AI-based noise generation diagnosis apparatus and method of PE components for electric vehicle

Patent metadata

Field	Value
Publication number	US-12112591-B2
Application number	US-202217869478-A
Country	US
Kind code	B2
Filing date	Jul 20, 2022
Priority date	Feb 11, 2022
Publication date	Oct 8, 2024
Grant date	Oct 8, 2024

Abstract

Embodiments of the present disclosure relate to an AI-based noise generation diagnosis apparatus and method of diagnosing noise of PE components for an electric vehicle, in which a plurality of microphones are disposed in a vehicle interior, and a calculation unit measures positions of noise sources from any one or more among the PE components comprising a main motor, a decelerator, an inverter, and an auxiliary motor from front, rear, and upward directions of the microphone, and at the same time, estimates a noise transmission position where the noise is transmitted.

First claim

What is claimed is:

1. An artificial intelligence (AI) based apparatus for diagnosing noise generation of power electric (PE) components for an electric vehicle, comprising:
a plurality of microphones disposed in a vehicle interior; and
a calculation unit collecting and calculating noise data measured by the microphones,
wherein the microphones acquire the noise data generated from any one or more of the PE components, and
wherein the calculation unit performs artificial intelligence learning for diagnosing failures and estimating a noise generation position and a noise transmission position to identify a position of the noise source and a noise leak portion.

2. The apparatus of claim 1 , wherein the PE components comprise a main motor, a decelerator, and an inverter positioned in a rear trunk, an auxiliary motor positioned in a front motor room.

3. The apparatus of claim 1 , wherein the same number of microphones is positioned in front and rear directions to estimate the position of the noise generation position and the noise transmission position.

4. The apparatus of claim 3 , wherein two or more microphones are additionally positioned in an upward direction to calculate a time difference at which the noises transmitted into the vehicle from the noise sources are delivered.

5. The apparatus of claim 3 , wherein a camera is positioned in the front and rear directions together with the microphones to visualize front and rear noise sources by the microphones.

6. The AI-based noise generation diagnosis apparatus of claim 3 ,
wherein the microphones are additionally positioned in an upward direction, and the camera is also additionally positioned in the upward direction.

7. An artificial intelligence (AI) based method of diagnosing noise generation of power electric (PE) components for an electric vehicle, comprising:
collecting noise data in front and rear directions from a plurality of microphones disposed in a vehicle interior;
extracting a feature vector of the noise data and preparing a data set labeled through data labeling;
performing, by a calculation unit, learning of a failure diagnosis model from the extracted feature vector and the labeled data set;
performing, by the calculation unit, learning estimating a noise generation position and a noise transmission position in addition to the failure diagnosis model; and
visualizing the noise generation position and the noise transmission position from image data acquired by a camera positioned in the front and rear directions together with the microphones.

8. The method of claim 7 , wherein the PE components comprise a main motor, a decelerator, and an inverter positioned in a rear trunk, an auxiliary motor positioned in a front motor room, or a combination thereof.

9. The method of claim 7 , wherein in the step of collecting of the noise data, the microphones are additionally positioned in an upward direction together with the microphones.

10. The method of claim 7 , wherein in the step of visualizing from the data, the camera is additionally positioned in an upward direction together with the microphones.

11. The method of claim 7 , wherein whine noise is separated by applying Cony-TASNet based on TCN and SepFormer model based on Transformer to the noise data measured by the microphones.

12. The method of claim 11 , wherein the step of extracting of the feature vector of the whine noise extracts and classifies an order component of the whine noise from any one or more noise sources among the PE component measured by the microphones.

13. The method of claim 7 , wherein the step of performing of the learning estimating the noise generation position and the noise transmission position performs learning about the noise data from any one or more of the PE components and the failure diagnosis, the noise generation position, and the noise transmission position from the normality/abnormality determination results for each noise data.

14. The method of claim 7 , wherein the step of estimating of the failure position extracts a failure position estimation feature vector as a model applying generalized cross correlation-phase transform (GCC-PHAT) by using a time difference calculation at which the noises collected by each of the microphones are delivered.

Assignees

• Hyundai Motor Co Ltd
• Kia Corp

Inventors

• Lee Dong-Chul
• Jung In-Soo

Classifications

• B60W2420/54

  Audio sensitive means, e.g. ultrasound · CPC title

• G01H3/125 Primary

  for representing acoustic field distribution (using optical means G01H9/002; sonar systems for imaging G01S7/56, G01S15/89; acoustic holography G03H3/00) · CPC title

• G10K11/26

  Sound-focusing or directing, e.g. scanning · CPC title

• G10K2210/3038

  Neural networks · CPC title

• G10K2210/1282

  Automobiles · CPC title