Detecting interference in a wireless network

Patent metadata

Field	Value
Publication number	US-12170908-B2
Application number	US-201917430207-A
Country	US
Kind code	B2
Filing date	Feb 15, 2019
Priority date	Feb 15, 2019
Publication date	Dec 17, 2024
Grant date	Dec 17, 2024

Abstract

A method 300 is disclosed for generating and training a model to detect interference conditions at a cell in a wireless cellular network and to classify the impact of detected interference conditions on performance of the wireless cellular network in the cell. The method comprises, for each of a plurality of cells in the wireless cellular network ( 360 ), obtaining data representing received signal power at a base station serving the cell over a period of time ( 310 ) and obtaining data representing a plurality of performance metrics for the cell over the time period ( 330 ). The method further comprises obtaining classifications of the received signal power data into one of a plurality of cell interference conditions ( 320 ) and the performance metric data into one of a plurality of cell impact classes ( 340 ). The method further comprises applying a Multi-Task Learning Machine Learning algorithm to a training data set comprising the classified received signal power and performance metric data to generate a model for classifying received signal power data into one of the plurality of cell interference conditions and for classifying performance metric data into one of the plurality of cell impact classes ( 350 ).

First claim

The invention claimed is:

1. A method for generating and training a model to detect interference conditions at a cell in a wireless cellular network and to classify the impact of detected interference conditions on performance of the wireless cellular network in the cell, the method comprising:
for each of a plurality of cells in the wireless cellular network:
obtaining data representing received signal power at a base station serving the cell over a period of time;
obtaining a classification of the received signal power data into one of a plurality of cell interference conditions;
obtaining data representing a plurality of performance metrics for the cell over the time period; and
obtaining a classification of the performance metric data into one of a plurality of cell impact classes;

the method further comprising:
applying a Multi-Task Learning (MTL) Machine Learning (ML) algorithm to a training data set comprising the classified received signal power and performance metric data to generate a model for classifying received signal power data into one of the plurality of cell interference conditions and for classifying performance metric data into one of the plurality of cell impact classes, wherein applying an MTL ML algorithm to a training data set comprising the classified received signal power and performance metric data comprises:
learning in parallel a feature representation for each task from the set of tasks comprising classifying received signal power data into one of the plurality of cell interference conditions and classifying performance metric data into one of the plurality of cell impact classes;
combining the feature representations learned for each of the tasks; and
jointly learning a shared feature representation for both tasks and parameters for the shared model to perform both tasks.

2. The method of claim 1 , wherein obtaining data representing received signal power and obtaining data representing a plurality of performance metrics over a time period comprise obtaining time series data representing the received signal power and obtaining time series data representing the plurality of performance metrics; the method further comprising, for each of a plurality of cells in the wireless cellular network:
converting the received signal power time series data into a received signal power image data set; and
converting the performance metric time series data into a performance metric image data set.

3. The method of claim 2 , wherein obtaining a classification of the received signal power data into one of a plurality of cell interference conditions comprises obtaining a classification of individual images of the received signal power image data set into one of the plurality of cell interference conditions; and wherein obtaining a classification of the performance metric data into one of a plurality of cell impact classes comprises obtaining a classification of individual images of the performance metric image data set into one of the plurality of cell impact classes.

4. The method of claim 1 , wherein applying an MTL ML algorithm to a training data set comprising the classified received signal power and performance metric data comprises applying the MTL ML algorithm to generate a model for classifying received signal power data into one of the plurality of cell interference conditions as a primary task and for classifying performance metric data into one of the plurality of cell impact classes as an auxiliary task.

5. The method of claim 1 , wherein combining the feature representations learned for each of the tasks comprises:
concatenating the feature representations learned for each of the tasks.

6. The method of claim 1 , wherein jointly learning a shared feature representation for both tasks and parameters for the shared model to perform both tasks comprises optimizing a loss function for the tasks.

7. The method of claim 6 , wherein images in the received signal power image data set comprise a first dimension representative of time and a second dimension representative of uplink received power of individual Physical Resource Blocks (PRBs) and wherein each image pixel represents a value of received signal power for a particular PRB at a particular time instant.

8. The method of claim 7 , wherein images in the performance metric image data set comprise a first dimension representative of time and a second dimension representative of performance metric, and wherein each image pixel represents a value for a particular performance metric at a particular time instant.

9. The method of claim 8 , wherein converting the received signal power time series data and performance metric time series data into received signal power and performance metric image data sets comprises:
partitioning the time period into a plurality of time windows; and
converting time series data for each time window into an individual data image.

10. The method of claim 9 , wherein converting the performance metric time series data into an image data set comprises normalizing the performance metric time series data between maximum and minimum values.

11. The method of claim 9 , wherein converting the received signal power time series data into an image data set comprises transforming the data to a standardized 2 dimensional array.

12. The method of claim 11 , wherein obtaining a classification of the received signal power data into one of a plurality of cell interference conditions and obtaining a classification of the performance metric data into one of a plurality of cell impact classes comprise at least one of:
obtaining a classification performed by a human expert;
classifying the received signal power data into one of a plurality of cell interference conditions and the performance metric data into one of a plurality of cell impact classes using an ML model.

13. A computer program product comprising a non-transitory computer readable medium storing a computer program comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out a method according to claim 1 .

14. A method for detecting interference conditions at a cell in a wireless cellular network and classifying the impact of detected interference conditions on performance of the wireless cellular network in the cell, the method comprising:
obtaining data representing received signal power at a base station serving the cell of the over a period of time;
obtaining data representing a plurality of performance metrics for the cell over the time period; and
applying a Machine Learned model to the obtained data to classify the received signal power data into one of a plurality of cell interference conditions and to classify the performance metric data into one of a plurality of cell impact classes, wherein the Machine Learned model comprises a model trained using a method comprising:
for each of a plurality of cells in the wireless cellular network:
obtaining data representing received signal power at a base station serving the cell over a period of time;
obtaining a classification of the received signal power data into one of a plurality of cell interference conditions;
obtaining data representing a plurality of performance metrics for the cell over the time period; and
obtaining a classification of the performance metric data into one of a plurality of cell impact classes;

the training method further comprising:

applying a Multi-Task Learning (MTL) Machine Learning (ML) algorithm to a training data set comprising the classified received signal power and performance metric data to generate a model for classifyin

Assignees

• Ericsson Telefon Ab L M

Inventors

• Eng Chin Lam
• Frank Philipp
• Martin Cuerdo Raul
• Ho Mitchell
• Ng Chee Wai

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/045

  Combinations of networks · CPC title

• H04W24/10

  Scheduling measurement reports {; Arrangements for measurement reports} · CPC title