TOUCH SENSITIVE DEVICE

• US 20130249831A1
• Filed: 07/25/2011
• Published: 09/26/2013
• Est. Priority Date: 07/23/2010
• Status: Abandoned Application

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Hiwave Technologies (UK) Ltd (Alphabet Inc.)
• Inventors: Harris, Neil
• Application Number:
  US13/811,904
• Publication Number:
  US 20130249831A1
• Time in Patent Office: Days
• US Class Current: 345/173
• CPC Class Codes: G06F 3/016 Input arrangements with force or tactile feedback as computer generated output to the user
• CPC Class Codes: G06F 3/041 Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means
• CPC Class Codes: G06F 3/0436 in which generating transducers and detecting transducers are attached to a single acoustic waves transmission substrate
• CPC Class Codes: G08B 6/00 Tactile signalling systems, e.g. personal calling systems

Abstract

A method of generating a set of filters comprising: determining an initial estimate of a filter to be applied to a signal associated with each transducer; defining a model of the system; calculating the vibration of the member as an output of the model of the system; calculating the output of the model by comparing the output of the model with a measured value; determining changed parameter values of parameters of the model; recalculating the error value for the output of the model by comparing the output of the model with the changed parameter values with the measured value; comparing the recalculated error value with the reference error value; setting the recalculated error value as the reference error value, setting the changed parameter values as the model parameters, and repeating the above steps, or outputting the model parameters; generating a set of new filters each using respective output model parameters.

Claims

1. A method of generating a set of filters for a touch sensitive device comprising a touch-sensitive member and a plurality of transducers mounted to the member, the method comprising:

a) determining an initial estimate of a filter to be applied to a respective signal associated with each transducer;

b) defining a model of the system whereby the relationship of vibration of the member to the respective signals can be calculated, the model having a plurality of parameters;

c) calculating the output of the model of the system;

d) calculating a reference error value for the output of the model by comparing the output of the model with a measured value;

e) determining changed parameter values of the parameters of the model;

f) recalculating the error value for the output of the model by comparing the output of the model with the changed parameter values with the measured value;

g) comparing the recalculated error value with the reference error value;

h) if the compared recalculated error value is less than the reference error value, setting the recalculated error value as the reference error value, setting the changed parameter values as the model parameters, and repeating the steps c) to h), or

if the compared recalculated error value is greater than the reference error value, outputting the model parameters;

generating a set of new filters each using respective output model parameters.

2. The method according to claim 1, wherein the filter has a plurality of coefficients, the number of filter coefficients being equal to the number of model parameters.

3. The method according to claim 1, wherein the model comprises an inverse of the filter.

4. The method according to claim 1, wherein the reference error value is calculated using:

5. The method according to claim 4, wherein determining changed parameter values of the parameters of the model comprises selecting parameters to reduce the value of SSE by finding the value of t that minimizes F(t) where

6. The method according to claim 5, comprising finding the value of t by estimating the vector gradient of error (grad (SSE)), changing all the model parameters proportionally to the estimated vector gradient of error, estimating F′

(t)=0 and F″

(t)=0 and determining t from the estimated derivatives of F(t) where

7. The method according to claim 6, comprising estimating the vector gradient of error with respect to each of the parameters by changing each parameter by a small amount, recalculating SSE and estimating the vector gradient of error using the difference between the reference value for SSE and the recalculated value of SSE.

8. The method according to claim 6, comprising estimating F′

(t=0) and F″

(t=0) by determining two new parameter sets, the first parameter set being changed proportionally to the estimated vector gradient of error by addition and the second parameter set being changed proportionally to the estimated vector gradient of error by addition, calculating SSE values for each of the two new parameter sets and calculating estimates for F′

(t=0) and F″

(t=0) from the new SSE values.

9. A method of generating a desired touch sensitivity in a touch sensitive device comprising a touch-sensitive member and a plurality of transducers mounted to the member, the method comprising:

generating a set of filters by carrying out the method of claim 1;

applying the set of filters to an output signal from each transducer to generate filtered output signals; and

using the filtered output signals to provide the desired touch sensitivity.

10. The method according to claim 9, wherein:

said determining an initial estimate comprises determining an initial estimate of a filter to be applied to a respective signal output from each transducer, whereby the filtered signal is generated by vibration of the member resulting from a physical touch on the member;

the model of the system is defined whereby the output signal can be calculated as a function of vibration of the member;

said calculating the output of the model comprises calculating the output signal as an output of the model of the system;

said calculating a reference error value comprises calculating a reference error value for the output of the model by comparing the output of the model with a measured value of the output signal when a known touch is applied to the member; and

said recalculating the error value comprises recalculating the error value for the output of the model by comparing the output of the model with a measured value of the output signal when a known touch is applied to the member.

11. The method according to claim 1, wherein the reference error value is calculated using:

12. The method according to claim 11, wherein recalculating the error value comprises perturbing each time impulse response h by a small amount α

at each transducer and calculating a new error value from

13. The method according to claim 12, wherein recalculating the error value comprises defining the gradient vector as;

follows;

14. The method according to claim 13, wherein recalculating the error value comprises determining the value at which the gradient vector is zero and recalculating the error value for this value.

15. The method according to claim 1, comprising determining the initial filter as a time-reversal filter.

16. The method according to claim 1, comprising determining the initial filter as a simultaneous multi-region filter.

17. The method according to claim 2, comprising determining the initial estimate using arbitrary filter coefficients.

18. The method according to claim 1, comprising determining the initial estimate using an infinite impulse response filter.

