Patent Application
20060262104
The present invention relates to devices that utilize vibrations propagating through a plate due to a contact to obtain information related to the contact, for example a vibration sensing touch input device.
Touch input devices can provide convenient and intuitive ways to interact with electronic systems including computers, mobile devices, point of sale and public information kiosks, entertainment and gaming machines, and so forth. Various touch input device technologies have been developed including capacitive, resistive, inductive, projected capacitive, surface acoustic wave, infrared, force, and others. It is also possible to form a touch input device from a touch plate provided with vibration sensors that detect vibrations propagating in the touch plate due to a touch input and determine the touch location from the detected vibrations.
The present invention provides a method that includes detecting vibrations propagating in a panel, developing a signal representative of the vibrations, generating an energy spectrum for the developed signal, and analyzing the energy spectrum for the presence of one or more features characteristic of ambient noise. From this analysis, noise signals can be distinguished from signals generated by valid panel contacts.
The present invention also provides a method for use with a vibration sensitive touch panel that includes the steps of characterizing a first feature set for energy spectra associated with panel vibrations caused by valid panel contacts, characterizing a second feature set different than the first for energy spectra associated with vibrations caused by noise, and comparing signals obtained from measured panel vibrations to the first feature set and the second feature set to determine whether the measured panel vibrations are indicative of a valid panel contact or a noise event.
Further, the present invention provides a vibration sensing touch panel system that includes vibration sensors coupled to a touch plate and configured to generate signals in response to vibrations propagating in the touch plate, and electronics in communication with the vibration sensors and configured to analyze an energy spectrum of the signals generated by the vibration sensors to determine the presence or absence of spectral features indicative of acoustic noise. The electronics can also be configured to determine contact location for signals determined not to originate from acoustic noise.
The above summary of the present invention is not intended to describe each embodiment or every implementation of the present invention. Advantages and attainments, together with a more complete understanding of the invention, will become apparent and appreciated by referring to the following detailed description and claims taken in conjunction with the accompanying drawings.
FIGS. 5(a) and 5(b) show the time domain and frequency domain, respectively, for detected vibrations caused by a noise event.
FIGS. 5(c) and 5(d) show the time domain and frequency domain, respectively, for detected vibrations caused by a valid touch contact.
FIGS. 6(a)-6(c) show a set of histograms indicating the number of measured occurrences of acoustic noise events that exhibit a particular spectral ratio p1, a particular impulse ratio p2, and a particular combination of p1 and p2.
FIGS. 7(a)-7(c) show a set of histograms indicating the number of measured occurrences of acoustic noise events that exhibit a particular spectral ratio p1, a particular impulse ratio p2, and a particular combination of p1 and p2.
While the invention is amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and will be described in detail. It is to be understood, however, that the intention is not to limit the invention to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the invention as defined by the appended claims.
The present invention relates to systems and methods for distinguishing vibrations propagating in a panel due to contact with the panel from vibrations propagating in the panel due to ambient acoustic noise coupled into the panel. For example, vibrations sensing touch panels that determine touch position based on the vibrations caused by the contact of a touch input on the panel can be susceptible to false recording of a touch event due to spurious panel vibrations caused by noise.
The present invention can be advantageously applied to discern between an acoustically generated noise event and a valid touch contact on a vibration sensing touch panel. Methods and systems of the present invention can be used alone or in combination with impulse reconstruction or other noise detection methods for improved rejection of spurious points under some circumstances. Examples of other noise detection methods include using a separate microphone that is acoustically isolated from the panel to continuously monitor for ambient noise. Better discernment of spurious points through use of the present invention may also translate to less rejection of valid touches by impulse reconstruction or alternative touch point validation methods, and consequently improved effective sensitivity to light touches.
One mode of operation of a vibration sensing, or bending wave, touch panel is the input of energy by the contact of a touch implement such as a finger or stylus on the touch plate of the touch panel. Energy from the contact propagates in the form of bending waves from the contact point to a set of vibration sensors positioned in various locations on the touch plate, for example one in each of the corners of a rectangular plate. Each vibration sensor can be used to develop signals, and the signals can be cross-correlated to determine the position of the contact. A more accurate determination of contact position can be achieved by correcting for dispersion of the bending waves propagating in the touch plate. The following documents, each of which is incorporated by reference, disclose one or more of various vibration sensing touch panels, vibration sensing transducers and transducer arrangements, and methods to locate the contact, or determine other information related to the contact, based on analysis of the vibrations signals: EP1240617B1, WO2003005292, U.S. Pat. No. 6,871,149, and commonly assigned U.S. patent application Ser. Nos. 10/440,650, 10/683,342, 10/739,471, 10/750,291, 10/750,502, 10/750,290, 10/850,324, 10/850,516, 10/957,364, 10/957,234, 60/615,469, 11/025,389, 11/032,572, and 11/116,463.
