Target grouping techniques for object fusion

• US 9,563,808 B2
• Filed: 01/14/2015
• Issued: 02/07/2017
• Est. Priority Date: 01/14/2015
• Status: Active Grant

• Current Assignee: GM Global Technology Operations LLC (General Motors Company)
• Original Assignee: GM Global Technology Operations LLC (General Motors Company)
• Inventors: Zeng, Shuqing, O'Dea, Kevin A., Feldman, Dmitriy, Nickolaou, James N.
• Primary Examiner(s): Bitar, Nancy
• Application Number:
  US14/597,108
• Publication Number:
  US 20160203374A1
• Time in Patent Office: 755 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G01S 13/726 Multiple target tracking
• CPC Class Codes: G01S 13/86 Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
• CPC Class Codes: G01S 13/867 Combination of radar systems with cameras
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01S 2013/9316 combined with communication equipment with other vehicles or with base stations
• CPC Class Codes: G01S 2013/9323 Alternative operation using light waves
• CPC Class Codes: G01S 2013/9324 Alternative operation using ultrasonic waves
• CPC Class Codes: G01S 2013/93271 in the front of the vehicles
• CPC Class Codes: G01S 2013/93272 in the back of the vehicles
• CPC Class Codes: G01S 2013/93274 on the side of the vehicles
• CPC Class Codes: G01S 2013/93276 in the windshield area
• CPC Class Codes: G01S 7/41 using analysis of echo signal for target characterisation; Target signature; Target cross-section
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30252 Vehicle exterior; Vicinity of vehicle
• CPC Class Codes: G06T 7/75 involving models
• CPC Class Codes: G06T 7/77 using statistical methods
• CPC Class Codes: G06V 10/44 : Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 20/588 : Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road

Abstract

A method is disclosed for improved target grouping of sensor measurements in an object detection system. The method uses road curvature information to improve grouping accuracy by better predicting a new location of a known target object and matching it to sensor measurements. Additional target attributes are also used for improved grouping accuracy, where the attributes includes range rate, target cross-section and others. Distance compression is also employed for improved grouping accuracy, where range is compressed in a log scale calculation in order to diminish errors in measurement of distant objects. Grid-based techniques include the use of hash tables and a flood fill algorithm for improved computational performance of target object identification, where the number of computations can be reduced by an order of magnitude.

Claims

1. A method for grouping object sensor measurements with target objects in an object detection system, said method comprising:

providing a list of target objects being tracked by the object detection system, where the list of target objects includes known targets identified by the object detection system in an area ahead of a host vehicle;

computing hypothesis locations and orientations for each known target in the list of target objects, where the hypothesis locations and orientations include a prediction of each known target'"'"'s movement since the list of target objects was previously computed, where computing hypothesis locations and orientations for each known target includes using digital map data to predict a location and orientation of each known target based on road curvature data and target velocity;

providing sensor measurement points from at least one object sensing system, where the sensor measurement points designate points at which an object has been detected in the area ahead of the host vehicle;

grouping, using a microprocessor, the sensor measurement points with the known targets at the hypothesis locations and orientations;

validating the hypothesis locations and orientations based on the grouping;

identifying new targets based on any clusters of sensor measurement points which do not correlate to one of the known targets; and

updating the list of target objects to include the known targets at the hypothesis locations and orientations, and any new targets identified.

2. The method of claim 1 wherein the list of target objects includes, for each of the known targets, attribute data including micro-Doppler data, target cross-section, signal-to-noise ratio in measurement points, and visual attributes if available from a camera-based object sensing system.

3. The method of claim 1 wherein the list of target objects includes, for each of the known targets, a range and azimuth angle relative to the host vehicle, and a range rate relative to the host vehicle.

4. The method of claim 3 wherein grouping the sensor measurement points with the known targets includes comparing both the range and the range rate of the points and the targets to establish correlations.

5. The method of claim 1 wherein computing hypothesis locations and orientations for each known target includes using object data provided by one or more sources external to the host vehicle, including object data provided by other vehicles and object data detected and provided by roadside infrastructure.

6. The method of claim 1 wherein grouping the sensor measurement points with the known targets includes using a grid-based technique which assigns each of the measurement points to a grid cell based on location, identifies a subset of grid cells in a neighborhood of each of the known targets, and groups the measurement points in the identified subset of grid cells with the known target.

