METHOD FOR REDUCING FALSE ACTIVATIONS IN REVERSE COLLISION AVOIDANCE SYSTEMS

• US 20150210293A1
• Filed: 01/23/2015
• Published: 07/30/2015
• Est. Priority Date: 01/29/2014
• Status: Active Grant

• Current Assignee: Continental Autonomous Mobility US LLC (Continental AG)
• Original Assignee: Continental Automotive Systems, Inc. (Continental AG)
• Inventors: Fletcher, Graham Lanier
• Granted Patent: US 9,493,170 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: B60W 2050/146 Display means
• CPC Class Codes: B60W 2554/00 Input parameters relating to objects
• CPC Class Codes: B60W 30/0956 the prediction being responsive to traffic or environmental parameters
• CPC Class Codes: B60W 50/14 Means for informing the driver, warning the driver or prompting a driver intervention
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

A collision avoidance system for a vehicle is disclosed that includes at least one sensor mounted to a vehicle to measure the vehicle'"'"'s environment in the reverse driving direction, a controller which creates a feature vector based on relationships among data provided by at least one sensor, and a non-volatile memory which stores feature vectors and related object configurations for a number of different environments. The system utilizes a comparison between the current feature vector and previously stored feature vectors to match the current environment with a previously observed environment, and then loads a previously stored object configuration which can improve collision detection performance to reduce the number of unnecessary system interventions.

Claims

1. A reverse collision avoidance system for a vehicle comprising:

at least one sensor mounted to the vehicle to detect objects proximate to a path of the vehicle;

a controller including instructions for;

analyzing data from the sensors to determine a current feature vector based upon data from the at least one sensor;

storing the current feature vector and associated object configuration for the environment in non-volatile memory for the collision avoidance system;

comparing the current feature vector with previously stored feature vectors; and

selecting an object configuration representing the environment surrounding the vehicle based on the comparison of the current feature vectors with previously stored feature vectors.

2. The reverse collision avoidance system of claim 1, wherein the controller includes instructions for:

comparing the current feature vector to previous feature vectors stored in a memory to determine similarities between the current feature vector and any of the previous feature vectors; and

selecting the object configuration from stored object configurations that corresponds with a previous feature vector that matches the current feature vector.

3. The reverse collision avoidance system of claim 1, wherein the controller further includes instructions for determining a probability of collision with the detected object.

4. The reverse collision avoidance system of claim 2, wherein the probability that the detected object is an obstacle is decreased when the current feature vector is matched to a previous feature vector and the detected object was previously identified as not an obstacle.

5. The reverse collision avoidance system of claim 2, wherein when a previous feature vector is determined to be a match with a current feature vector, a relevance value for the previous feature vector is increased.

6. The reverse collision avoidance system as recited in claim 1, wherein the current feature vector comprises a relationship between proximate objects.

7. The reverse collision avoidance system as recited in claim 1, wherein the at least one sensor comprises a camera mounted within the vehicle.

8. The reverse collision avoidance system as recited in claim 1, wherein the at least one sensor comprises a proximity sensor mounted to the vehicle.

9. A collision avoidance system for a vehicle comprising:

at least one sensor mounted to the vehicle to view a vehicle path for the vehicle;

a controller which analyzes data from the sensors to determine a current feature vector based upon relationships among data detected by the at least one sensor; and

a non-volatile memory which stores the current feature vector and associated object configuration in the collision avoidance system, wherein a probability that an object is an obstacle within the vehicle path is modified based upon a comparison between the current feature vector and a previous feature vector and an object configuration that corresponds with the previous feature vector.

10. The collision avoidance system as recited in claim 9, wherein, the controller compares the current feature vector to previous feature vectors stored in the memory to determine if the current feature vector includes similar characteristics with any of the previous feature vectors.

11. The collision avoidance system of claim 9, wherein the at least one sensor includes a camera which provides a view of the vehicle path, wherein the controller analyzes an image from the camera to identify objects proximate to the vehicle.

12. The collision avoidance system of claim 9, wherein the at least one sensor comprises a proximity sensor.

13. The collision avoidance system of claim 9, wherein the probability that the object is an obstacle to be avoided is decreased when the current feature vector is matched to a previous feature vector and the object was previously identified as not an obstacle.

14. The collision avoidance system of claim 9, wherein when a previous feature vector is determined to be a match with a current feature vector, a relevance value for the previous feature vector is increased.

15. A method of avoiding a collision while operating a vehicle in reverse comprising:

detecting a plurality of objects proximate to a vehicle with at least one sensor, when the vehicle is placed in a reverse drive gear;

analyzing data from the sensors with a controller for a collision avoidance system to determine a current feature vector based upon relationships among data from the at least one sensor;

storing the current feature vector in non-volatile memory for the collision avoidance system;

comparing the current feature vector to previous feature vectors stored in the memory to determine if the current feature vector matches any of the previous feature vectors;

modifying the probability that an object is an obstacle based upon a match between the current feature vector and a previous feature vector; and

providing at least one vehicle collision avoidance action when the probability that the object is determined to be an obstacle exceeds a predetermined threshold.

16. The method of claim 15, further comprising determining a probability of collision with the obstacle, and providing at least one vehicle collision avoidance action when the probability that the object is determined to be an obstacle that requires avoidance exceeds a predetermined threshold and when the probability of collision exceeds a predetermined collision threshold.

17. The method of claim 15, further comprising decreasing the probability that the object is an obstacle when the current feature vector is matched to a previous feature vector and the object was previously identified as not an obstacle.

18. The method of claim 15, further comprising increasing a relevance value for a previous feature vector every time that previous feature vector is determined to be a match to the current feature vector.