SIGNAL PROCESSING SYSTEMS

• US 20140071095A1
• Filed: 02/22/2013
• Published: 03/13/2014
• Est. Priority Date: 08/27/2010
• Status: Active Grant

• Current Assignee: InputDynamics Limited
• Original Assignee: InputDynamics Limited
• Inventors: GODSILL, Simon
• Granted Patent: US 9,459,733 B2
• Time in Patent Office: Days
• US Class Current: 345/177
• CPC Class Codes: G06F 3/043 using propagating acoustic waves
• CPC Class Codes: G06F 3/0433 in which the acoustic waves are either generated by a movable member and propagated within a surface layer or propagated within a surface layer and captured by a movable member

Abstract

This invention relates to methods, apparatus, and computer programme code for processing acoustic signal data to determine where an object has been tapped with a stylus, finger nail or the like. The method involved storing a set of labelled training data comprising digitised waveforms from a sensor for taps at a plurality of different locations. The labelled training data is then processed to determine mean value and covariance data for the waveforms, which is afterwards used in conjunction with a digitised waveform of a tap at an unknown location to identify the location of the tap. Preferably the covariance is decomposed into a plurality of basis functions for each region each with a respective weighting, which are used to represent captured data for an unknown tap and parameters of the representation are classified to locate the tap.

Claims

1. A method of detecting the location of a tap on an object having at least one acoustic sensor, the method comprising:

storing a set of labelled training data, said labelled training data comprising digitised waveform data of waveforms captured from said sensor for taps at a plurality of different locations on said object in combination with location data indicating for each said waveform the location of the tap;

processing said labelled training data to determine, for each of a plurality of tap-sensing regions of said object, mean value data and covariance data for at least two said waveforms captured from taps in the said region, said mean value data defining a mean of said at least two waveforms and said covariance data defining covariance of said at least two waveforms;

capturing tap data comprising a digitised waveform of a tap at an unknown location captured by said sensor; and

determining a tapped region of said object from said tap data and from said mean value data and said covariance data from said plurality of regions.

2. A method as claimed in claim 1 wherein said determining of said tapped region comprises determining from said captured tap data and from said mean value data and said covariance data from said plurality of regions, probability data defining a probability, for each of said regions, that said unknown location is in said region;

and

determining a said tapped region of said object for said unknown location from said probability data.

3. A method as claimed in claim 2 wherein said determining of said probability data comprises determining, for each said region and for each of a plurality of possible tap times, a probability of capturing said captured tap data.

4. A method as claimed in claim 3 further comprising pre-processing said captured tap data to identify a subset of said possible tap times for said determining of said probability data.

5. A method as claimed in claim 3 wherein said determining of said tapped region further comprises scaling said probability for each of said regions by one or both of:

i) a probability of no tap having been captured; and

ii) a prior probability of a said tap in the region for each of said plurality of possible tap times.

6. A method as claimed in claim 1 wherein said processing of said labelled training data further comprises decomposing said covariance of said at least two waveforms into a plurality of basis functions each with a respective weighting, wherein said covariance data comprises data defining said respective weightings and providing basis function data defining said basis functions;

and wherein said determining of said tapped region comprises determining said tapped region from said tap data and from said mean value data and covariance data from said plurality of regions and from said basis function data.

7. A method as claimed in claim 6 wherein a different set of said functions is determined for each said region.

8. A method as claimed in claim 6 further comprising determining at least one said basis function common to said plurality of tap-sensing regions.

9. A method as claimed in claim 6 wherein said determining of said tapped region comprises, for each said region, estimating a set parameters to represent said captured data in terms of said basis functions for said region, and determining said tapped region by classifying said estimated sets of parameters for said regions.

10. A method as claimed in claim 6 wherein said determining of said tapped region comprises operating on said captured tap data with a filter using each of said basis functions.

11. A method as claimed in claim 6 wherein said determining of said tapped region comprises determining, for each said region, j, and for a set of possible tap times n₀, and for captured tap data y, a said probability p(y|j,n₀), or a log likelihood ratio dependent on said probability p(y|j,n₀), by applying to said data y a first FIR filter having coefficients determined by values of t, where t represents a said set of basis functions for region j, and a second FIR filter having coefficients determined by values of said mean value data.

