METHOD AND SYSTEM FOR DETECTING, TRACKING AND ESTIMATING STATIONARY ROADSIDE OBJECTS

• US 20150293216A1
• Filed: 04/15/2014
• Published: 10/15/2015
• Est. Priority Date: 04/15/2014
• Status: Active Grant

• Current Assignee: GM Global Technology Operations LLC (General Motors Company)
• Original Assignee: GM Global Technology Operations LLC (General Motors Company)
• Inventors: O'DEA, Kevin A., ZENG, Shuqing, NICKOLAOU, James N., HAMNER, Bradley
• Granted Patent: US 9,834,207 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: B60W 2420/408 Radar; Laser, e.g. lidar
• CPC Class Codes: B60W 2420/60 Doppler effect
• CPC Class Codes: B60W 2520/10 Longitudinal speed
• CPC Class Codes: B60W 2520/105 Longitudinal acceleration
• CPC Class Codes: B60W 30/02 Control of vehicle driving stability
• CPC Class Codes: B60W 30/10 Path keeping cruise control for automatically following a preceding vehicle B60W30/165
• CPC Class Codes: B60W 30/12 Lane keeping
• CPC Class Codes: B60W 30/14 Adaptive cruise control
• CPC Class Codes: B60W 40/06 Road conditions
• CPC Class Codes: G01S 13/726 Multiple target tracking
• CPC Class Codes: G01S 13/867 Combination of radar systems with cameras
• CPC Class Codes: G01S 13/87 Combinations of radar systems, e.g. primary radar and secondary radar
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01S 2013/932 using own vehicle data, e.g. ground speed, steering wheel direction
• CPC Class Codes: G01S 2013/9323 Alternative operation using light waves
• CPC Class Codes: G01S 2013/93271 in the front of the vehicles

Abstract

A system and method for selectively reducing or filtering data provided by one or more vehicle mounted sensors before using that data to detect, track and/or estimate a stationary object located along the side of a road, such as a guardrail or barrier. According to one example, the method reduces the amount of data by consolidating, classifying and pre-sorting data points from several forward looking radar sensors before using those data points to determine if a stationary roadside object is present. If the method determines that a stationary roadside object is present, then the reduced or filtered data points can be applied to a data fitting algorithm in order to estimate the size, shape and/or other parameters of the object. In one example, the output of the present method is provided to automated or autonomous driving systems.

Claims

1. A method for evaluating roadside stationary objects, comprising the steps of:

receiving sensor readings from one or more vehicle mounted sensor(s), the sensor readings include a plurality of data points and each data point corresponds to an object that is within a field-of-view of one of the sensors;

consolidating the plurality of data points into a smaller set of consolidated data points;

classifying the consolidated data points into a smaller set of classified data points, the classified data points correspond to one or more stationary objects;

pre-sorting the classified data points into first and second smaller sets of pre-sorted data points, the first set of pre-sorted data points correspond to one or more stationary objects located only on a left side of the road and the second set of pre-sorted data points correspond to one or more stationary objects located only on a right side of the road;

determining if there is one or more stationary objects present on the left side of the road by using the first set of pre-sorted data points, and determining if there is one or more stationary objects present on the right side of the road by using the second set of pre-sorted data points; and

when there is a stationary object present on the left side of the road then applying the first set of pre-sorted data points to a data fitting algorithm that estimates a size, shape and/or location of the stationary object, and when there is a stationary object present on the right side of the road then applying the second set of pre-sorted data points to a data fitting algorithm that estimates a size, shape and/or location of the stationary object.

2. The method of claim 1, wherein each data point corresponds to an object that is within a field-of-view of one of the sensors and is provided in the form of coordinates that indicate the relative position of the object to the sensor.

3. The method of claim 1, wherein the consolidating step further comprises consolidating the plurality of data points into a smaller set of consolidated data points by reducing the number of data points through the application of one or more of the following technique(s):

a clustering technique, a grouping technique, or a mapping technique.

4. The method of claim 1, wherein the classifying step further comprises classifying the consolidated data points into a smaller set of classified data points by using Doppler measurements in conjunction with dynamic information gathered from sensors on the host vehicle to determine if each data point corresponds to a stationary object.

5. The method of claim 1, wherein the classifying step further comprises classifying the consolidated data points into a first set of classified data points that correspond to stationary objects and a second set of classified data points that correspond to moving objects, providing the second set of classified data points to a separate target tracking algorithm, and then discarding the second set of classified data points so that only the first set of classified data points are provided to the pre-sorting step.

6. The method of claim 1, wherein the pre-sorting step further comprises identifying a classified data point that corresponds to a left side of the road and adding that data point to the first set of pre-sorted data points, and identifying a classified data point that corresponds to a right side of the road and adding that data point to the second set of pre-sorted data points.

