IP Library › Granted Patent US 12,460,994
Granted Patent B2
US 12,460,994 · App. 17/388,505 · Granted Nov 4, 2025

Abnormality detection device

Inventor: Atsushi Moribe (Kariya, JP)
Assignee: DENSO CORPORATION
G01M17/007G05B23/02G06F18/2133G06F18/214G06N3/044G06N3/045G06N3/08
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,460,994
App. No.
17/388,505
Filed
Jul 29, 2021
Granted
Nov 4, 2025
Kind
B2
Art Unit
3657
USPC
701/29.1
Abstract

An abnormality detection device, method, or a storage medium acquires learning target data and monitoring target data, generates a state observer by using a variable in an input variable configuration, generates a threshold, calculates an abnormality degree by combining a second state observation value and the monitoring target data and inputting a combined result to the competitive neural network, and calculates a determination result.

Claims (59)

1 . An abnormality detection device comprising:

a signal acquisition portion configured to acquire learning target data and monitoring target data, the learning target data is data used for machine learning and the monitoring target data is data of a monitoring target;

a state observer generation portion configured to generate a state observer by using a variable in an input variable configuration;

a normal model generation portion configured to generate a threshold by combining a first state observation value obtained by input of the learning target data to the state observer and the learning target data and inputting a combined result to a competitive neural network;

an abnormality degree calculation portion configured to calculate an abnormality degree by combining a second state observation value obtained by input of the monitoring target data to the state observer and the monitoring target data and inputting a combined result to the competitive neural network; and

a determination portion configured to calculate a determination result by comparing the threshold with the abnormality degree, wherein

the abnormality degree is a difference between the learning target data and the second state observation value, which are input to the competitive neural network, and neuron weight data of a winning unit of the competitive neural network.

2 . The abnormality detection device according to claim 1 , wherein:

the state observer is configured to reflect a function of a device to be detected by the abnormality detection device or a malfunction mechanism of the device.

3 . The abnormality detection device according to claim 1 , wherein:

the variable in the variable configuration is a function of a device to be detected by the abnormality detection device or a factor of a malfunction mechanism of the device.

4 . The abnormality detection device according to claim 1 , wherein:

the state observer generation portion is configured to generate the state observer by linearly combining the variable in the variable configuration and inputting the learning target data.

5 . The abnormality detection device according to claim 1 , further comprising:

an input portion configured to input the variable configuration.

6 . The abnormality detection device according to claim 1 , wherein:

the state observer includes a first state observer and a second state observer; and

a correlation between the first state observer and the second state observer is maximized.

7 . The abnormality detection device according to claim 1 , wherein:

the state observer includes a first state observer, a second state observer, and a third state observer; and

a sum of a first correlation between the first state observer and the second state observer, a second correlation between the second state observer and the third state observer, and a third correlation between the third state observer and the first state observer is maximized.

8 . The abnormality detection device according to claim 6 , wherein:

the first state observer is a driving system state observer for a driving system of an automobile;

the second state observer is an air-fuel ratio system state observer for a ratio of air and fuel supplied to an engine;

the variable configuration of the driving system state observer includes, as the variable, an engine rotation speed and a turbine rotation speed; and

the variable configuration of the air-fuel ratio system state observer includes, as the variable, an oxygen sensor voltage and an air-fuel ratio sensor current.

9 . The abnormality detection device according to claim 7 , wherein:

the first state observer is a driving system state observer for a driving system of an automobile;

the second state observer is an air-fuel ratio system state observer for a ratio of air and fuel supplied to an engine;

the variable configuration of the driving system state observer includes, as the variable, an engine rotation speed and a turbine rotation speed; and

the variable configuration of the air-fuel ratio system state observer includes, as the variable, an oxygen sensor voltage and an air-fuel ratio sensor current.

10 . The abnormality detection device according to claim 1 , further comprising:

a factor analysis portion configured to specify a cause of an abnormality by using the second state observation value and the monitoring target data that cause determination indicating abnormality when the determination result indicates the abnormality.

11 . The abnormality detection device according to claim 10 , wherein:

the factor analysis portion is configured to specify the cause of the abnormality by further using the second state observation value and the monitoring target data that are earlier or later in time than the second state observation value and the monitoring target data that cause the determination of the abnormality.

12 . The abnormality detection device according to claim 10 , further comprising:

a display portion configured to display information for specifying the cause of the abnormality.