In particular, a method referred to as “impulse reconstruction” has been disclosed in above-mentioned and incorporated U.S. Ser. No. 10/750,290. Impulse reconstruction can be used as a consistency check to verify the validity of a touch input point reported to the system. In this method, a scaling and phase factor can be applied to the signals from each transducer signal channel to reverse the effects of the propagation in the panel (e.g., dispersion), thereby “reconstructing” the original impulse that was created by the touch contact event. For example, given a determined touch location, preferably after removing dispersion effects, the signals received at each transducer can be time reversed back to the point of the original contact, thereby reconstructing the original impulse.
One use of impulse reconstruction is to distinguish between valid inputs, which result in similar reconstructed impulses from each sensing channel, and spurious points generated by noise events. Such noise events can include mechanical events such as contacts to the bezel of an integrated touch screen, which can couple vibrations into the touch plate through the supporting gasket, and acoustic events where ambient sound is incident on the touchscreen, generating bending waves in the panel. Both of these noise sources can generate transient signals in the panel that may incorrectly return a touch input location when the signals are analyzed by the location algorithm. The impulse reconstruction method helps to discern between a true contact location and a spurious noise-generated point. As discussed in this document, however, certain acoustic noise conditions may be difficult to discern using only impulse reconstruction. In at least these cases, methods and systems of the present invention can be used to discern true plate contact events from noise events.
The present invention involves analyzing the shape of the energy spectrum developed from signals detected by vibration sensing transducers in response to vibrations propagating in the touch plate. Methods of the present invention take advantage of the phenomenon that the vibrations generated in the touch plate due to typical acoustic noise events are distinguishable from those generated due to typical touch contact events. In particular, a touch contact to the panel generally creates greater low frequency energy content than does an acoustic noise event. While this is partially dictated by the typical frequency spectra of acoustic noise versus touch contact events, it is also caused by frequency dependent, or “microphonic,” coupling of the acoustic noise to the panel, which disfavors coupling of lower frequencies.
There are two distinct frequency bands over which the mechanism for microphonic pickup of acoustic vibrations by a panel differs, those frequencies that are less than the coincidence frequency, termed “below coincidence,” and those frequencies that greater than the coincidence frequency, termed “above coincidence.” The coincidence frequency is the frequency at which the speed of bending wave propagation in the panel is the same as that in the ambient medium, typically air, which will be assumed to be the ambient medium in this document without loss of generality. Below the coincidence frequency the bending wavespeed is less than that in air, whereas above the coincidence frequency the bending wavespeed is greater than that in air.
The coincidence frequency is a function of properties of the panel, including material and thickness. For 2 mm thick glass, a typical thickness for bending wave touch panels, the bending wave velocity is given by the following dispersion relation:
k=0.53×√{square root over (ω)},
where ω is angular frequency and k is the wavevector of the bending wave, and 0.53 is a factor that folds in various physical properties of the panel. The wavevector relates to the bending wave velocity, vB, by the following equation:
The frequency at which the bending wave velocity equals the speed of sound in air (343 meters per second) is therefore 5.3 kHz.
For below coincidence frequencies, there is no direct matching between the sound wave in air and a bending wave in the panel. Any coupling of sound below coincidence is characterized by an approximately omni-directional response. For above coincidence frequencies, there can be a direct match between the waves in the air and vibrations in the panel and a directional response, as indicated by
Microphonic pickup above coincidence (λa less than λb) is significantly more efficient than below coincidence (λa greater than λb), and has a directional response. At coincidence (λa equal to λb), the most efficient angle for microphonic pickup is along the plane of the panel (Θ=90°). As the frequency increases (and the panel bending wavespeed increases), the matching angle moves towards normal incidence (Θ=0°).
If the noise source 360 is relatively far away from the panel (e.g., on the scale of the panel size or more) then the wavefront 370 will be relatively spread out and flat by the time it reaches the panel. As such, there is likely no detectable single point of incidence, and any attempt to reconstruct an impulse based on a reported touch point would likely yield a spread out impulse, indicating a false touch event that can be ignored or cancelled. Furthermore, ambient sounds from noise sources far away from the panel and that are not centered on the panel will yield dissimilar signals between signal correlation channels associated with different transducer pairs. Comparing the reconstructed impulse to the impulse between channels and the sharpness of the impulse would likely reject these cases.
