# US 2025/0094830 A1 · Frequency-domain signal clustering

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0094830 A1
- Document ID: US20250094830A1
- Status: Published application
- Kind: A1
- Application number: 18/370,101
- Filed: 2023-09-19
- Published: 2025-03-20
- Earliest claimed: 2023-09-19
- Assignees: Oracle International Corporation
- Applicants: Oracle International Corporation
- Inventors: Keyang RU; Ruixian LIU; Kuei-Da LIAO; Guang Chao WANG; Matthew T. GERDES; Kenny C. GROSS
- Web page: https://patentgrep.com/patent/US20250094830A1

## CPC

- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06F 18/23: Clustering techniques
- G06F 18/2433: Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Systems, methods, and other embodiments associated with clustering of time series signals based on frequency domain analysis are described. In one embodiment, an example method includes accessing time series signals to be separated into clusters. The example method also includes determining similarity in the frequency domain among the time series signals. The example method further includes extracting a cluster of similar time series signals from the time series signals based on the similarity in the frequency domain. And, the example method includes training a machine learning model to detect anomalies based on the cluster.

## Claims

### Claim 1

A computer-implemented method, comprising:
accessing time series signals to be separated into clusters;
determining similarity of power spectral density among the time series signals;
extracting a cluster of similar time series signals from the time series signals based on the similarity of power spectral density; and
training a machine learning model to detect anomalies based on the cluster.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, wherein determining similarity of power spectral density among the time series signals further comprises performing a comparison of the time series signals in a frequency domain.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein extracting a cluster of similar time series signals from the time series signals based on the similarity of power spectral density further comprises choosing those of the time series signals below a jump of dissimilarity to be the cluster of time series signals.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein determining similarity of power spectral density among the time series signals further comprises:
selecting one of the time series signals to be a reference signal;
generating cross power spectral densities between the reference signal and the other time series signals;
normalizing the cross power spectral densities to a periodogram of the reference signal; and
generating cumulative mean absolute errors of the normalized cross power spectral densities with respect to the periodogram of the reference signal, wherein the similarity of power spectral density is the cumulative mean absolute error of the cross power spectral density.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, wherein extracting a cluster of similar time series signals from the time series signals based on the similarity of power spectral density further comprises:
sorting the time series signals based on the similarity of power spectral density of a time series signal with other time series signals;
detecting a change point in the similarities for the time series signals; and
selecting the time series signals below the change point to be in the cluster.

### Claim 6 (depends on 5)

The computer-implemented method of claim 5, wherein the change point in the similarities for the time series signals is detected by applying a Mann-Kendall test to a mean cumulative function of the similarities in the sorted order of the time series signals.

### Claim 7 (depends on 5)

The computer-implemented method of claim 5, wherein the steps of determining similarity of power spectral density and extracting a cluster of similar time series signals are repeated until no further change point is detected at a given confidence level, the method further comprising placing the remaining time series signals that were not extracted into a final cluster once no further change point is detected.

### Claim 8 (depends on 1)

The computer-implemented method of claim 1, further comprising:
detecting an anomaly with the trained machine learning model; and
generating an electronic alert that the anomaly was detected in the cluster of the time series signals.

### Claim 9 (depends on 1)

The computer-implemented method of claim 1, wherein noise on one or more of the time series signals is in excess of 50%.

### Claim 10 (depends on 1)

The computer-implemented method of claim 1, wherein the machine learning model is a multivariate state estimation technique model, the method further comprising assigning the time series that are in the cluster to be inputs of the multivariate state estimation technique model, and not assigning the time series that are excluded from the cluster to be inputs of the multivariate state estimation technique model.

### Claim 11

One or more non-transitory computer-readable media that includes stored thereon computer-executable instructions that when executed by at least a processor of a computer system cause the computer system to:
access a collection of time series signals that are undifferentiated with respect to clusters;
determine similarity in a frequency domain among the time series signals;
chooses a cluster of similar time series signals from the time series signals based on the similarity in the frequency domain; and
route the cluster to a destination that is discrete from destinations for other clusters detected in the time series signals.

### Claim 12 (depends on 11)

The non-transitory computer-readable medium of claim 11, wherein the instructions for determining similarity of power spectral density among the time series signals further cause the computer system to perform a comparison of the time series signals in a frequency domain.

### Claim 13 (depends on 11)

The non-transitory computer-readable medium of claim 11, wherein the instructions for choosing a cluster of similar time series signals from the time series signals based on the similarity in the frequency domain further cause the computer system to:
apply a Mann-Kendall test to detect a change point in the time series signals, wherein the time series signals are in sorted in order of similarity; and
choose those of the time series signals below the change point to be the cluster of time series signals.

### Claim 14 (depends on 11)

The non-transitory computer-readable medium of claim 11, wherein the instructions further cause the computer system to:
repeat determining similarity in the frequency domain and extracting a cluster of similar time series signals until no further change point is detected at a given confidence level; and
place the remaining time series signals that were not extracted into a final cluster once no further change point is detected.

### Claim 15 (depends on 10)

The non-transitory computer-readable medium of claim 10, wherein the instructions to route the cluster to a destination that is discrete from destinations for other clusters further cause the computer system to:
train a machine learning model to detect anomalies in the cluster, wherein the machine learning model is specific to the cluster;
monitor a surveillance phase of the cluster with the trained machine learning model to detect an anomaly; and
in response to detecting the anomaly, generate an electronic alert that the anomaly has occurred in the cluster.

### Claim 16

A computer system, comprising:
at least one processor;
at least one memory connected to the at least one processor;
one or more non-transitory computer-readable media including instructions stored thereon that when executed by at least the processor cause the computing system to:
access a training range of time series signals that include unidentified clusters;
determine measures of correlation among the time series signals based on analysis in the frequency domain;
transfer time series signals into a cluster based on the measures of correlation;
train a machine learning model to detect anomalies based on the training range of the cluster; and
in response to detecting an anomaly in a surveillance range of the cluster, transmit an electronic alert that the anomaly has occurred in the cluster.

### Claim 17 (depends on 16)

The computer system of claim 16, wherein the instructions for transferring time series signals into a cluster based on the measures of correlation further cause the computer system to:
sort the time series signals in ascending order of the measures of correlation; and
transfer into the cluster those of the time series signals having a measure of correlation below a discontinuity in the measures of correlation for the sorted time series signals.

### Claim 18 (depends on 16)

The computer system of claim 16, wherein the measures of correlation among the time series signals are based on cumulative mean absolute errors of cross power spectral density between one of the time series signals that is selected as a reference and others of the time series signals.

### Claim 19 (depends on 16)

The computer system of claim 16, wherein the machine learning model is a multivariate anomaly detection model.

### Claim 20 (depends on 16)

The computer system of claim 16, further comprising:
an asset; and
one or more sensors that are configured to produce the time series signals as descriptions of physical states of the asset over time;
wherein the training range of the cluster describes a first physical state of the asset;
wherein the surveillance range of the cluster describes a second physical state of the asset; and
the anomaly in the surveillance range of the cluster indicates that a degradation of the asset is occurring.
