Sensor misalignment correction

Patent metadata

Field	Value
Publication number	US-12235385-B2
Application number	US-202217674921-A
Country	US
Kind code	B2
Filing date	Feb 18, 2022
Priority date	Feb 18, 2022
Publication date	Feb 25, 2025
Grant date	Feb 25, 2025

Abstract

A computer includes a processor and a memory storing instructions executable by the processor to predict a quantity of misalignment of an optical sensor based on a projected motion of a vehicle, predict an error of the predicted quantity of misalignment, and actuate the vehicle based on the predicted quantity of misalignment and the predicted error. The vehicle includes the optical sensor.

First claim

The invention claimed is:

1. A computer comprising a processor and a memory, the memory storing instructions executable by the processor to:
execute an algorithm to predict a numerical quantity of misalignment of an optical sensor, a future projected motion of a vehicle being an input to the algorithm, the vehicle including the optical sensor;
predict an error of the predicted numerical quantity of misalignment, the error representing an uncertainty of the predicted numerical quantity of misalignment; and
actuate at least one of a propulsion, a brake system, or a steering system of the vehicle using the predicted numerical quantity of misalignment and the predicted error.

2. The computer of claim 1 , wherein an upcoming profile of a road on which the vehicle is traveling is an input to the algorithm for the predicted numerical quantity of misalignment.

3. The computer of claim 1 , wherein an upcoming profile of a road on which the vehicle is traveling is an input to predicting the error.

4. The computer of claim 1 , wherein the future projected motion of the vehicle is an input to predicting the error.

5. The computer of claim 1 , wherein predicting the error is performed without the predicted quantity of misalignment.

6. The computer of claim 1 , wherein at least one weather condition is an input to the algorithm for the predicted numerical quantity of misalignment.

7. The computer of claim 1 , wherein at least one weather condition is an input to predicting the error.

8. The computer of claim 1 , wherein a motion state of the vehicle is an input to the algorithm for the predicted numerical quantity of misalignment, and the motion state of the vehicle is an input to predicting the error.

9. The computer of claim 1 , wherein the instructions further include instructions to, in response to the predicted error being below a threshold, adjust data from the optical sensor using the predicted quantity of misalignment.

10. The computer of claim 9 , wherein actuating the at least one of the propulsion, the brake system, or the steering system of the vehicle includes actuating the at least one of the propulsion, the brake system, or the steering system of the vehicle using the adjusted data from the optical sensor.

11. The computer of claim 1 , wherein the instructions further include instructions to, in response to the predicted error exceeding a threshold, modify a plan for motion of the vehicle.

12. The computer of claim 11 , wherein actuating the vehicle includes actuating the vehicle using the modified plan for the motion of the vehicle.

13. The computer of claim 1 , wherein the algorithm for the predicted numerical quantity of misalignment is a machine-learning algorithm.

14. The computer of claim 13 , wherein the instructions further include instructions to determine the quantity of misalignment of the optical sensor, and train the machine-learning algorithm using the determined quantity of misalignment of the optical sensor.

15. The computer of claim 13 , wherein the machine-learning algorithm is a first machine-learning algorithm, and predicting the error includes executing a second machine-learning algorithm distinct from the first machine-learning algorithm.

16. The computer of claim 15 , wherein the first machine-learning algorithm is a different model than the second machine-learning algorithm.

17. The computer of claim 1 , wherein predicting the error includes executing a machine-learning algorithm.

18. The computer of claim 17 , wherein the instructions further include instructions to determine the quantity of misalignment of the optical sensor, and train the machine-learning algorithm using the determined quantity of misalignment of the optical sensor.

19. The computer of claim 18 , wherein training the machine-learning algorithm further uses the predicted quantity of misalignment of the optical sensor.

20. A method comprising:
executing an algorithm to predict a numerical quantity of misalignment of an optical sensor, a future projected motion of a vehicle being an input to the algorithm, the vehicle including the optical sensor;
predicting an error of the predicted numerical quantity of misalignment, the error representing an uncertainty of the predicted numerical quantity of misalignment; and
actuating at least one of a propulsion, a brake system, or a steering system of the vehicle using the predicted numerical quantity of misalignment and the predicted error.

Assignees

• Ford Global Tech Llc

Inventors

• Herman david michael
• Amodeo Catherine Marie
• Shanku Alexander George

Classifications

• G01S13/931

  of land vehicles · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/047

  Probabilistic or stochastic networks · CPC title