7. The method of claim 6, wherein the pre-sorting step further comprises comparing a lateral position of a classified data point to a left side offset and when the lateral position is left of the left side offset then adding that data point to the first set of pre-sorted data points, and comparing a lateral position of a classified data point to a right side offset and when the lateral position is right of the right side offset then adding that data point to the second set of pre-sorted data points.

8. The method of claim 1, wherein the determining step is carried out separately to determine if there is one or more stationary objects present on the left side of the road and to determine if there is one or more stationary objects present on the right side of the road.

9. The method of claim 1, wherein the pre-sorting step or the determining step further comprises determining if a data point was sensed during a current sensor cycle (new data point) or a past sensor cycle (old data point), and discarding the data point as an old data point if it was sensed during a past sensor cycle.

10. The method of claim 1, wherein the pre-sorting step or the determining step further comprises comparing a lateral position of a data point to a lateral threshold and comparing a longitudinal position of the same data point to a longitudinal threshold, and discarding the data point as an outlier if either the lateral position is not within the lateral threshold or the longitudinal position is not within the longitudinal threshold.

11. The method of claim 1, wherein the pre-sorting step or the determining step further comprises statistically validating a group of data points as a whole.

12. The method of claim 11, wherein the pre-sorting step or the determining step further comprises statistically validating the group of data points as a whole by comparing a lateral position variance for the group to a group threshold, and discarding data points if the lateral position variance for the group is greater than the group threshold.

13. The method of claim 1, wherein the pre-sorting step or the determining step further comprises statistically validating each data point in a group of data points individually.

14. The method of claim 13, wherein the pre-sorting step or the determining step further comprises statistically validating each data point individually by comparing a lateral position value Y for each data point to an individual threshold, and discarding the data point if the lateral position value Y is greater than the individual threshold, the lateral position value Y=((Lateral Position of Data Point-Average)²/Variance).

15. The method of claim 1, wherein the pre-sorting step or the determining step further comprises combining a group of statistically validated data points gathered during a current sensor cycle (new data points) with a group of statistically validated data points previously gathered during one or more previous sensor cycles (old data points) into a group of combined data points, wherein the combined data points corresponds to objects on either the left side of the road or the right side of the road.

16. The method of claim 15, wherein the pre-sorting step or the determining step further comprises sorting the group of combined data points according to their longitudinal position.

17. The method of claim 1, wherein the pre-sorting step or the determining step further comprises identifying data points that are located longitudinally ahead of the host vehicle, and adding such data points to a data point list (New Data Point List (I)).

18. The method of claim 17, wherein if there are no data points that are located ahead of the host vehicle, then the determining step further comprises determining that there are no stationary objects present on either left side of the road or the right side of the road.

19. The method of claim 1, wherein the pre-sorting step or the determining step further comprises determining if an overall longitudinal extent or distance (D_LONG) of data points is greater than a longitudinal threshold, and discarding data points if the overall longitudinal extent or distance (D_LONG) is not greater than the longitudinal threshold.

20. The method of claim 1, wherein the pre-sorting step or the determining step further comprises determining if an overall lateral extent or distance (D_LAT) of data points is less than a lateral threshold, and discarding data points if the overall lateral extent or distance (D_LAT) is not less than the lateral threshold.

21. The method of claim 1, wherein the pre-sorting step or the determining step further comprises calculating longitudinal and lateral inter-data point spacing of data points and comparing the inter-data point spacing to longitudinal and lateral spacing thresholds, and discarding data points if the longitudinal inter-data point spacing is greater than the longitudinal spacing threshold or if the lateral inter-data point spacing is greater than the lateral spacing threshold.

22. The method of claim 1, wherein the determining step further comprises comparing the number of data points in a data point list (New Data Point List (II)) to a point threshold, and determining that there is a stationary object present on either the left side of the road or the right side of the road when the number of data points in the data point list is greater than the point threshold.

23. The method of claim 1, wherein the determining step further comprises employing a chi-square test to determine if a stationary object is present.

24. The method of claim 1, wherein the determining step further comprises evaluating a timer that corresponds to a previously detected old stationary object from a past sensor cycle to determine if the old stationary object is still present or if the host vehicle has driven past the old stationary object.

25. The method of claim 1, wherein the applying step further comprises applying at least one of the first or second sets of pre-sorted data points to a data fitting algorithm that estimates a size, shape and/or location of a stationary guardrail located on either the left or the right side of the road.

26. The method of claim 1, further comprising the step of:

providing the output of the data fitting algorithm to a separate algorithm that controls one or more aspects of operating the vehicle in an automated driving mode.

27. The method of claim 1, further comprising the step of:

providing a visual representation to the driver that includes a grid or area of interest surrounding the host vehicle, wherein the output of the data fitting algorithm is represented in the grid and is updated as the host vehicle drives along the road.