Anomaly detection system and method

Patent metadata

Field	Value
Publication number	US-2018038954-A1
Application number	US-201615552225-A
Country	US
Kind code	A1
Filing date	Feb 16, 2016
Priority date	Feb 20, 2015
Publication date	Feb 8, 2018
Grant date	—

Abstract

An acoustic array system for anomaly detection is provided. The acoustic array system ( 100 ) performs a scan (or a progressive scan of frequencies) of a given volume by transmitting one or more signals, and receives one or more reflected signals from objects within the volume. The reflected signals are then amplified and converted to a set of digital signals. Features of the set of digital signals are extracted both in time and frequency domains. The acoustic array system ( 100 ) further performs a comparison of these set of digital extracted features with the reflected signals via machine learning techniques. Based on the comparison, the acoustic array system detects one or more anomalies.

First claim

1 . A method comprising:
(i) insonifying at a first time interval, by using an acoustic array system, a volume at one or more predetermined frequencies from a first location by transmitting one or more signals; (ii) receiving one or more reflected signals from one or more objects in said first location based on said one or more transmitted signals; (iii) repeating the steps (i) till (ii) until a last location in said volume is insonified to obtain a first set of reflected signals; (iv) amplifying said first set of reflected signals to obtain a set of amplified signals; (v) converting said set of amplified signals to a set of digital signals; (vi) performing a comparison of said set of digital signals with said first set of reflected signals; and (vii) detecting, at said first time interval, one or more anomalies based on said comparison.

2 . The method of claim 1 , further comprising extracting one or more features from said set of digital signals in at least one of a time domain and a frequency domain, wherein said one or more extracted features comprises at least one of an amplitude or a duration per each insonified frequency, power spectral density and frequency.

3 . The method of claim 2 , wherein performing a comparison comprises analyzing for one or more variations in said one or more extracted features of said set of digital signals with respect to one or more features of at least one of said one or more transmitted signals and said first set of reflected signals, and wherein said one or more anomalies are detected based on said one or more variations.

4 . The method of claim 3 , further comprising generating an acoustic map specific to said volume based on said one or more extracted features of the set of digital signals and with respect to said first set of reflected signals.

5 . The method of claim 4 , further comprising training said acoustic array system based on said acoustic map generated, wherein when said acoustic array system is trained, said acoustic array system comprises training data.

6 . The method of claim 5 , further comprising determining, using said trained acoustic array system, number of frequency scans required for insonifying said volume with specific direction at a second time interval based on at least one of said training data, said one or more variations, addition of one or more objects, removal of said one or more objects, a change in location of said one or more objects in said volume, and said first set of reflected signals; and
detecting, using said trained acoustic array system, one or more anomalies at said second time interval in said volume, wherein said first set of reflected signals are a set of baseline signals that are used as a reference and compared with a second set of reflected signals obtained at said second time interval to detect said one or more anomalies in said volume during said second time interval.

7 . The method of claim 1 , wherein range of said one or more predetermined frequencies is greater than or equal to 20 kilo Hertz (kHz).

8 . An acoustic array system, comprising:
one or more transmitters that are configured to transmit, at a first time interval, one or more signals at one or more predetermined frequencies, for insonifying a volume from a first location; one or more receivers that are configured to receive one or more reflected signals from one or more objects in said first location based on said one or more transmitted signals, wherein said one or more signals are transmitted until a last location in said volume is insonified to obtain a first set of reflected signals; a microphone pre-amplification unit that is configured to amplify said first set of reflected signals to obtain a set of amplified signals; a multi-channel synchronous analog to digital converter (ADC) that is configured to convert said set of amplified signals to a set of digital signals; and a processor that is configured to perform a comparison of said set of digital signals with said first set of reflected signals, and detect, at said first time interval, one or more anomalies based on said comparison.

9 . The acoustic array system of claim 8 , wherein one or more features are extracted from said set of digital signals in at least one of a time domain and a frequency domain, wherein said one or more extracted features comprises at least one of an amplitude or a duration per each insonified frequency, power spectral density and frequency.

10 . The acoustic array system of claim 9 , wherein said processor is configured to perform a comparison by analyzing for one or more variations in said one or more features of said set of digital signals with respect to one or more features of at least one of said one or more transmitted signals and said set of reflected signals, and wherein said one or more anomalies are detected based on said one or more variations.

11 . The acoustic array system of claim 9 , wherein said processor is further configured to generate an acoustic map specific to said volume based on said one or more extracted features and said first set of reflected signals with respect to said one or more transmitted signals.

12 . The acoustic array system of claim 11 , wherein said processor is further configured to execute, one or more machine learning techniques stored in a memory, to train said acoustic array system in at least one of said time domain and said frequency domain based on said acoustic map generated, and wherein when said acoustic array system is trained, said acoustic array system comprises training data.

13 . The acoustic array system of claim 12 , wherein said trained acoustic array system is configured to determine number of frequency scans required for insonifying said volume with specific direction at a second time interval based on at least one of said training data, said one or more variations, addition of one or more objects, removal of said one or more objects, a change in location of said one or more objects in said volume, and said first set of reflected signals, and detect one or more anomalies at said second time interval in said volume, wherein said set of reflected signals are a set of baseline signals that are used as a reference and compared with another set of reflected signals obtained at said second time interval to detect said one or more anomalies.

14 . The acoustic array system of claim 8 , wherein range of said one or more predetermined frequencies is greater than or equal to 20 kilo Hertz (kHz).

15 . The acoustic array system of claim 8 , wherein said one or more anomalies are detected based on at least one of a position and a distance of said one or more receivers from said one or more objects.

16 . One or more non-transitory machine readable information storage mediums comprising one or more instructions which when executed by one or more acoustic array systems causes:
(i) insonifying at a first time interval, by using an acoustic array system, a volume at one or more predetermined frequencies from a first location by transmitting one or more signals; (ii) receiving one or more reflected signals from one or more objects in the first location based on the one or more transmitted signals; (iii) repeating the steps (i) till (ii) until a last location in the volume is insonified to obtain a first set of reflected signals; (iv) amplifying the first set of reflected signals to obtain a set of amplified signals; (v) converting the set of amplified signals to a set of digital signals; (vi) performing a comparison of the set of digital signals with the first set of reflected signals; and (vii) dete

Assignees

• Tata Consultancy Services Ltd

Inventors

• Deshpande Parijat
• Vempada Ramu
• Dasgupta Ranjan
• Pal Arpan
• Roy Dibyendu

Classifications

• G08B13/1672 Primary

  using sonic detecting means, e.g. a microphone operating in the audio frequency range · CPC title

• G08B29/186

  Fuzzy logic; neural networks · CPC title

• G08B13/1636

  using pulse-type detection circuits · CPC title

• G01S15/523 Primary

  for presence detection (burglar, theft or intruder alarms G08B13/00, e.g. G08B13/16) · CPC title

• G08B13/1681

  using infrasonic detecting means, e.g. a microphone operating below the audible frequency range · CPC title