公开(公告)号：US20230004863A1

公开(公告)日：2023-01-05

申请号：US17682225

申请日：2022-02-28

Applicant: KABUSHIKI KAISHA TOSHIBA

Inventor： Yasuhiro KANISHIMA ,  Takashi SUDO ,  Hiroyuki YANAGIHASHI

IPC: G06N20/00 ,  G06N7/00 ,  G06K9/62

Abstract: According to one embodiment, a learning apparatus includes a processor. The processor acquires data with a label indicating whether the data is normal data or anomalous data. The processor calculates an anomaly degree indicating a degree to which the data is the anomalous data using an output of a model for the data. The processor calculates a loss value related to the anomaly degree using a loss function based on an adjustment parameter based on a previously calculated loss value and the label. The processor updates a parameter of the model so as to minimize the loss value.

公开(公告)号：US20230296475A1

公开(公告)日：2023-09-21

申请号：US17942275

申请日：2022-09-12

Applicant: KABUSHIKI KAISHA TOSHIBA

Inventor： Takashi SUDO ,  Yasuhiro KANISHIMA ,  Hiroyuki YANAGIHASHI

IPC: G01M13/00 ,  G01P15/00 ,  G01P3/00

CPC classification number: G01M13/00 ,  G01P15/00 ,  G01P3/00

Abstract: According to one embodiment, a condition monitoring apparatus includes a processing circuitry. The processing circuitry is configured to collect a sensor signal output from a sensor that monitors a condition of a mechanical device that is at least partially mobile. The processing circuitry is configured to diagnose a presence or absence of an anomaly in the mechanical device based on the sensor signal. The processing circuitry is configured to cut out the sensor signal in a time width according to any one or more of a speed, an acceleration, and a jerk of the mechanical device. The processing circuitry is configured to determine the presence or absence of an anomaly based on the cut out sensor signal.

公开(公告)号：US20240262653A1

公开(公告)日：2024-08-08

申请号：US18454241

申请日：2023-08-23

Applicant: KABUSHIKI KAISHA TOSHIBA

Inventor： Hiroyuki YANAGIHASHI ,  Takashi SUDO ,  Kouta NAKATA ,  Tsukasa IKE

IPC: B66B5/00 ,  B66B1/34 ,  B66B3/00 ,  H04R1/02 ,  H04R1/40 ,  H04R3/00

CPC classification number: B66B5/0018 ,  B66B1/3492 ,  B66B3/002 ,  H04R1/028 ,  H04R1/406 ,  H04R3/005

Abstract: According to one embodiment, an abnormal noise detection device includes a first processing circuit. The first processing circuit is configured to collect an operating noise of a moving body by a first microphone installed in the moving body; detect an abnormal noise in the operating noise; input position information of the moving body; and store history information in which a detection result of detecting the abnormal noise and the position information are associated with each other in a first memory. The first processing circuit is configured to determine presence or absence of abnormality of the moving body by integrating the detection result for each piece of the position information on a basis of the history information.

公开(公告)号：US20210065918A1

公开(公告)日：2021-03-04

申请号：US16809271

申请日：2020-03-04

Applicant: KABUSHIKI KAISHA TOSHIBA

Inventor： Yasuhiro KANISHIMA ,  Hiroyuki YANAGIHASHI ,  Takashi SUDO ,  Kazunori IMOTO

IPC: G16Y40/10 ,  G16Y20/20 ,  G06K9/00 ,  G06N20/00

Abstract: According to one embodiment, a condition monitoring device includes a processor. The processor is configured to acquire a time-series signal about a condition of a monitor target from a first sensor, acquire operation timing information indicating start of operation of the monitor target, detect a first operation segment signal from the time-series signal based on the operation timing information, detect a second operation segment signal from the first operation segment signal based on a waveform feature of the first operation segment signal, and determine the condition of the monitor target based on the second operation segment signal.

公开(公告)号：US20210010909A1

公开(公告)日：2021-01-14

申请号：US16807611

申请日：2020-03-03

Applicant: KABUSHIKI KAISHA TOSHIBA

Inventor： Hiroyuki YANAGIHASHI ,  Takashi SUDO ,  Kazunori IMOTO ,  Yasuhiro KANISHIMA

IPC: G01M99/00 ,  G10L25/51 ,  G06N3/08 ,  G06F17/11 ,  G01H17/00

Abstract: According to one embodiment, a learning apparatus includes a memory and a hardware processor connected to the memory which learns a transformation function to extract a feature value of an input signal. The hardware processor updates the transformation function based on a signal indicative of a first condition and a signal indicative of a second condition which is different from the first condition, using a first loss function on the signal indicative of the first condition and a second loss function on the signal indicative of the second condition. The second loss function is designed such that the second condition becomes distant from the first condition.