Noise data artificial intelligence apparatus and pre-conditioning method for identifying source of problematic noise

Patent metadata

Field	Value
Publication number	US-11494643-B2
Application number	US-201916687499-A
Country	US
Kind code	B2
Filing date	Nov 18, 2019
Priority date	Dec 13, 2018
Publication date	Nov 8, 2022
Grant date	Nov 8, 2022

Abstract

A noise data artificial intelligence learning method for identifying the source of problematic noise may include a noise data pre-conditioning method for identifying the source of problematic noise including: selecting a unit frame for the problematic noise among noises sampled with time; dividing the unit frame into N segments; analyzing frequency characteristic for each segment of the N segments and extracting a frequency component of each segment by applying Log Mel Filter; and outputting a feature parameter as one representative frame by averaging information on the N segments, wherein an artificial intelligence learning by the feature parameter extracted according to a change in time by the noise data pre-conditioning method applies Bidirectional RNN.

First claim

What is claimed is:

1. A noise data artificial intelligence learning method of identifying and diagnosing a source of problematic noise in a power train of a vehicle including a memory for storing noise data and vibration data and a processor configured to execute instructions stored in the memory, the method comprising:
a noise data pre-conditioning method for identifying the source of the problematic noise, including:
selecting, by the processor, a unit frame for the problematic noise among noises sampled with time;
dividing, by the processor, the unit frame into N segments;
analyzing, by the processor, frequency characteristic for each segment of the N segments and extracting a frequency component of each segment by applying Log Mel Filter; and
outputting, by the processor, a feature parameter as one representative frame by averaging information on the N segments,
wherein the output feature parameter is used as an input value for an artificial intelligence learning of an artificial intelligence learning machine,
wherein the artificial intelligence learning by the feature parameter extracted according to a change in time by the noise data pre-conditioning method applies a Bidirectional Recurrent Neural Network (RNN) including a Gated Recurrent Unit (GRU),
wherein the artificial intelligence learning additionally is configured to apply a Deep Neural Network (DNN) for improving a probabilistic accuracy of a Many to one diagnosis technique,
wherein the artificial intelligence learning is configured to additionally apply an Attention Mechanism for improving an information loss of past time data,
wherein the artificial intelligence learning is configured to additionally apply an Early stage ensemble algorithm for emphasizing importance of the past time data among problems of the RNN, and
wherein when a time axis of problematic noise learning data is constantly collected, an Ensemble model of jointly trained RNN algorithms using a time-frequency map and an engine RPM-frequency map for improving accuracy is additionally applied.

2. The noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein the sampling is configured to sample in a range of twice a problematic frequency band.

3. The noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein an overlap is set between the unit frame with time and a next unit frame.

4. A noise data artificial intelligence diagnostic apparatus implemented with the noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein a noise of a vehicle or a powertrain is directly measured by an input device of the apparatus, or stored noise data is provided through a storage medium.

5. A noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein an artificial intelligence logic of the artificial intelligence learning machine is configured to finally identify the source of the problematic noise by use of the feature parameter.

6. The noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein the Many to one diagnosis technique is used for diagnosing the source of the problematic noise for time flow.

7. The noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein the Attention Mechanism is used for establishing an accurate learning model for the problematic noise by additionally applying weights of an important portion of learning data.

8. The noise data artificial intelligence learning method of identifying the source of the problematic noise of claim 1 ,
wherein the Early stage ensemble algorithm is used for maintaining information by dividing weights for an initial time and a final time evenly in the learning data.

Assignees

• Hyundai Motor Co Ltd
• Kia Motors Corp
• Iucf Hyu
• Iucf Hyu Industry Univ Corporation Foundation Hanyang Univ

Inventors

• Lee Dong-Chul
• Jung In-Soo
• Chang Joon-Hyuk
• Noh Kyoung-Jin

Classifications

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06N3/045 Primary

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G10K11/178 Primary

  by electro-acoustically regenerating the original acoustic waves in anti-phase · CPC title

• G10L25/30

  using neural networks · CPC title