Search

EP-3237922-B1 - EXTRACTING FEATURE GEOMETRIES FOR LOCALIZATION OF A DEVICE

EP3237922B1EP 3237922 B1EP3237922 B1EP 3237922B1EP-3237922-B1

Inventors

  • MODICA, Leo
  • STENNETH, Leon

Dates

Publication Date
20260506
Application Date
20151218

Claims (13)

  1. A method for using one or more processors, comprising: receiving (S101), from a device configured to collect three-dimensional depth map data, a three-dimensional depth map of a location in a path network, wherein the pathway network comprises a road network; identifying a physical structure within the depth map; dividing (S105) the three-dimensional depth map, at a physical structure, into one or more horizontal slices at an elevation from a road level; and extracting (S107) a two-dimensional feature geometry from the one or more horizontal slices using a linear regression algorithm or a curvilinear regression algorithm, georeferencing points of the extracted feature geometry to a geographic location for determination of the geographic location of the device, wherein each point is matched with location reference information for latitude, longitude, and at least one of elevation from the road level, and altitude.
  2. The method of claim 1, further comprising: encoding the georeferenced points of the extracted feature geometry in a fingerprint database.
  3. The method of claim 1 or claim 2, wherein the two-dimensional feature geometry is a line or a set of connected lines, and the extracting uses the linear regression algorithm, further comprising: georeferencing points of the line or the set of connected lines to a geographic location, wherein each point is matched with location reference information for latitude, longitude, elevation from the road level, and altitude.
  4. The method of claim 1 or claim 2, wherein the two-dimensional feature geometry is a curve, arc or spline, and the extracting uses the curvilinear regression algorithm.
  5. The method of claim 4, further comprising: georeferencing a radius of the arc and points along the arc to a geographic location, wherein each point is matched with location reference information for latitude, longitude, elevation from the road level, and altitude.
  6. The method of claim 5, further comprising: georeferencing knots of the spline and points along the spline to a geographic location, wherein each point is matched with location reference information for latitude, longitude, elevation from the road level, and altitude.
  7. The method of any preceding claim, wherein receiving the three-dimensional depth map comprises collecting (S201), by an end-user device, the three-dimensional depth map at the location in the path network, further comprising: identifying, using a processor of the end-user device, a physical structure within the three-dimensional depth map, and wherein the depth map three-dimensional is divided, at the physical structure, into a horizontal plane at an elevation from a road level.
  8. The method of claim 7, further comprising: transmitting the extracted feature geometry to an external processor; and receiving, from the external processor (140), a geographic location of the end-user device (201) or selected portion of feature geometries through a comparison of the feature geometry and a database of feature geometries for the path network, wherein the geographic location of the end-user device (201) is determined using measured distances between control point data of at least three control points of the selected portion of feature geometries and the end-user device.
  9. The method of any preceding claim, wherein each of the one or more horizontal slices has a thickness of 0.1-2 meters.
  10. The method of any preceding claim, wherein the one or more horizontal slices includes data at a defined, selected elevation, as well as data that exists within a range above and below the selected elevation.
  11. The method of any preceding claim, comprising: analyzing a first horizontal plane at a first elevation to identify two-dimensional feature geometries; extracting two-dimensional feature geometries from a second horizontal plane at a second elevation if no useful two-dimensional geometric features are identified in the first horizontal plane.
  12. An apparatus (100; 201) comprising: at least one processor (140; 205); and at least one memory (144; 209) including computer program code for one or more programs (148); the at least one memory (144; 209) and the computer program code configured to, with the at least one processor (140; 205), cause the apparatus (100; 201) to at least perform the method of any preceding claim.
  13. A computer-readable medium (120) including computer program code for one or more programs; the computer program code configured to cause at least one processor to at least perform the method of any of claims 1 to 11.

Description

FIELD The following disclosure relates to developing a fingerprint database and determining the geographic location of an end-user device (e.g., vehicle, mobile phone, smart watch, etc.) using a multilateration calculation. BACKGROUND Vehicle localization using Global Positioning Systems (GPS), local area wireless technology (e.g., WiFi), and short-wavelength radio waves (e.g., Bluetooth) may be imprecise. In the case of GPS, multi-pathing causes timing variations for signals received by the vehicle. In the case of WiFi and Bluetooth, signal strengths are an unreliable means for localization due primarily to occlusion and lack of precision in the location of the transmitting stations in three-dimensional (3D) space. In such cases, the references upon which multilateration is performed are not precise enough to produce lane level, or in some case road level, positioning. US 2014/225771 (A1) relates to a method in which the position of a mobile device is determined based, at least on part, of the relative locations of two or more nearby points. Distance data, received from a range finding sensor, corresponds to distances to the two or more nearby points from the mobile device. A predetermined model that includes previously recorded locations for objects is accessed to receive location data for two or more nearby points. A position of the mobile device is calculated based on the location data and the distance data. The nearby points may be building edges or building corners. The calculation may involve a series of equations. The series of equations may include satellite-based positioning equations in addition to equations based on the predetermined model. WO 2011/ 023246 A1 discloses a vehicle navigation or mapping system comprising at least one sensor located on or in a vehicle and adapted to perform measurements to obtain sensor measurement data, a data store for storing reference sensor measurement data; and a processing resource adapted to determine whether the sensor measurement data matches the reference sensor measurement data and, if the sensor measurement data matches the reference sensor measurement data, to determine a relative or absolute spatial location of the vehicle from stored location data associated with the stored reference sensor measurement data. SUMMARY OF THE INVENTION The present invention is defined by the claims. SUMMARY OF THE DISCLOSURE Example systems, apparatuses, and methods are provided for developing a fingerprint database and extracting feature geometries for determining the geographic location of a device are disclosed herein. In one example, the method comprises receiving a depth map of a location in a path network. The method further comprises identifying, using a processor, a physical structure within the depth map. The method further comprises dividing the depth map, at the physical structure, into a horizontal plane at an elevation from a road level. The method further comprises extracting a two-dimensional feature geometry from the horizontal plane using a linear regression algorithm, a curvilinear regression algorithm, or a machine learning algorithm. In another example, the method comprises collecting, by an end-user device, a depth map at a location in a path network. The method further comprises identifying, using a processor of the end-user device, a physical structure within the depth map. The method further comprises dividing the depth map, at the physical structure, into a horizontal plane at an elevation from a road level. The method further comprises extracting a two-dimensional feature geometry from the horizontal plane using a linear regression algorithm, a curvilinear regression algorithm, or a machine learning algorithm. In yet another example, the apparatus comprises at least one processor and at least one memory including computer program code for one or more programs; the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to at least perform: (1) receive a depth map of a location in a path network; (2) identify a physical structure within the depth map; (3) divide the depth map, at the physical structure, into a horizontal plane at an elevation from a road level; and (4) extract a two-dimensional feature geometry from the horizontal plane using a linear regression algorithm, a curvilinear regression algorithm, or a machine learning algorithm. BRIEF DESCRIPTION OF THE DRAWINGS Exemplary examples are described herein with reference to the following drawings. FIG. 1 illustrates an example of a depth map image with extracted horizontal slices at multiple elevations, and identified two-dimensional images from the extracted slices.FIG. 2 illustrates an example of an encoded fingerprint in a fingerprint database.FIG. 3 illustrates an example of a line feature geometry of the encoded fingerprint of FIG. 2.FIG. 4 illustrates an example of an arc feature geometry of the encoded fingerprint of FIG