Ambient sound classification based message routing for local security and remote internet query systems

Patent metadata

Field	Value
Publication number	US-2019180735-A1
Application number	US-201715841045-A
Country	US
Kind code	A1
Filing date	Dec 13, 2017
Priority date	Dec 13, 2017
Publication date	Jun 13, 2019
Grant date	—

Abstract

Systems and methods to monitor for diverse audio sounds. A sound signal is received. The sound signal is processed to classify sounds represented by the sound signal as a determined sound type that is one sound type within a plurality of sound types. The plurality of sound types comprising at least spoken words and non-voice security event related sounds. A security event notification is sent to a security monitor based on classifying the sounds as non-voice security event related sounds, and an internet based query is sent to an internet query server based on classifying the sounds as spoken words.

First claim

What is claimed is:

1 . A method comprising:
receiving a sound signal; processing the sound signal to classify sounds represented by the sound signal as a determined sound type that is one sound type within a plurality of sound types, the plurality of sound types comprising at least spoken words and non-voice security event related sounds; sending a security event notification to a security monitor based on classifying the sounds as non-voice security event related sounds; and sending an internet based query to an internet query server based on classifying the sounds as spoken words.

2 . The method according to claim 1 , wherein the sounds represented by the sound signal are sounds received in an area, wherein the sounds received within the area comprise a sounding alarm, and wherein classifying the sounds as security event related sounds comprises processing the sound signal to determine that the sound signal comprises a sound of the sounding alarm within the sound signal.

3 . The method according to claim 1 , further comprising:
receiving, based on sending the internet based query to the internet query server, a response from the internet query server; and presenting the response.

4 . The method according to claim 1 ,
wherein the plurality of sound types further comprises distant non-voice security event related sounds, the distant non-voice security event related sounds comprising sounds that are associated with security events and that originate outside of an area proximate to a sound receiver receiving the sounds and producing the sounds signal, the method further comprising: sending a distant security event notification to the security monitor based on classifying the sounds as distant non-voice security event related sounds.

5 . The method according to claim 1 ,
wherein the plurality of sound types further comprises an audible alert signal generated by a safety sensor, wherein the sending the security event notification further comprises sending, based on classifying the sounds as the audible alert signal, the security event notification to a monitoring system that monitors a premises.

6 . The method according to claim 1 , wherein the sounds represented by the sound signal are sounds received in an area, wherein the sounds received within the area comprise spoken words, and wherein classifying the sounds as spoken words comprises processing the sound signal to determine that the sounds comprises a speaking voice, and
wherein the sending the internet based query comprises sending content of the sounds to the internet query server.

7 . The method of claim 6 , wherein sending content of the sounds comprising sending a digitized representation of the sounds to the internet query server.

8 . The method of claim 6 , further comprising performing, based on classifying the sounds as spoken words, speech recognition on the sounds to extract data expressed in the spoken words, and
wherein sending content of the sounds comprises sending the data expressed in the spoken words to the internet query server.

9 . The method according to claim 1 , wherein
the plurality of sound types further comprises a non-voice non-security event related sound, and the method further comprises sending a non-security event notification to a user device based on classifying the sounds as non-voice non-security event related sounds.

10 . The method according to claim 9 further comprising:
receiving, prior to receiving the sound signal, a first new sound signal representing a first new sound;
training a process processing the sound signal to classify the first new sound as the non-voice security event related sound within the plurality of sound types;
receiving, prior to receiving the sound signal, a second new sound signal representing a second new sound;
training the process processing the sound signal to classify the second new sound as the non-voice non-security event related sound within the plurality of sound types; and
receiving a user input indicative of the user device for sending the non-voice non-security event notification to the user device.

11 . A diverse ambient sound detector, comprising:
an audio processor that, when operating, receives a sound signal; and a sound classifier that, when operating:
processes the sound signal to classify sounds represented by the sound signal as a determined sound type that is one sound type within a plurality of sound types, the plurality of sound types comprising at least spoken words and non-voice security event related sounds;
sends a security event notification to a security monitor based on classifying the sounds as non-voice security event related sounds; and
sends an internet based query to an internet query server based on classifying the sounds as spoken words.

12 . The diverse ambient sound detector according to claim 11 , wherein the sounds represented by the sound signal are sounds received in an area, wherein the sounds received within the area comprise a sounding alarm, and
wherein the sound classifier classifies the sounds as security event related sounds by at least processing the sound signal to determine that the sound signal comprises a sound of the sounding alarm within the sound signal.

13 . The diverse ambient sound detector according to claim 11 , further comprising:
a response processor that, when operating:
receives, based on sending the internet based query to the internet query server, a response from the internet query server; and
presents the response.

14 . The diverse ambient sound detector according to claim 11 ,
wherein the plurality of sound types further comprises distant non-voice security event related sounds, the distant non-voice security event related sounds comprising sounds that are associated with security events and that originate outside of an area proximate to a sound detector receiving the sounds and producing the sounds signal, wherein the sound classifier, when operating, further sends a distant security event notification to the security monitor based on classifying the sounds as distant non-voice security event related sounds.

15 . The diverse ambient sound detector according to claim 11 wherein the sounds represented by the sound signal are sounds received in an area, wherein the sounds received within the area comprise spoken words, and
wherein the sound classifier:
classifies the sounds as spoken words by at least processing the sound signal to determine that the sounds comprises a speaking voice, and
sends the internet based query by at least sending content of the sounds to the internet query server.

16 . The diverse ambient sound detector of claim 15 , wherein the sound classifier, when operating:
further performs, based on classifying the sounds as spoken words, speech recognition on the sounds to extract data expressed in the spoken words, and sends content of the sounds by at least sending the data expressed in the spoken words to the internet query server.

17 . A security system comprising:
an installable diverse ambient sound detector coupled to a monitoring system, the installable diverse ambient sound detector comprising: an audio processor that, when operating, receives a sound signal; and a sound classifier that, when operating:
processes the sound signal to classify sounds represented by the sound signal as a determined sound type that is one sound type within a plurality of sound types, the plurality of sound types comprising at least spoken words and non-voice security event related sounds;

Assignees

• Florida Power & Light Co

Inventors

• Mcelhaney Patrick G

Classifications

• G10L15/08 Primary

  Speech classification or search · CPC title

• G08B1/08

  using electric transmission {; transformation of alarm signals to electrical signals from a different medium, e.g. transmission of an electric alarm signal upon detection of an audible alarm signal} · CPC title

• G10L25/54

  for retrieval · CPC title

• G10L15/02

  Feature extraction for speech recognition; Selection of recognition unit · CPC title

• H04R2420/07

  Applications of wireless loudspeakers or wireless microphones · CPC title