13 . An abnormality detection method comprising:

acquiring learning target data, the learning target data is data used for machine learning;

generating a state observer by using a variable in an input variable configuration;

generating a threshold by combining a first state observation value obtained by input of the learning target data to the state observer and the learning target data and inputting a combined result to a competitive neural network;

acquiring monitoring target data, the monitoring target data is data of a monitoring target;

calculating an abnormality degree by combining a second state observation value obtained by input of the monitoring target data to the state observer and the monitoring target data and inputting a combined result to the competitive neural network; and

calculating a determination result by comparing the threshold with the abnormality degree, wherein

the abnormality degree is a difference between the learning target data and the second state observation value, which are input to the competitive neural network, and neuron weight data of a winning unit of the competitive neural network.

14 . A computer-readable non-transitory storage medium storing an abnormality detection program configured to cause a computer to:

acquire learning target data, the learning target data is data used for machine learning;

generate a state observer by using a variable in an input variable configuration;

generate a threshold by combining a first state observation value obtained by input of the learning target data to the state observer and the learning target data and inputting a combined result to a competitive neural network;

acquire monitoring target data, the monitoring target data is data of a monitoring target;

calculate an abnormality degree by combining a second state observation value obtained by input of the monitoring target data to the state observer and the monitoring target data and inputting a combined result to the competitive neural network; and

calculate a determination result by comparing the threshold with the abnormality degree, wherein

the abnormality degree is a difference between the learning target data and the second state observation value, which are input to the competitive neural network, and neuron weight data of a winning unit of the competitive neural network.

15 . The abnormality detection device according to claim 1 , wherein:

the signal acquisition portion, the state observer generation portion, the normal model generation portion, the abnormality degree calculation portion, and the determination portion correspond to a processor coupled to a memory.

16 . The abnormality detection device according to claim 5 , wherein:

the input portion corresponds to the processor.

17 . The abnormality detection device according to claim 9 , wherein:

the factor analysis portion is configured to specify the cause of the abnormality by further using the second state observation value and the monitoring target data that are earlier or later in time than the second state observation value and the monitoring target data that cause the determination of the abnormality.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 29, 2021
From: MORIBE, ATSUSHI
To: DENSO CORPORATION
Reel/Frame 057021/0746 →
Priority Claims (1)
JP 2019-021084 · Feb 7, 2019 · national
Continuity (2)
Continuation PCTJP2019050704 · Dec 24, 2019
Related Publication 20210357727A1 · Nov 18, 2021
References Cited (27)
US 10496515B2 · Maya · 2019 [cited by examiner]
US 10605188B2 · Matsumoto · 2020 [cited by examiner]
US 10816951B2 · Abe · 2020 [cited by examiner]
US 20080301077A1 · Fung · 2008 [cited by examiner]
US 20120277949A1 · Ghimire et al. · 2012 [cited by applicant]
US 20160310046A1 · Heinrich · 2016 [cited by examiner]
US 20170287125A1 · Tanaka · 2017 [cited by examiner]
US 20180172624A1 · Aoki · 2018 [cited by examiner]
US 20180231493A1 · Aoki · 2018 [cited by examiner]
US 20180276569A1 · Oguri · 2018 [cited by applicant]
US 20190301979A1 · Kawanoue · 2019 [cited by examiner]
US 20190308640A1 · Miller · 2019 [cited by examiner]
US 20200019852A1 · Yoon et al. · 2020 [cited by applicant]
US 20200039454A1 · Hsieh · 2020 [cited by examiner]
US 20200150159A1 · Takeda · 2020 [cited by examiner]
US 20200162642A1 · Oniki · 2020 [cited by examiner]
US 20200334554A1 · Takahashi · 2020 [cited by examiner]
US 20200376667A1 · Götvall · 2020 [cited by examiner]
US 20230128023A1 · Moribe · 2023 [cited by applicant]
JP 2007198918A · 2007 [cited by applicant]
JP 4832609B1 · 2011 [cited by applicant]
JP 2014518580A · 2014 [cited by applicant]
JP 2015018389A · 2015 [cited by applicant]
JP 2017102826A · 2017 [cited by applicant]
JP 2017192277A · 2017 [cited by applicant]
KR 101940029B1 · 2019 [cited by applicant]
WO WO2018230645A1 · 2018 [cited by applicant]