morlock, Shaundrea Margaret Newman Kenyon
Life of this patent
Estimating the term…
Abstract
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.
Claims24
A method of pre-processing probe data, the method comprising:
- examining a sequential dataset for a route of travel between two points;
- identifying at least one anomaly in the sequential dataset;
- and
- identifying at least a first and a second datum of the sequential dataset, wherein said first datum occupies 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 first and second datum.
(canceled)
(canceled)
14-15. (canceled)
A method of pre-processing probe data, the method comprising:
- examining a plurality of sequential location measurements for a route of travel between two points:
- fitting a best fit mathematical curve to the plurality of sequential location measurements;
- identifying at least one anomaly by identifying at least one of the plurality of sequential location measurements that does not fall near the best fit mathematical curve within a statistical threshold;
- removing the identified at least one anomaly;
- and
- inserting sequential location data corresponding to the best fit mathematical curve in place of the anomaly.
25-29. (canceled)
Dependencies are inferred from the claim text.