7. The method of claim 6 wherein the grid-based technique includes identifying the subset of grid cells in the neighborhood of each of the known targets by establishing a nine-cell square of grid cells surrounding a centroid of each of the known targets.

8. The method of claim 6 wherein the grid-based technique includes using a flood fill algorithm to identify the subset of grid cells in the neighborhood of each of the known targets, where the known targets can have any arbitrary shape and orientation.

9. The method of claim 1 wherein providing sensor measurement points from at least one object sensing system includes providing sensor measurement points from a radar-based object sensing system and from a camera-based object sensing system.

10. The method of claim 1 further comprising using the list of target objects in a collision warning system on the host vehicle.

11. A method for grouping object sensor measurements with target objects in an object detection system, said method comprising:

providing a list of target objects being tracked by the object detection system, where the list of target objects includes known targets identified by the object detection system in an area ahead of a host vehicle;

computing hypothesis locations and orientations for each known target in the list of target objects, where the hypothesis locations and orientations include a prediction of each known target'"'"'s movement since the list of target objects was previously computed based on road curvature data and target velocity;

providing sensor measurement points from at least one object sensing system, where the sensor measurement points designate points at which an object has been detected in the area ahead of the host vehicle;

grouping, using a microprocessor, the sensor measurement points with the known targets at the hypothesis locations and orientations, including comparing both a range and a range rate of the points and the targets to establish correlations, and further including using a mapped range value for the measurement points and the known targets, where the mapped range value is computed from an actual range value using a logarithmic scale;

validating the hypothesis locations and orientations based on the grouping;

identifying new targets based on any clusters of sensor measurement points which do not correlate to one of the known targets; and

updating the list of target objects to include the known targets at the hypothesis locations and orientations, and any new targets identified.

12. The method of claim 11 wherein grouping the sensor measurement points with the known targets includes using a grid-based technique which assigns each of the measurement points to a grid cell based on location, identifies a subset of grid cells in a neighborhood of each of the known targets, and groups the measurement points in the identified subset of grid cells with the known target.

13. An object detection system comprising:

at least one object sensing system onboard a host vehicle, said object sensing system providing sensor measurement points which designate points at which an object has been detected in an area ahead of the host vehicle;

a memory module for storing a list of target objects being tracked by the object detection system, where the list of target objects includes known targets identified by the object detection system in the area ahead of the host vehicle; and

an object detection processor in communication with the memory module and the at least one object sensing system, said object detection processor being configured to;

compute hypothesis locations and orientations for each known target in the list of target objects, where the hypothesis locations and orientations include a prediction of each known target'"'"'s movement since the list of target objects was previously computed, wherein computing hypothesis locations and orientations for each known target by the processor includes using digital map data to predict a location and orientation of each known target based on road curvature data and target velocity;

group the sensor measurement points with the known targets at the hypothesis locations and orientations;

validate the hypothesis locations and orientations based on the grouping;

identify new targets based on any clusters of sensor measurement points which do not correlate to one of the known targets; and

update the list of target objects to include the known targets at the hypothesis locations and orientations, and any new targets identified.

14. The object detection system of claim 13 wherein the list of target objects includes, for each of the known targets, attribute data including micro-Doppler data, target cross-section, signal-to-noise ratio in measurement points, and visual attributes if available from a camera-based object sensing system.

15. The object detection system of claim 13 wherein the list of target objects includes, for each of the known targets, a range and azimuth angle relative to the host vehicle, and a range rate relative to the host vehicle, and where grouping the sensor measurement points with the known targets by the processor includes comparing both the range and the range rate of the points and the targets to establish correlations.

16. The object detection system of claim 13 wherein grouping the sensor measurement points with the known targets by the processor includes using a grid-based technique which assigns each of the measurement points to a grid cell based on location, identifies a subset of grid cells in a neighborhood of each of the known targets, and groups the measurement points in the identified subset of grid cells with the known target.

17. The object detection system of claim 16 wherein the grid-based technique includes using a flood fill algorithm to identify the subset of grid cells in the neighborhood of each of the known targets, where the known targets can have any arbitrary shape and orientation.

18. The object detection system of claim 13 wherein grouping the sensor measurement points with the known targets by the processor includes using a mapped range value for the measurement points and the known targets, where the mapped range value is computed from an actual range value using a logarithmic scale.