# US 2022/0019788 A1 · Determination of noise in a signal

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

- Number: US 2022/0019788 A1
- Document ID: US20220019788A1
- Status: Published application
- Kind: A1
- Application number: 17/358,550
- Filed: 2021-06-25
- Published: 2022-01-20
- Earliest claimed: 2020-06-26
- Assignees: Orange
- Applicants: Orange
- Inventors: Esteban Selva; Apostolos Kountouris
- Web page: https://patentgrep.com/patent/US20220019788A1

## CPC

- G06F 18/24137: {Distances to cluster centroïds}
- H04B 17/345: Interference values ({signal-to-interference ratio [SIR] or carrier-to-interference ratio [CIR]} H04B17/336)
- G06F 18/23213: with fixed number of clusters, e.g. K-means clustering
- H04B 1/1027: {assessing signal quality or detecting noise/interference for the received signal}
- H04B 17/26: using historical data, averaging values or statistics
- H04L 1/20: using signal quality detector
- G06F 2218/04: Denoising
- G06F 2218/14: by matching peak patterns
- G06F 2218/22: Source localisation; Inverse modelling
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04B: TRANSMISSION
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A method for determining noise associated with a received signal in a telecommunications network, the received signal being sampled beforehand in the form of a succession of data. The method includes: selecting a portion of the data in the received signal and determining a given number of possible partitions of the selected portion of the data in the received signal, the number of partitions being greater than or equal to two; partitioning all of the data in the received signal into the number of partitions; estimating mean energies of the signal in each of the partitions and identifying a partition from among the partitions having a minimum mean, the identified partition being a noise partition; and estimating a variance of the noise partition, the noise associated with the received signal being a function of the variance of the noise partition.

## Claims

### Claim 1

A method comprising:
determining noise associated with a signal received by an equipment connected to a telecommunications network, said received signal being sampled beforehand in the form of a succession of data, the determining being implemented by a device and comprising:
selecting a portion of the data in the received signal and determining a number of possible partitions of the selected portion of the data in the received signal, the number of partitions being greater than or equal to two,
partitioning all of the data in the received signal into said number of partitions,
estimating mean energies of the signal in each of the partitions and identifying a partition from among said partitions having a minimum mean, the identified partition being a noise partition, and
estimating a variance of the noise partition, the noise associated with the received signal being a function of said variance of the noise partition.

### Claim 2 (depends on 1)

The method as claimed in claim 1, comprising:
refining the variance estimation by determining aberrant data from among the data in the received signal in the noise partition and weighting said aberrant data in the noise partition.

### Claim 3 (depends on 1)

The method as claimed in claim 1, wherein determining the number of partitions comprises:
varying the number of partitions between given values and, for each value of the number of partitions:
partitioning the portion of the data in the received signal according to the value of the number of partitions,
calculating a quality coefficient that measures the quality of the partitioning of the portion of the data in the received signal,
associating the value of the number of partitions with the quality coefficient; and
said method furthermore comprising selecting the value of an optimum number of partitions corresponding to the number of partitions associated with the highest quality coefficient.

### Claim 4 (depends on 1)

The method as claimed in claim 1, wherein the partitioning of the data in the received signal is performed using a k-means method comprising the preliminary generation of a center of gravity for each partition and iterative steps of:
for each center of gravity, assigning the data in the received signal that are closest to said center of gravity into the partition corresponding to said center of gravity, and
for each partition, updating the center of gravity of said partition on the basis of the mean energy of the signal in said partition.

### Claim 5 (depends on 3)

The method as claimed in claim 3, wherein the partitioning of the data in the received signal is performed using a k-means method comprising the preliminary generation of a center of gravity for each partition and iterative steps of:
for each center of gravity, assigning the data in the received signal that are closest to said center of gravity into the partition corresponding to said center of gravity, and
for each partition, updating the center of gravity of said partition on the basis of the mean energy of the signal in said partition, and wherein the quality coefficient is a silhouette coefficient.

### Claim 6 (depends on 4)

The method as claimed in claim 4, wherein the partitioning of the data in the received signal is terminated when a number of iterations of the iterative steps is reached or when the centers of gravity of the partitions are unchanged from one iteration to the next.

### Claim 7 (depends on 1)

The method as claimed in claim 1, comprising, beforehand, preprocessing of the received signal comprising:
estimating energies of the received signal for a plurality of sub-bands,
windowing said energies of the received signal on the basis of a window of given length, and
generating the data in the received signal by summing and weighting the windowed energies of the received signal.

### Claim 8 (depends on 1)

The method as claimed in claim 1, wherein the selection of a received signal data portion is random.

### Claim 9 (depends on 1)

The method as claimed in claim 1, wherein the variance of the noise partition is the mean energy of the noise partition.

### Claim 10 (depends on 1)

The method as claimed in claim 1, further comprising:
estimating a signal-to-noise ratio of the received signal, comprising:
estimating the noise on the basis of the determined variance, and
estimating the signal-to-noise ratio on the basis of the noise and of the received signal.

### Claim 11 (depends on 1)

The method as claimed in claim 1, further comprising:
detecting presence of the signal in the telecommunications network, comprising:
after estimating the variance of the noise associated with the signal:
estimating a detection threshold on the basis of the estimated variance,
receiving an energy for a given acquisition duration, and
determining the presence of the signal when the received energy is greater than the estimated detection threshold.

### Claim 12 (depends on 11)

The method as claimed in claim 11, further comprising:
estimating a spectral occupancy level of the sampled received signal, comprising:
for each frequency sub-band of a plurality of sub-bands of the received signal, determining a carrier of the signal by detecting presence of the signal in said frequency sub-band according to the method as claimed in claim 11, and
estimating the spectral occupancy level on the basis of the one or more determined carriers of the signal.

### Claim 13

A device for determining noise associated with a signal received by said device, comprising:
a processing circuit configured to:
determine noise associated with a signal received by an equipment connected to a telecommunications network, said received signal being sampled beforehand in the form of a succession of data, the determining being implemented by a device and comprising:
selecting a portion of the data in the received signal and determining a number of possible partitions of the selected portion of the data in the received signal, the number of partitions being greater than or equal to two,
partitioning all of the data in the received signal into said number of partitions,
estimating mean energies of the signal in each of the partitions and identifying a partition from among said partitions having a minimum mean, the identified partition being a noise partition, and
estimating a variance of the noise partition, the noise associated with the received signal being a function of said variance of the noise partition.

### Claim 14

A non-transitory computer program storage medium storing instructions for implementing a method of determining noise associated with a signal received by an equipment connected to a telecommunications network, when said instructions are executed by a processor of a processing circuit, said received signal being sampled beforehand in the form of a succession of data, wherein the instructions configure the processing circuit to:
select a portion of the data in the received signal and determining a number of possible partitions of the selected portion of the data in the received signal, the number of partitions being greater than or equal to two,
partition all of the data in the received signal into said number of partitions,
estimate mean energies of the signal in each of the partitions and identify a partition from among said partitions having a minimum mean, the identified partition being a noise partition, and
estimate a variance of the noise partition, the noise associated with the received signal being a function of said variance of the noise partition.