19. The method according to claim 1, wherein the desired touch sensitivity is to produce a maximum response to a touch at one location on the touch sensitive member and a minimum response to a touch at a second location on the touch sensitive member.

20. A touch sensitive device comprising

a touch-sensitive member,

a plurality of transducers mounted to the member, and

a processor configured to carry out the method according to claim 1.

43. A computer program comprising program code which, when executed on a processor of a touch sensitive device, will cause the touch sensitive device to carry out the method claim 1.

21. A method of generating a set of filters for a touch sensitive device comprising a touch-sensitive member and a plurality of transducers mounted to the member, the method comprising:

choosing a set of frequencies for use in the filters;

determining an impulse response of a filter for each respective transducer to be applied to a signal associated with each transducer;

calculating the transfer function of each filter, wherein each filter has a transfer function with at least one pole and at least one zero, and calculating the transfer function of each said filter comprises,

determining at least one pole coefficient which determines at least one pole;

determining, using said at least one pole coefficient, a pole representation of the transfer function which filters said input signal using said at least one pole;

using said respective impulse response, calculating at least one zero coefficient which determines at least one zero; and

combining said pole representation of the transfer function with said at least one zero coefficient to calculate said transfer function of said filter; and

generating a set of filters comprising said calculated filters.

22. The method according to claim 21 wherein each said filter is an infinite impulse response filter.

23. The method according to claim 22 wherein said infinite impulse response filter has a transfer function of the form:

24. The method according to claim 23, wherein the pole coefficients are expressed as

a_k,0=−

2Re(p_k) and a_k,1=|p_k|² and the transfer function is written as

25. The method according to claim 23 or claim 21, comprising determining said pole representation of the transfer function from:

u_0,k;

=1 u_1,k;

=−

(a_k,0·

u_0,k) u_i+2,k;

=−

(a_k,0·

u_i+1,k+a_k,1·

u_i,k)

26. The method according to claim 21, comprising determining the pole coefficients from the chosen set of frequencies.

27. The method according to claim 21, comprising calculating the zero coefficients using a generalized inverse matrix with the zeros being defined by:

28. The method according to claim 21, comprising calculating zero coefficients for zeros which are not paired with said determined pole coefficients.

29. A method of generating a desired touch sensitivity in a touch sensitive device comprising a touch-sensitive member and a plurality of transducers mounted to the member, the method comprising:

generating a set of filters by carrying out the method of claim 21;

applying the set of filters to an output signal from each transducer to generate filtered output signals; and

using the filtered output signals to provide the desired touch sensitivity.

30. The method according to claim 29, wherein:

said determining an impulse response of the filter comprises determining an impulse response of the filter for each respective transducer to be applied to a signal output by said transducer, wherein the output signal is generated by vibration of the member resulting from a physical touch on the member.

31. The method according to claim 21 wherein the impulse response of each filter is derived from a measured value of vibration of the member.

32. A touch sensitive device comprising

a touch-sensitive member,

a plurality of transducers mounted to the member, and

a processor configured to carry out the method according to claim 21.

33. A method of generating a set of filters for a touch sensitive device comprising a touch-sensitive member and a plurality of transducers mounted to the member, the method comprising:

choosing a set of frequencies for use in the filters;

calculating a set of transfer functions of respective filters for each transducer to be applied to a signal associated with each transducer;

wherein each filter has a transfer function with at least one pole and at least one zero and calculating the transfer function of each said filter comprises,

determining at least one pole coefficient which determines said at least one pole;

determining, using said at least one pole coefficient, a pole representation of the transfer function which filters said input signal using said at least one pole;

using an eigenvector method to determine at least one zero coefficient which determines said at least one zero; and

combining said pole representation of the transfer function with said at least one zero coefficient to determine said transfer function for said filter; and

generating a set off filters comprising said calculated filters.

34. The method according to claim 33 wherein said filter is an infinite impulse response filter.

35. The method according to claim 33, further comprising determining zero coefficients for zeros which are not paired with said determined pole coefficients.

36. A method of generating a desired touch sensitivity in a touch sensitive device comprising a touch-sensitive member and a plurality of transducers mounted to the member, the method comprising:

generating a set of filters by carrying out the method of claim 33;

applying the set of filters to an output signal from each transducer to generate filtered output signals; and

using the filtered output signals to provide the desired touch sensitivity.

37. The method according to claim 36, wherein:

said calculating a set of transfer functions of respective filters for each transducer comprises calculating a set of transfer functions of respective filters for each transducer to be applied to a signal output by each transducer, wherein the output signal is generated by vibration of the member resulting from a physical touch on the member.

38. The method of claim 37 wherein the eigenvector method comprises:

defining a first parameter of the desired touch sensitivity;

using an eigenvector method to identify sets of possible filter transfer functions which satisfy the first parameter;

defining a second parameter of the desired touch sensitivity;

using an eigenvector method to select from the identified sets of possible filter transfer functions which satisfy the first parameter, sets of possible filter transfer functions which satisfy the second parameter.

39. The method of claim 38 wherein the eigenvector method further comprises:

defining at least one further parameter of the desired touch sensitivity; and

using an eigenvector method to select from the previously selected sets of possible filter transfer functions which satisfy the first parameter and second parameter, sets of possible filter transfer functions which best satisfy the at least one further parameter.

40. The method of claim 39 wherein the selection selects a combination of possible filter transfer functions.

41. The method of claim 40 wherein the selection selects a linear combination of possible filter transfer functions.

42. A touch sensitive device comprising

a touch-sensitive member,

a plurality of transducers mounted to the member, and

a processor configured to carry out the method according to claim 33.