When the sound source 360 is relatively close to the panel 310 (e.g., within a distance smaller than the size of the panel), the situation can be more problematic for previously implemented solutions such as impulse reconstruction to address. In this case the sound 370 is likely to propagate out from the sound source 360 over the surface of the panel 310, resulting in stronger coupling into the panel 310 above the coincidence frequency. Such strong coupling may trigger the system to attempt to determine the location of what at first seems to be a touch contact, even for relatively weak sound sources. Furthermore, at and around the frequencies of this strong coupling, the speed of propagation of sound waves in the air is similar to the propagation speed in the panel, and as such the signals detected at each vibration sensing transducer 320 are similar to what would be detected when a touch contact to the panel 310 occurred at a location directly under the sound source 260. As a result, the impulse reconstruction method may interpret the noise as an approximate impulse event corresponding to a location under the sound source, thereby seeming to confirm a touch input rather than indicating a false touch due to noise.
In the present invention, it is recognized that coupled ambient acoustic noise has a characteristic energy spectrum distinct from the energy spectrum of touch contact events, even for cases in which the acoustic noise pickup events give rise to a signal that as a reconstructed impulse looks like a real touch and is erroneously not rejected. For acoustic noise coupled into a touch plate, a general rise in high frequency pickup over low frequencies would be expected due to more efficient pickup above coincidence. In addition, for sound sources that are relatively close to the panel and having a relatively smooth frequency output (i.e., not strongly peaked), the coupled vibrations would be expected to show a maximum for frequencies close to the panel coincidence frequency. For sound sources that are strongly peaked around a frequency band, as is often the case when like objects are struck together to generate the sound, a maximum near the coincidence frequency may not be readily observed.
The shape of the energy spectrum of detected noise-induced vibrations will also depend on the spectrum of the sound output from the noise source. In principle, the noise source could have a frequency response that is strongly weighted towards low frequencies, which in turn could compensate for the coincidence effect, yielding a more even spectral shape in the panel pickup signal. In such a case, however, the low frequency airborne sound would spread out from the contact point significantly faster in air than any induced bending wave in the panel. The likely end result is a signal that when reconstructed would yield a spread out impulse that would be recognized as a false touch by the impulse reconstruction algorithm and rejected. Observing a spectral characteristic of higher contribution from above coincidence frequencies and lower contribution from below coincidence frequencies is therefore likely to reveal spurious points generated by acoustic noise that are not likely to be rejected by the impulse reconstruction technique. Conversely, the cases in which the distinction between above and below coincidence frequencies is muted, the impulse reconstruction technique is likely to catch and reject the spurious point. As such, spectral shape methods can be combined in a complementary fashion with impulse reconstruction to better discern common noise events while better detecting valid contacts.
Once noise events are discerned, the spectral characteristics of the signal can be further analyzed to determine the type of noise event (e.g., hand clap, clinking of metal objects, etc.) in cases where it is desirable to do so. For example, the spectral content of the signal can be compared to various sample signals recorded during a calibration step.
When a true, or valid, contact occurs on the panel, the typical actions of the user and the implement used to contact the panel give rise to a wide and varied bandwidth of induced bending waves, typically including a high level of low frequency energy. Such low frequency energy contains relatively little useful location information because of the long spatial wavelength of the bending waves propagating in the panel at these frequencies, which tends to blur out spatial resolution. Indeed, when determining touch position, the low frequency energy is preferably filtered so as to emphasize the higher frequency energy, thereby reducing the dynamic range requirements on the signal chain. However, as discussed, detecting the low frequency vibrations can be useful in distinguishing valid touches from noise.
Signals detected from the valid contact, when processed through the location algorithm and impulse reconstruction, are expected to have a significantly greater level of low frequency energy than for ambient acoustic noise events (normalized for similar high frequency levels). The actual spectral shape of signals will also depend on the electronic and/or digital filtering in the signal chain. Even so, a true touch should be well characterized by an increased ratio of below coincidence energy to above coincidence energy. In one embodiment of the present invention, a threshold ratio of below coincidence energy to above coincidence energy can be used to distinguish true touches from acoustic noise, providing an improved touch sensor through enhanced rejection of acoustically generated spurious points.
The signals upon which measurements of the spectral shape are based can be one or both of: pickup from a separate sensor, for example a dedicated transducer optimized for pickup of low frequency energy (exemplary transducers include those disclosed in commonly assigned U.S. patent application Ser. Nos. 10/683,342 and 10/957,364, previously referred to and incorporated by reference); and pickup from the sensing channels that are optimised for contact location.
