Estimating a number of occupants in a region

Patent metadata

Field	Value
Publication number	US-10607147-B2
Application number	US-201615182901-A
Country	US
Kind code	B2
Filing date	Jun 15, 2016
Priority date	Jun 15, 2016
Publication date	Mar 31, 2020
Grant date	Mar 31, 2020

Abstract

A method for estimating a number of occupants in a region comprises receiving a time series of sensor values detected over a period of time by a motion sensor sensing motion in the region. A spread parameter indicative of the spread of the sensor values is determined. The number of occupants in the region is estimated based on the spread parameter.

First claim

We claim:

1. A method for quantitatively estimating a number of occupants in a region, comprising:
detecting with at least one motion sensor a series of sensor values over a period of time,
wherein the at least one motion sensor comprises a passive infrared (PIR) sensor for detecting the sensor values indicative of infrared radiation in said region;
using processing circuitry, determining a spread parameter indicative of a spread of the sensor values detected over said period of time, wherein the spread parameter is indicative of a spread of a probability distribution corresponding to the sensor values detected over said period of time, said spread of the probability distribution indicative of a width of a distribution of frequencies of occurrence of different sensor values within said series of sensor values detected over said period of time; and
using processing circuitry, determining, when the region is occupied by at least one occupant, a quantitative estimate of the number of occupants in the region in dependence on said spread parameter.

2. The method of claim 1 , wherein the probability distribution comprises a Laplace distribution.

3. The method of claim 1 , comprising, for each of a plurality of windows of time, estimating the number of occupants in dependence on the sensor values.

4. The method of claim 1 , wherein the number of occupants is estimated by applying a regression model to map the spread parameter to the number of occupants.

5. The method of claim 4 , comprising selecting, in dependence on the spread parameter or a previously estimated number of occupants of the region, one of a first regression model and a second regression model for mapping the spread parameter to the estimated number of occupants.

6. The method of claim 5 , wherein the first regression model is selected when the spread parameter or the previously estimated number of occupants is less than a predetermined threshold, and the second regression model is selected when the spread parameter or the previously estimated number of occupants is greater than the predetermined threshold.

7. The method of claim 5 , comprising estimating the number of occupants for each of a plurality of windows of time in dependence on the sensor values detected during the corresponding window of time;
in response to detecting that the number of occupants estimated using the first regression model is greater than a first threshold number of occupants for more than a first number of previous windows of time, switching to using the second regression model for a subsequent window of time; and
in response to detecting that the number of occupants estimated using the second regression model is less than a second threshold number of occupants for more than a second number of previous windows of time, switching to using the first regression model for a subsequent window of time.

8. The method of claim 5 , wherein the first regression model comprises a linear regression model.

9. The method of claim 5 , wherein the second regression model comprises a log-linear regression model.

10. The method of claim 1 , comprising excluding selected sensor values from the determination of the spread parameter.

11. The method of claim 1 , comprising clustering segments of the sensor values into a plurality of clusters, and excluding segments of the sensor values corresponding to one or more selected clusters from the determination of the spread parameter.

12. The method of claim 11 , wherein the clustering is performed using a model obtained by machine learning based on training data.

13. The method of claim 12 , comprising adapting the model in response to the sensor values detected by the at least one motion sensor in a period of time for which the number of occupants is being estimated.

14. The method of claim 13 , wherein the machine learning based on the training data uses a first inference technique; and
said adapting the model in response to the sensor values uses a second inference technique.

15. The method of claim 14 , wherein the first inference technique comprises a Markov Chain Monte Carlo inference technique, Beam sampling inference technique or Gibbs sampling inference technique; and
the second inference technique comprises an iterative Maximum-a-posteriori inference technique.

16. The method of claim 12 , wherein the model comprises one of:
a Hidden Markov Model, and
an infinite Hidden Markov Model.

17. The method of claim 1 , wherein said at least one motion sensor comprises a single motion sensor.

18. A data processing apparatus comprising a circuitry configured to perform the method of claim 1 .

19. A non-transitory computer-readable storage medium storing a computer program to control a data processing apparatus to perform the method of claim 1 .

20. The method of claim 1 , wherein said sensor values detected by the PIR sensor are not attributed to any particular spatial position within the region.

21. The method of claim 1 , comprising controlling a device using the quantitative estimate of the number of occupants in the region.

22. The method of claim 1 , comprising at least one of:
controlling a lighting, heating or air conditioning system based on the quantitative estimate of the number of occupants in the region;
controlling a burglar alarm system based on the quantitative estimate of the number of occupants in the region; and
providing the quantitative estimate of the number of occupants in the region to an emergency responder in an emergency response scenario.

23. A system comprising:
at least one motion sensor to detect motion in a region by detecting a series of sensor values over a period of time, wherein the at least one motion sensor comprises a passive infrared (PIR) sensor for detecting the sensor values indicative of infrared radiation in said region; and
processing circuitry to determine a spread parameter indicative of a spread of the temporal series of sensor values detected over said period of time, wherein the spread parameter is indicative of a spread of a probability distribution corresponding to the sensor values detected over said period of time, said spread of the probability distribution indicative of a width of a distribution of frequencies of occurrence of different sensor values within said series of sensor values detected over said period of time, and in dependence on said spread parameter, to determine, when the region is occupied by at least one occupant, a quantitative estimate of a number of occupants in the region.

Assignees

• Advanced Risc Mach Ltd

Inventors

• RAYKOV Yordan Petrov
• Özer Emre
• Dasika Ganesh Suryanarayan

Classifications

• G06N20/00

  Machine learning · CPC title

• G06Q50/16

  Real estate · CPC title

• G06Q10/00

  Administration; Management · CPC title

• G01J5/0025

  Living bodies (ear thermometers G01J5/0011; detecting, measuring or recording for diagnostic purposes A61B5/00) · CPC title

• G06N7/005 Primary

  Physics · mapped topic