Detecting and resolving information contradictions for a vehicle

Patent metadata

Field	Value
Publication number	US-12304512-B2
Application number	US-202218057461-A
Country	US
Kind code	B2
Filing date	Nov 21, 2022
Priority date	Nov 21, 2022
Publication date	May 20, 2025
Grant date	May 20, 2025

Abstract

A system for resolving information contradictions in information received by a vehicle includes a plurality of vehicle sensors for receiving the information and a controller in electrical communication with the plurality of vehicle sensors. The controller is programmed to receive a first piece of information from a first of the plurality of vehicle sensors and receive a second piece of information from a second of the plurality of vehicle sensors. The controller is further programmed to determine a contradiction status between the first piece of information and the second piece of information. The controller is further programmed to resolve a contradiction between the first piece of information and the second piece of information using a third piece of information from a third of the plurality of vehicle sensors in response to determining that the contradiction status is a contradicting status.

First claim

What is claimed is:

1. A system for resolving information contradictions in information received by a vehicle, the system comprising:
a plurality of vehicle sensors for receiving the information, wherein the plurality of vehicle sensors includes at least a camera;
a controller in electrical communication with the plurality of vehicle sensors, wherein the controller is programmed to:
determine a first piece of information using a first of the plurality of vehicle sensors;
determine a second piece of information using a second of the plurality of vehicle sensors;
determine a contradiction status between the first piece of information and the second piece of information, wherein the contradiction status includes a contradicting status and a non-contradicting status;
construct if-then clauses from one of the first piece of information and the second piece of information in response to determining that the contradiction status is the contradicting status;
determine a third piece of information using a third of the plurality of vehicle sensors in response to determining that the contradiction status is the contradicting status, wherein the third of the plurality of vehicle sensors is the camera, and wherein the third piece of information includes an image of an environment surrounding the vehicle;
resolve a contradiction between the first piece of information and the second piece of information to determine a true piece of information by evaluating the if-then clauses using a visual question answering algorithm based on the third piece of information in response to determining that the contradiction status is the contradicting status, wherein the true piece of information is one of: the first piece of information and the second piece of information; and
adjust an operation of the vehicle based on the true piece of information in response to resolving the contradiction.

2. The system of claim 1 , wherein to determine the contradiction status, the controller is further programmed to:
determine the contradiction status between the first piece of information and the second piece of information using at least a contradiction detection algorithm.

3. The system of claim 2 , wherein the contradiction detection algorithm is a transformer model machine learning algorithm.

4. The system of claim 3 , wherein the transformer model machine learning algorithm is configured to receive the first piece of information and the second piece of information as inputs and provide a contradiction probability as an output.

5. The system of claim 1 , wherein the plurality of vehicle sensors includes at least: a microphone for receiving voice commands from an occupant of the vehicle and a vehicle communication system for communicating with systems external to the vehicle.

6. The system of claim 5 , wherein the first, second, and third pieces of information are at least one of: a voice command from the occupant of the vehicle received using the microphone, an image of the environment surrounding the vehicle captured using the camera, and information retrieved from an external system using the vehicle communication system.

7. The system of claim 5 , wherein to resolve the contradiction between the first piece of information and the second piece of information, the controller is further programmed to:
retrieve the third piece of information from the third of the plurality of vehicle sensors, wherein the third piece of information includes at least one of: the image of the environment surrounding the vehicle captured using the camera, information about a remote vehicle in the environment surrounding the vehicle retrieved using the vehicle communication system, and information retrieved from an external system using the vehicle communication system;
determine a truth score for each of the first and second pieces of information based on evaluating the if-then clauses using the third piece of information; and
resolve the contradiction based on the truth score for each of the first and second pieces of information.

8. The system of claim 7 , wherein to evaluate the if-then clauses, the controller is further programmed to:
evaluate the if-then clauses using the information about the remote vehicle, wherein the information about the remote vehicle includes at least one of: a speed and a heading of the remote vehicle.

9. The system of claim 7 , wherein to evaluate the if-then clauses, the controller is further programmed to:
evaluate the if-then clauses using the information retrieved from the external system, wherein the information retrieved from the external system includes at least one of: map data, driving law data, and vehicle operator manual data.

10. The system of claim 7 , wherein to resolve the contradiction based on the truth score for each of the first and second pieces of information, the controller is further programmed to:
compare the truth score of the first piece of information to the truth score of the second piece of information;
determine the first piece of information to be the true piece of information in response to determining that the truth score of the first piece of information is greater than the truth score of the second piece of information and that a difference between the truth score of the first piece of information and the truth score of the second piece of information is greater than or equal to a predetermined truth score difference threshold;
determine the second piece of information to be the true piece of information in response to determining that the truth score of the second piece of information is greater than the truth score of the first piece of information and that the difference between the truth score of the first piece of information and the truth score of the second piece of information is greater than or equal to the predetermined truth score difference threshold; and
prompt the occupant of the vehicle with the first piece of information and the second piece of information in response to determining that the difference between the truth score of the first piece of information and the truth score of the second piece of information is less than the predetermined truth score difference threshold.

11. A method for resolving information contradictions for a vehicle, the method comprising:
determining a first piece of information using a first of a plurality of vehicle sensors;
determining a second piece of information using a second of the plurality of vehicle sensors;
determining a contradiction status between the first piece of information and the second piece of information, wherein the contradiction status includes a contradicting status and a non-contradicting status;
constructing if-then clauses from one of the first piece of information and the second piece of information in response to determining that the contradiction status is the contradicting status;
determining a third piece of information from a third of the plurality of vehicle sensors in response to determining that the contradiction status is the contradicting status, wherein the third of the plurality of vehicle sensors is a camera, and wherein the third piece of information includes an image of an environment surrounding the vehicle; and
resolving a contradiction between the first piece of information and the second piece of information to determine a true piece of information by evaluating the if-then clauses using a visual question answering algorithm based on the third piece of information in response to determining that the contradiction status is the contradicting status, wherein the true piece of information is one of: the first piece of information and the second piece of information; and
adjust

Assignees

• Gm Global Tech Operations Llc

Inventors

• Esna Ashari Esfahani Alireza
• Mudalige Upali P

Classifications

• B60W50/0205

  Diagnosing or detecting failures; Failure detection models · CPC title

• G05B13/0265

  the criterion being a learning criterion · CPC title

• B60W2050/0215

  Sensor drifts or sensor failures · CPC title

• G01S13/931

  of land vehicles · CPC title

• G01S13/867

  Combination of radar systems with cameras · CPC title