LEARNING ALGORITHM TO DETECT HUMAN PRESENCE IN INDOOR ENVIRONMENTS FROM ACOUSTIC SIGNALS

• US 20150380013A1
• Filed: 06/21/2015
• Published: 12/31/2015
• Est. Priority Date: 06/30/2014
• Status: Active Grant

• Current Assignee: Rajeev Conrad Nongpiur
• Original Assignee: Rajeev Conrad Nongpiur
• Inventors: Nongpiur, Rajeev Conrad
• Granted Patent: US 10,068,587 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G10L 17/02 Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction
• CPC Class Codes: G10L 17/26 Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices
• CPC Class Codes: G10L 25/15 the extracted parameters being formant information
• CPC Class Codes: G10L 25/51 for comparison or discrimination
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals

Abstract

A system is described that constantly learns the sound characteristics of an indoor environment to detect the presence or absence of humans within that environment. A detection model is constructed and a decision feedback approach is used to constantly learn and update the statistics of the detection features and sound events that are unique to the environment in question. The learning process may not only rely on acoustic signal, but may also make use of signals derived from other sensors such as range sensor, motion sensors, pressure sensors, and video sensors.

Claims

1. A method for detecting human presence or occupancy in an indoor environment comprising:

providing a database of models for determining human presence based upon acoustic events;

detecting the acoustic events in an audio signal;

determining human presence by comparing the acoustic events with the models; and

refining the models when human presence is verified, so that the models over time become more accurate and more specific to the acoustic events present in the audio signals of the particular indoor environment.

2. The method of claim 1, further comprising:

determining the probabilistic score where the acoustic events include;

indoor speech;

footsteps;

household sounds;

electronic audio or music.

3. The method of claim 1, wherein detecting the acoustic events in the audio signal includes:

extracting directions of the audio signals;

extracting magnitude of the audio signals

extracting frequency spectrogram of the audio signals;

extracting autocorrelation coefficients of the audio signals; and

extracting time information of the audio signals.

4. The method of claim 1, wherein determining human presence includes:

computing the likelihood scores for human presence and human absence;

determining human presence if the ratio of human-presence score to human-absence score is greater than a prescribed threshold.

5. The method of claim 4, wherein determining the likelihood score for human presence or absence in an indoor environment includes:

computing the detection probabilities of the acoustic events during human presence or absence;

computing from the human-presence model or human-absence model the likelihood score for human presence or absence from the detection probabilities of the acoustic events.

6. The method of claim 5, wherein computing the detection probabilities of acoustic events during human presence or absence includes:

computing the features corresponding to the acoustic events from microphone array signals;

computing the probability of features during human presence or absence by taking into account the prior probability distributions of features during human presence or absence in the indoor environment.

7. The method of claim 1, wherein refining the human-presence model or human-absence model includes:

initializing a generic model for human presence or absence;

computing the probabilistic score for human presence or absence from the joint probability of the acoustic events;

updating the prior probability distribution of the input features for the human-presence model or human-absence model.

8. The method for claim 7, wherein updating the prior probability distribution of the input features for the human-presence model or human-absence model includes:

confirmation of human presence or absence from other sensors;

compute score of human-presence or absence from the human-presence model or human-absence model.

9. The method of claim 8 further comprising:

when human-presence or absence is confirmed or the score of human presence absence exceeds a prescribed threshold;

updating the joint probability of sound events and human presence or absence;

updating the prior distribution of input features for the human-presence model or human absence model.

10. The method of claim 9, wherein the input features with updated prior distributions include:

direction of sound sources;

magnitude of the sound signal;

frequency spectrum;

autocorrelation coefficients;

time-of-day, day-of-week, and month-of-year.

11. The method of claim 9, wherein the sound events with updated joint probability include:

indoor human speech;

human footsteps

human-made household sounds.

12. A system for identifying individual(s) in an indoor environment, comprising:

a microphone array for receiving an audio signal;

a vocal-tract length estimator that takes in the signals from the microphone array;

a pitch estimator that takes in the signals from the microphone array;

an individual detector coupled to the pitch and vocal-tract-length estimators and configured to identify individual speaker(s) present within the indoor environment.

13. The system of claim 12, wherein the vocal-tract length estimator further includes:

a detector for schwa-like vowels that takes in the signal from the microphone array.

14. The system of claim 13, further comprising:

when schwa-like vowels are detected;

determining the formant peaks of vowels

determining the direction-of-arrival of the vowels;

determining the number of speakers from the direction-of-arrival of the vowels; and

storing the formant peaks for classification;

wherein;

if multiple speakers are present, configuring the formant peaks of the speakers to be mutually exclusive between the speakers.

15. The system of claim 12, wherein the pitch estimator further includes:

a detector for vowels of speech.

wherein;

when vowel are detected in speech;

extracting the pitch features of the vowel;

determining the direction-of-arrival of the vowels;

determining the number of speakers from the direction-of-arrival of the vowels; and

storing the pitch features for classification.

16. The system of claim 15, wherein storing the pitch features for classification includes:

if multiple speakers are present, configuring the pitch features from the speakers to be mutually exclusive between the speakers.

17. The system of claim 12, further comprising:

determining the pitch and vocal-tract-length features from the newly acquired data;

storing the features in a database for semi-supervised classification;

wherein;

if the features come from multiple speakers;

configuring the features to be mutually exclusive between the speakers during semi-supervised classification.