Signal analysis systems and methods for features extraction and interpretation thereof

• US 10,664,698 B2
• Filed: 02/21/2018
• Issued: 05/26/2020
• Est. Priority Date: 11/03/2017
• Status: Active Grant

• Current Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Original Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Inventors: Banerjee, Snehasis, Chattopadhyay, Tanushyam, Mukherjee, Ayan
• Primary Examiner(s): Lau, Tung S
• Application Number:
  US15/900,987
• Publication Number:
  US 20190138806A1
• Time in Patent Office: 825 Days
• Field of Search: None
• CPC Class Codes: G06F 16/24568 Data stream processing; Continuous queries
• CPC Class Codes: G06F 18/213 Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
• CPC Class Codes: G06F 2218/04 Denoising
• CPC Class Codes: G06F 2218/08 Feature extraction
• CPC Class Codes: G06N 3/10 Interfaces, programming languages or software development kits, e.g. for simulating neural networks

Abstract

Development of sensor data based descriptive and prescriptive system involves machine learning tasks like classification and regression. Any such system development requires the involvement of different stake-holders for obtaining features. Such features typically obtained are not interpretable for 1-D sensor signals. Embodiments of the present disclosure provide systems and methods that perform signal analysis for features extraction and interpretation thereof wherein input is raw signal data where origin of a feature is traced to signal data, and mapped to domain/application knowledge. Feature(s) are extracted using deep learning network(s) and machine learning (ML) model(s) are implemented for sensor data analysis to perform causality analysis for prognostics. Layer(s) (say last layer) of Deep Network(s) contains the automatically derived features that can be used for ML tasks. Parameter(s) tuning is performed based on the set of features that were recommended by the system to determined performance of systems (or applications) under consideration.

Claims

1. A processor implemented method, comprising:

obtaining (202), via one or more hardware processors, an input data from one or more sensors;

pre-processing (204), via the one or more hardware processors, the input data to filter at least one of noise and one or more anomalies and obtain a filtered data;

extracting (206), by using a window based processing technique on the filtered data, a first set of features from the filtered data, wherein each feature from the first set of features is unique and corresponds to one or more domains;

applying (208) metadata to the first set of features;

generating (210), using a selection and recommendation system, a second set of features from the first set of features upon applying the metadata;

generating (212) one of more feature values for the second set of features based on one or more time instances observed in the window based processing technique;

performing (214), using one or more structured knowledge representation methods based search, a comparison of the one or more generated features values with one or more corresponding pre-defined domain values; and

mapping (216) the one or more generated feature values to one or more corresponding domain specific templates based on the comparison.

2. The processor implemented method of claim 1, wherein the step of extracting, by using a window based processing technique on the filtered data, a first set of features from the filtered data comprises:

analyzing a window size associated with each window applied using the window based processing technique; and

extracting, by using the analyzed window size, the first set of features associated with each window.

3. The processor implemented method of claim 2, wherein the window size is analysed based on a weighted aggregation of at least one of a domain specified window and a statistically identified window size.

4. The processor implemented method of claim 1, further comprising tuning the one or more feature values corresponding to each feature from the second set of features to determine performance of each feature in the one or more domains.

5. The processor implemented method of claim 1, further comprising upon applying the window based processing technique on the filtered data, identifying one or more duplicate features from each window;

and filtering the one or more duplicate features from the first set of features.

6. A system (100), comprising:

a memory (102) storing instructions;

one or more communication interfaces (106); and

one or more hardware processors (104) coupled to the memory (102) via the one or more communication interfaces (106), wherein the one or more hardware processors (104) are configured by the instructions to;

obtain an input data from one or more sensors;

pre-process the input data to filter at least one of noise and one or more anomalies and obtain a filtered data;

extract, by using a window based processing technique on the filtered data, a first set of features from the filtered data, wherein each feature from the first set of features is unique and corresponds to one or more domains;

apply metadata to the first set of features;

generate, using a selection and recommendation system, a second set of features from the first set of features upon applying the metadata;

generate one of more feature values for the second set of features based on one or more time instances observed in the window based processing technique;

perform, using one or more structured knowledge representation methods based search, a comparison of the one or more generated features values with one or more corresponding pre-defined domain values; and

map the one or more generated feature values to one or more corresponding domain specific templates based on the comparison.

7. The system of claim 6, wherein the first set of features are extracted from the filtered data by:

analyzing a window size associated with each window applied using the window based processing technique; and

extracting, by using the analyzed window size, the first set of features associated with each window.

8. The system of claim 7, wherein the window size is analysed based on a weighted aggregation of at least one of a domain specified window and a statistically identified window size.

9. The system of claim 6, wherein the one or more hardware processors are further configured to tune the one or more feature values corresponding to each feature from the second set of features to determine performance of each feature in the one or more domains.

10. The system of claim 6, wherein upon the window based processing technique being applied on the filtered data, the one or more hardware processors are further configured to identify one or more duplicate features from each window, and filter the one or more duplicate features from the first set of features.

11. One or more non-transitory machine readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors causes:

obtaining, via the one or more hardware processors, an input data from one or more sensors;

pre-processing, via the one or more hardware processors, the input data to filter at least one of noise and one or more anomalies and obtain a filtered data;

extracting, by using a window based processing technique on the filtered data, a first set of features from the filtered data, wherein each feature from the first set of features is unique and corresponds to one or more domains;

applying metadata to the first set of features;

generating, using a selection and recommendation system, a second set of features from the first set of features upon applying the metadata;

generating one of more feature values for the second set of features based on one or more time instances observed in the window based processing technique;

performing, using one or more structured knowledge representation methods based search, a comparison of the one or more generated features values with one or more corresponding pre-defined domain values; and

mapping the one or more generated feature values to one or more corresponding domain specific templates based on the comparison.

12. The one or more non-transitory machine readable information storage mediums of claim 11, wherein the step of extracting, by using a window based processing technique on the filtered data, a first set of features from the filtered data comprises:

analyzing a window size associated with each window applied using the window based processing technique; and

extracting, by using the analyzed window size, the first set of features associated with each window.

13. The one or more non-transitory machine readable information storage mediums of claim 12, wherein the window size is analysed based on a weighted aggregation of at least one of a domain specified window and a statistically identified window size.

14. The one or more non-transitory machine readable information storage mediums of claim 11, wherein the one or more instructions further cause tuning the one or more feature values corresponding to each feature from the second set of features to determine performance of each feature in the one or more domains.

15. The one or more non-transitory machine readable information storage mediums of claim 11, wherein the one or more instructions further cause

upon applying the window based processing technique on the filtered data, identifying one or more duplicate features from each window;

and

filtering the one or more duplicate features from the first set of features.