12. A method of detecting the location of a tap on an object having at least one acoustic sensor, the method comprising:

storing a set of labelled training data, said labelled training data comprising digitised waveform data of waveforms captured from said sensor for taps at a plurality of different locations on said object in combination with location data indicating for each said waveform the location of the tap;

processing said labelled training data using a statistical model having a plurality of different parameters associated with each said region to determine from said waveforms, for each said region, statistical data representing the variability of said parameters resulting from tapping said object in said region;

capturing tap data comprising a digitised waveform of a tap at an unknown location captured by said sensor; and

determining a tapped region of said object from said tap data and said statistical data.

13. A method as claimed in claim 12 wherein said statistical model comprises a plurality of basis functions each with a respective weighting, and wherein said statistical data represents the variability of one or both of said basis functions and said weightings.

15. A method as claimed in claim 13, wherein a different set of said functions is determined for each said region;

and

wherein said determining of said tapped region comprises determining probability data defining a probability, for each of said regions, that said unknown location is in said region; and

determining a said tapped region of said object for said unknown location from said probability data;

wherein said determining of said probability data comprises determining for each said region and for each of a plurality of possible tap times, a probability of capturing said captured tap data.

17. A method as claimed in claim 12 wherein said statistical model includes a scale parameter representing a strength of a said tap;

wherein said determining of said tapped region comprises determining probability data defining a probability, for each of said regions, that said unknown location is in said region, and determining a said tapped region of said object for said unknown location from said probability data; and

wherein said determining of said probability data comprises determining, for each said region and for each of a plurality of possible tap times and for each of a plurality of possible values of said scale parameter, a probability of capturing said captured tap data.

14. (canceled)

16. (canceled)

18-33. -33. (canceled)

34. A method of detecting a tap on an object having at least one acoustic/vibration sensor, the method comprising:

capturing tap data comprising a digitised waveform of a tap on said object captured by said sensor;

processing said captured tap data by applying a tap detection procedure to generate time series tap metric data comprising data providing a detection metric of a potential said tap at a succession of times;

determining a classifier to classify said detection metrics of potential said taps into one or more detected taps; and

applying said classifier to said tap metric data to detect said tap on said object;

wherein said determining of said classifier comprises identifying a region of background noise in said captured tap data; and

determining said classifier from data in said region of background noise.

35. A method as claimed in claim 34 wherein said identifying said region of background noise comprises:

identifying potential said taps for which said detection metric is greater than a threshold value; and

defining said region of background noise to exclude time windows around each said identified potential said tap.

38. A method as claimed in claim 34 wherein said determining said classifier data in said regions of background noise comprises:

generating synthetic tap data comprising a combination of said data in said region of background noise and stored tap data representing taps on said object;

applying said tap detection procedure to both data in said region of background noise and to said synthetic tap data to provide, respectively, background noise tap metric data and synthetic-tap tap metric data; and

determining said classifier from said background noise tap metric data and said synthetic-tap tap metric data.

39. A method as claimed in any claim 34 wherein said classifier comprises a threshold level to classify a potential tap as a detected tap, and

wherein said determining said classifier data in said regions of background noise comprises:

setting a value of said threshold level;

generating synthetic tap data comprising a combination of said data in said region of background noise and stored tap data representing taps on said object;

applying said tap detection procedure to both data in said region of background noise and to said synthetic tap data to provide, respectively, background noise tap metric data and synthetic-tap tap metric data; and

determining said classifier from said background noise tap metric data and said synthetic-tap tap metric data.

41. A method as claimed in claim 34 for detecting the location of said tap on said object, wherein said stored tap data comprises data for taps at a plurality of different locations on said object, wherein said tap detection procedure is configured for determining a tapped region of said object from said tap data when said object is tapped in one of a plurality of tap-sensing regions of the object;

and wherein the method comprises determining a plurality of said classifiers, one for each said tap-sensing region.

36-37. -37. (canceled)

40. (canceled)

42-64. -64. (canceled)