In a particular embodiment, whether detected signals derive from an acoustic noise source can be determined by calculating sums of the amplitudes in two ranges in the frequency domain, which can be illustrated in reference to
Amplitude sums of the signals over each of these frequency domains can be calculated, represented as S(f1 . . . f2) and S(f3 . . . f4). Then, by determining a parameter p1 representing a ratio of the two sums, namely p1=S(f1 . . . f2)/S(f3 . . . f4), it can be determined that the signals represented by FIGS. 5(a) and 5(b) includes more high frequency content than the signals represented by FIGS. 5(c) and 5(d). Further, it can be determined whether p1 exceeds a threshold value, in this case the threshold value being set so that the signals represented by FIGS. 5(a) and 5(b) do not exceed the threshold, indicating acoustic noise, and the signals represented by FIGS. 5(c) and 5(d) exceed the threshold, indicating a potential valid touch input. This ratio and threshold approach can also be used in combination with other observations or measurements to make more accurate judgments and/or to further distinguish and filter noise events from valid touch events.
For example, if the frequency limits in
Additional parameters can optionally be folded into the analysis. For example, an additional parameter, p2, can be obtained from impulse reconstruction algorithms, where p2 represents a measure of the alignment of the reconstructed impulses for each of the signal channels, a higher degree of alignment indicating a higher likelihood of a impulse event, and therefore a valid contact.
For example, starting with a set of reconstructed impulses, one for each signal channel, an alignment function can be calculated for slices of time that range over an interval shorter than the duration spanned by the reconstructed impulse data. For each time slice, the minimum value of the reconstructed impulses across all the channels within that time slice is divided by the sum of the absolute values of the reconstructed impulses over all the channels and from the first sample up to the sample some interval below the time slice being evaluated (e.g., up to the sample 20 microseconds before the time slice being evaluated). Calculating the alignment function in this manner yields a sharp positive spike at the point where the reconstructed impulses best align, and highest magnitude of the spike is a measure of how well the reconstructed impulses are aligned. As such, the parameter p2 can be defined as the maximum value of the alignment function over the interval calculated.
In comparison, if only p1 was used and a threshold of about 0.07 to 0.08 was set, all or nearly all valid touch contacts would be correctly interpreted as touches, but some noise events would not be interpreted as noise. If only p2 was used and a threshold of about 0.07 to 0.08 was set, all or nearly all noise events would be correctly interpreted as noise, but some touch contacts would not be interpreted as valid touches. By using a suitable formula that combines p1 and p2, occurrences of correct valid touch interpretations and correct noise event interpretations can each be increased.
In addition to distinguishing between signals from valid touch contacts and signals from acoustic noise events, the present invention may have application to sensing tracing movements over the panel. During tracing, the movement of the tracing implement (such as a stylus, pen or finger) on the panel generates a noise-like signal. Dispersion corrected correlation functions may be used to locate the position of the contact, for example using methods disclosed in WO2003005292, previously referred to and incorporated by reference. The spectral shape of the detected input will be related to the contact pressure, the velocity of the movements on the plate, the implement type, and so forth. As such, typical spectra may be correlated to known classes of movements, for example by comparing detected spectra against a table of characteristics of different movements, pressures, and contact implements that may be recorded during a calibration procedure.
Steady state acoustic noise can also be picked up by the panel and correctly interpreted as noise rather than as the presence of a moving contact. In some cases, however, steady state noise might be interpreted as a moving trace on the panel. The most commonly occurring spurious event is a reported touch moving around the central portion of the touch panel when no such movement is occurring. This can happen in typical operating environments where a sound far from the touch screen generates similar noise-like signals on each of the sensing channels, resulting in central correlation function peaks similar to what would result from a valid touch in the middle of the screen. If the noise source is close to the screen, then peaks indicative of a touch location under the contact can result as discussed above with respect to impulsive acoustic noise. In these circumstances, a measure based on the spectral shape of the pickup signals may help distinguish between a valid contact, whose spectral shape and trace characteristics fit a pre-determined template, and steady state acoustic noise, which has a different spectral shape than that of a typical finger or stylus contact traced on the panel with a tracing movement similar to what is exhibited by the spurious contact. As discussed above, the signals created by the noise event generally exhibit less low frequency content in its energy spectrum than for valid touches.
The foregoing description of the various embodiments of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teaching. It is intended that the scope of the invention be limited not by this detailed description, but rather by the claims appended hereto.
