# US 2019/0317848 A1 · Sensor data anomaly detection

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

- Number: US 2019/0317848 A1
- Document ID: US20190317848A1
- Status: Published application
- Kind: A1
- Application number: 16/455,186
- Filed: 2019-06-27
- Published: 2019-10-17
- Earliest claimed: 2017-02-15
- Assignees: Sap SE
- Applicants: Sap SE
- Inventors: Robert Meusel; Jaakob Kind; Atreju Florian Tauschinsky; Janick Frasch; Minji Lee; Michael Otto
- Web page: https://patentgrep.com/patent/US20190317848A1

## CPC

- G06F 11/079: {Root cause analysis, i.e. error or fault diagnosis (in a hardware test environment G06F11/22; in a software test environment G06F11/36)}
- G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks}
- G06F 11/0721: {within a central processing unit [CPU]}
- G06F 16/2477: {Temporal data queries}
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Some embodiments include reception of a time-series of a respective data value generated by each of a plurality of sensors, calculation of a regression associated with a first sensor of the plurality of sensors based on the received plurality of time-series, the regression being a function of the respective data values of the others of the plurality of data sources, reception of respective data values associated with a time from and generated by each the plurality of respective sensors, determination of a predicted value associated with the time for the first sensor based on the regression associated with the first sensor and on the respective data values associated with the time, comparison of the predicted value with the received value associated with the time and generated by the first sensor, and determination of a value indicating a likelihood of an anomaly based on the comparison.

## Claims

### Claim at position 1

-20. (canceled)

### Claim 21

A system comprising:
a plurality of data sources, each of the plurality of data sources to generate a time-series of a respective data value;
a server device to:
receive the plurality of time-series of respective data values;
receive one or more configuration parameters;
calculate a regression associated with a first data source of the plurality of data sources based on the received plurality of time-series and the received one or more configuration parameters;
receive respective data values associated with a time from and generated by each of the plurality of respective data sources;
determine a predicted value associated with the time for the first data source based on the regression associated with the first data source, received one or more configuration parameters and on the respective data values associated with the time;
compare the predicted value with the received value associated with the time and generated by the first data source; and
determine a value indicating a likelihood of an anomaly based on the comparison.

### Claim 22 (depends on 21)

The system of claim 21, wherein the one or more configuration parameters are window size, target column, weight by uncertainty, normalizer type and normalizer value.

### Claim 23 (depends on 22)

The system of claim 22, wherein the window size parameter indicates the number of most-recent observations per time-series.

### Claim 24 (depends on 23)

The system of claim 23, wherein a first window size parameter is applied to a first sensor and a second window size parameter is applied to a second sensor.

### Claim 25 (depends on 24)

The system of claim 24, wherein:
the first window size parameter is different from the second window size parameter; and
the first sensor is different from the second sensor.

### Claim 26 (depends on 22)

The system of claim 22, wherein the target column parameter indicates at least one column for which the regression is to be calculated.

### Claim 27 (depends on 26)

The system of claim 26, wherein the at least one column represents a target set of data sources.

### Claim 28 (depends on 26)

The system of claim 26, wherein the at least one column is selected based on at least one of a relationship of the data sources, and a relevance of the data sources to detection of the anomaly.

### Claim 29 (depends on 22)

The system of claim 22, wherein the normalizer type parameter specifies a type of normalizer to apply to the value indicating the likelihood of the anomaly.

### Claim 30 (depends on 29)

The system of claim 29, wherein the normalizer value indicates a value which is used for the normalization.

### Claim 31 (depends on 30)

The system of claim 30, wherein the normalizer value is one quantile, threshold and none.

### Claim 32 (depends on 31)

The system of claim 31, wherein the quantile normalizer value causes an internal calculation of an effective threshold value based on training data.

### Claim 33

A computer-implemented method comprising:
receiving a time-series of a respective data value generated by each of a plurality of sensors;
receiving one or more configuration parameters;
calculating a regression associated with a first sensor of the plurality of sensors based on the received plurality of time-series and the received one or more configuration parameters;
receiving respective data values associated with a time from and generated by each the plurality of respective sensors;
determining a predicted value associated with the time for the first sensor based on the regression associated with the first sensor, received one or more configuration parameters and on the respective data values associated with the time;
comparing the predicted value with the received value associated with the time and generated by the first sensor; and
determining a value indicating a likelihood of an anomaly based on the comparison.

### Claim 34 (depends on 33)

The method of claim 33, wherein the one or more configuration parameters are window size, target column, weight by uncertainty, normalizer type and normalizer value.

### Claim 35 (depends on 34)

The method of claim 34, wherein the window size parameter indicates the number of most-recent observations per time-series.

### Claim 36 (depends on 35)

The method of claim 35, wherein a first window size parameter is applied to a first sensor and a second window size parameter is applied to a second sensor, and wherein the first window size parameter is different from the second window size parameter and the first sensor is different from the second sensor.

### Claim 37 (depends on 34)

The method of claim 34, wherein the target column parameter indicates at least one column for which the regression is to be calculated; and
wherein the at least one column represents a target set of data sources.

### Claim 38 (depends on 34)

The method of claim 34, wherein the normalizer type parameter specifies a type of normalizer to apply to the value indicating the likelihood of the anomaly; and
the normalizer value indicates a value which is used for the normalization.