Probabilistic lane assignment method

• US 20100076684A1
• Filed: 09/24/2008
• Published: 03/25/2010
• Est. Priority Date: 09/24/2008
• Status: Active Grant

• Current Assignee: Aptiv Technologies AG
• Original Assignee: Delphi Technologies, Inc. (BorgWarner Incorporated)
• Inventors: Schwartz, David A., Schiffmann, Jan K.
• Granted Patent: US 8,055,445 B2
• Time in Patent Office: Days
• US Class Current: 701/301
• CPC Class Codes: B60T 2201/08 Lane monitoring; Lane Keeping Systems
• 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 13/931 of land vehicles
• CPC Class Codes: G01S 2013/932 using own vehicle data, e.g. ground speed, steering wheel direction
• CPC Class Codes: G01S 2013/9321 Velocity regulation, e.g. cruise control
• CPC Class Codes: G01S 2013/93271 in the front of the vehicles

Abstract

An improved probabilistic lane assignment method for detected objects in the scene forward of a host vehicle. Road/lane model parameters, preferably including an angular orientation of the host vehicle in its lane, are estimated from host vehicle sensor systems, taking into account measurement uncertainty in each of the constituent parameters. A probabilistic assignment of the object'"'"'s lane is then assessed based on the road/lane model parameters and object measurements, again taking into account measurement uncertainty in both the road/lane model and object measurements. According to a first embodiment, the probabilistic assignment is discrete in nature, indicating a confidence or degree-of-belief that the detected object resides in each of a number of lanes. According to a second embodiment, the probabilistic assignment is continuous in nature, providing a lateral separation distance between the host vehicle and the object, and a confidence or degree-of-belief in the lateral separation distance.

Claims

1. A probabilistic lane assignment method for an object in a scene forward of a host, where the host is traveling in one lane of a multiple lane road, the method comprising the steps of:

sensing downrange and lateral coordinates of the object relative to a host coordinate system, and estimating a mean of said lateral coordinate and a variance that reflects uncertainty in the sensing of said lateral coordinate;

determining an apparent curvature of the host ego lane, and estimating a mean of said curvature and a variance that reflects uncertainty in the determination of said curvature;

determining a lateral coordinate of the host'"'"'s ego lane center at the obtained downrange coordinate of the object based on at least the estimated mean and variance of said apparent curvature;

determining a width of said one lane, and estimating a mean of said width and a variance that reflects uncertainty in the determination of said width;

combining the means and variances of at least the lateral coordinate of the object, the determined lateral coordinate of the host'"'"'s ego lane center at the downrange coordinate of the object, and the determined lane width to form a lane assignment for the object, said lane assignment including a lane designation and an indication of confidence or degree-of-belief in the lane designation; and

utilizing the calculated lane assignment to assess a threat the object poses to the host.

2. The probabilistic lane assignment method of claim 1, where:

the variance that reflects uncertainty in the determination of said apparent curvature is represented by covariance matrix elements that reflect a difference between said apparent curvature and the estimated mean of said apparent curvature.

3. The probabilistic lane assignment method of claim 2, including the steps of:

low-pass filtering the apparent curvature with a relatively long time constant to estimate the mean of the apparent curvature; and

low-pass filtering the squared difference between said apparent curvature and the estimated mean of said apparent curvature with a relatively short time constant to estimate said covariance matrix elements.

4. The probabilistic lane assignment method of claim 1, including the steps of:

determining an angular orientation of the host within said one lane, and estimating a mean of said angular orientation and a variance that reflects uncertainty in the determination of the angular orientation; and

determining said lateral coordinate of the ego lane center at the downrange coordinate of the object based at least on the estimated means and variances of said apparent curvature and said determined angular orientation.

5. The probabilistic lane assignment method of claim 4, where:

the variance that reflects uncertainty in the determination of said angular orientation is represented by covariance matrix elements that reflect a difference between said angular orientation and the estimated mean of said angular orientation.

6. The probabilistic lane assignment method of claim 5, including the steps of:

low-pass filtering the angular orientation with a relatively long time constant to estimate the mean of the angular orientation; and

low-pass filtering the squared difference between said angular orientation and the estimated mean of said angular orientation with a relatively short time constant to estimate said covariance matrix elements.

7. The probabilistic lane assignment method of claim 1, where:

said lane assignment includes a confidence or degree-of-belief that said object is within a given lane of said road.

8. The probabilistic lane assignment method of claim 7, including the steps of:

determining coordinates of lane markers that bound said lanes, and estimating means of said coordinates and variances that reflect uncertainty in the determination of said lane marker coordinates; and

combining the means and variances of at least the lateral coordinate of the object, the determined lateral coordinate of the ego lane center at the downrange coordinate of the object, and the determined lane marker coordinates to form a lane assignment for the object.

9. The probabilistic lane assignment method of claim 7, including the steps of:

determining a probability P_i that the object is i lanes in a given direction away from the host according to;

$P_i=\underset{j=-1}{\overset{1}{\sum <br><br>}}<br><br>H_j<br><br>O_{j+i}$

where H_j is the probability that the host is j lanes in the given direction away from said one lane, and O_j is the probability that the object is j lanes in the given direction away from said one lane.

10. The probabilistic lane assignment method of claim 9, including the steps of:

forming probability mass distributions for the lane marker coordinates bounding said one lane based on the estimated means and variances of said lane marker coordinates; and

apportioning the probability mass distributions among said one lane and the lanes adjacent said one lane to determine H<sub>−

1</sub>, H₀ and H₊₁.

11. The probabilistic lane assignment method of claim 1, where:

said lane assignment includes a lane number L indicating a lateral offset distance of said object relative to said host, and a confidence or degree-of-belief in said lane number L.

12. The probabilistic lane assignment method of claim 11, including the steps of:

determining an off-center distance of the host from a center of said one lane, and estimating a mean of said off-center distance and a variance that reflect uncertainty in the determination of said off-center distance; and

determining said lane number L by combining the means of the lateral coordinate of the object, the projected lateral coordinate of the host, the off-center distance and the determined lane width, and the variance of the determined lane width.

13. The probabilistic lane assignment method of claim 11, including the step of:

determining a variance of the lane number L to form the confidence or degree-of-belief in said lane number L.

14. The probabilistic lane assignment method of claim 13, where the step of determining the variance of lane number L includes the steps of:

determining an off-center distance of the host from a center of said one lane, and estimating a mean of said off-center distance and a variance that reflect uncertainty in the determination of said off-center distance; and

determining the variance of said lane number L by combining the means and variances of the lateral coordinate of the object, the projected lateral coordinate of the host, the off-center distance and the determined lane width.