Detection of a hands-off condition through machine learning

Patent metadata

Field	Value
Publication number	US-11912283-B2
Application number	US-202016931351-A
Country	US
Kind code	B2
Filing date	Jul 16, 2020
Priority date	Jul 25, 2019
Publication date	Feb 27, 2024
Grant date	Feb 27, 2024

Abstract

Technologies and techniques for automatically generating labeled steering torque data, with which an artificial intelligence (AI) unit is trained to detect hands-off conditions when the vehicle is being operated.

First claim

The invention claimed is:

1. A process for training a learning algorithm for a driver assistance system for a vehicle, comprising:
recording, via first sensors, steering torque data for the vehicle;
recording, via second sensors, distance data associated with a driver's hands and a steering wheel;
detecting hands-off conditions for associated time intervals, based on at least the recorded distance data and one or more of (i) steering wheel movement data, (ii) data describing the course of the roadway, (iii) lane detection data, and/or (iv) vehicle movement data;
linking the detected hands-off conditions with the recorded steering torque data for the time intervals; and
compiling the linked hands-off conditions to generate labeled steering torque data, configured to be processed as training data for detecting future hands-off conditions in the learning algorithm for the vehicle.

2. The process of claim 1 , wherein compiling the linked hands-off conditions comprises compiling data for the driver assistance system via an adaptive algorithm.

3. The process of claim 1 , wherein the second sensors comprise capacitive distance sensors.

4. The process of claim 1 , further comprising processing the compiled data in the driver assistance system for detecting future hands-off conditions for the vehicle.

5. The process of claim 1 , wherein compiling the linked hands-off condition comprises storing the linked hand-off condition on a data storage medium and further comprising testing the stored data with an adaptive algorithm.

6. A system for training a learning algorithm for a driver assistance system for a vehicle, comprising:
first sensor for recording steering torque data for the vehicle;
second sensor for recording distance data associated with a driver's hands and a steering wheel;
a storage medium for storing recorded steering torque data and distance data; and
a control unit, operatively coupled to the storage medium, wherein the control unit and storage medium are configured to
detect hands-off conditions for associated time intervals, based on at least the recorded distance data and one or more of (i) steering wheel movement data, (ii) data describing the course of the roadway, (iii) lane detection data, and/or (iv) vehicle movement data;
link the detected hands-off conditions with the recorded steering torque data for the time intervals; and
compile the linked hands-off conditions and generating labeled steering torque data, configured to be processed as training data for detecting future hands-off conditions in the learning algorithm for the vehicle.

7. The system of claim 6 wherein the control unit and storage medium are configured to compile the linked hands-off conditions by compiling data for the driver assistance system via an adaptive algorithm.

8. The system of claim 6 , wherein the second sensor comprises a capacitive distance sensor.

9. The system of claim 6 , wherein the control unit and storage medium are configured to process the compiled data in the driver assistance system for detecting future hands-off conditions for the vehicle.

10. The system of claim 6 , wherein the control unit and storage medium are configured to test the stored data with an adaptive algorithm.

11. An article of manufacture for training a learning algorithm for a driver assistance system for a vehicle, the article comprising:
a non-transitory computer-readable medium having stored therein instructions executable by one or more processors to:
record, via first sensors, steering torque data for the vehicle;
record, via second sensors, distance data associated with a driver's hands and a steering wheel;
detect hands-off conditions for associated time intervals, based on at least the recorded distance data and one or more of (i) steering wheel movement data, (ii) data describing the course of the roadway, (iii) lane detection data, and/or (iv) vehicle movement data;
link the detected hands-off conditions with the recorded steering torque data for the time intervals; and
compile the linked hands-off conditions and generating labeled steering torque data, configured to be processed as training data for detecting future hands-off conditions in the learning algorithm for the vehicle.

12. The article of manufacture of claim 11 , wherein the non-transitory computer-readable medium compiles the linked hands-off conditions by compiling data for the driver assistance system via an adaptive algorithm.

13. The article of manufacture of claim 11 , wherein the second sensors comprise capacitive distance sensors.

14. The article of manufacture of claim 11 , wherein the non-transitory computer-readable medium processes the compiled data in the driver assistance system for detecting future hands-off conditions for the vehicle.

15. The article of manufacture of claim 11 , wherein the non-transitory computer-readable medium compiles the linked hands-off condition by storing the linked hand-off condition on a data storage medium and further comprising testing the stored data with an adaptive algorithm.

Assignees

• Volkswagen Ag

Inventors

• Eigel Thomas
• Stahl Felix

Classifications

• G06N3/09

  Supervised learning · CPC title

• B60W40/08 Primary

  related to drivers or passengers · CPC title

• B60W40/06

  Road conditions · CPC title

• B60W50/0097

  Predicting future conditions · CPC title

• G06V20/588

  Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road · CPC title