PRUD'HOMME, Pierre-Benoît (20 rue de Rosheim, Strasbourg, Strasbourg, F-67000, FR)
AUFRERE, Etienne (17 rue du faisan, Geispolsheim, Geispolsheim, F-67118, FR)
HUONG, Quang (15 rue de la Bourgogne, Erstein, Erstein, F-67150, FR)
GRANDEMANGE, Eric (15 Boulevard de la Marne, Strasbourg, Strasbourg, F-67000, FR)
PRUD'HOMME, Pierre-Benoît (20 rue de Rosheim, Strasbourg, Strasbourg, F-67000, FR)
AUFRERE, Etienne (17 rue du faisan, Geispolsheim, Geispolsheim, F-67118, FR)
HUONG, Quang (15 rue de la Bourgogne, Erstein, Erstein, F-67150, FR)
| CLA I MS An algorithm for detecting activation of a tactile pressure sensor having a mechanic structure comprising the steps consisting of: measuring periodically an input quantity (V0; fo) of the sensor, when it is not pressed, the period being set according to sensor dynamic requirements; computing a current idle quantity (Vidie ; fidie) depending on the input quantity (V0; fo) measured at step a); computing an activation threshold (AVP; ΔίΡ) based on a quantity characterization of the sensor mechanic structure and depending on the idle quantity (Vidie; ) defined at step b); comparing the sensor input quantity (V0; fo) with the last defined idle quantity (Vidie; ) increased by the activation threshold (AVP; ΔίΡ) computed in step c) in order to determine whether the sensor is pressed or not. The algorithm according to claim 1 , wherein the quantity measured at step a) is the input frequency (f0) defining a current idle frequency (fidie) at step b) and computing an activation threshold (ΔίΡ) based on a frequency characterization of the sensor mechanic structure and depending on the idle frequency (fidie) defined at step b). The algorithm according to claim 1 , wherein the quantity measured at step a) is the input voltage (V0) defining a current idle voltage (Vidie) at step b) and computing an activation threshold (ΔνΡ) based on a voltage characterization of the sensor mechanic structure and depending on the idle voltage (Vidie) defined at step b). 4. The algorithm according to any of claims 1 to 3, wherein the period at step a) is dynamically set upon detection of environmental constraint changes. 5. The algorithm according to claim 4, wherein the period is dynamically set to a fast period mode defined by a shorter period, when either a push has been detected as released or the idle quantity has dropped under a predetermined threshold. 6. The algorithm according to any of claims 1 to 5, wherein the idle quantity (Vidie; fidie) in step b) is defined as the average value between the previous defined value of the idle quantity and the last measured input voltage in step a). 7. The algorithm according to any of claims 1 to 6, wherein the quantity characterization of the sensor mechanic structure is made all over the range of strength used by sensor. 8. The algorithm according to any of claims 1 to 7, wherein computation in step c) is based on an equation of the sensor behaviour determined by the quantity characterization setting the activation threshold (ΔνΡ; ΔίΡ) to be near the strength request for such activation. 9. The algorithm according to any of claims 1 to 7, wherein computation in step c) is based on a correspondence table between the idle quantity (Vidie; fidie) determined at step b) and a corresponding activation threshold (ΔνΡ; ΔίΡ) determined by the quantity characterization of the sensor mechanic structure. 10. The algorithm according to any of the preceding claims, wherein step c) further consists in computing a non-activation threshold (AVR; AfR) to determine when the sensor is released after having been pressed, said non-activation threshold being strictly lower than the activation threshold (ΔνΡ; ΔίΡ). 1 1 . The algorithm according to claim 10, wherein the non-activation threshold (AVR; AfR) is either directly based on the activation threshold (AVP; ΔίΡ) or computed from the quantity characterization of the sensor mechanic structure. 12. The algorithm according to any of the preceding claims, for detecting activation of a tactile pressure sensor among a plurality of tactile pressure sensors being mechanically dependant on each other's, wherein it further comprises the step of determining which sensor has been pressed by comparing for each sensor the gap between the sensor input quantity (V0; fo) with the last defined idle quantity (Vidie; fidie) and selecting the sensor with the highest gap. 13. The algorithm according to claim 12, wherein it further comprises the steps of calibrating each sensor and storing for each sensor a quantity deviation for a given force applied, said quantity deviation being used for computing the activation threshold at step b). 14. The algorithm according to claim 12 or 13, wherein it further comprises the step of adjusting the actuation threshold based on a temperature characterization of each sensor mechanic structure. 15. The algorithm according to claim 13 or 14, wherein each gap computation is adjusted by a predetermined coefficient for each sensor based on their stored quantity deviation and/or their temperature characterization. |
BUTTON
TECHNICAL FIELD
The present invention relates to an algorithm for detecting activation of a push button comprising a tactile pressure sensor. Such algorithm is applicable to all products that contain Force Sensing Resistor® (FSR) technology working in preloaded condition and how to control the force to detect sensor activation through a rigid mechanical part (e.g. detect activation above a specified actuation force, regardless the velocity of the actuation).
BACKGROUND OF THE INVENTION
FSR® Integration Guide & Evaluation Parts Catalog With Suggested Electrical Interfaces, which is enclosed herewith by reference, provides an overview of the Force Sensing Resistors technology along with some basic electrical interfaces using such FSRs. In particular, figure 17 of this document shows an FSR current-to-voltage converter described by the following equation:
VOUT = VREF/2 X [1 + RG/R F SR] Another example is given in Figure 18 of this document showing a simple force to frequency converter with an FSR device as the feedback element around a Schmitt trigger. At zero force, the FSR is an open circuit. Depending on the last stage of the trigger, the output remains constant, either high or low. When the FSR is pressed, the oscillator starts, its frequency increasing with increasing force.
It is known from document WO 2009/070503, use of a force sensing resistor where an FSR output which is a function of the resistance is measured. Whether a change in magnitude of the FSR output during a time interval is greater than a threshold is determined. A touch applied on the FSR is detected during the time interval if the change is greater than the threshold. This document nevertheless presents some drawbacks among with the fact that the FSR is a pre-loaded sensor which does not take into account the environment in which the FSR is integrated decreasing the reliability of the detection of the FSR activation.
It is also known from the document US 5 440 237, a method and apparatus for normalizing electronic sensor data to correct for variations in individual sensor transfer characteristics which are not known in advance. A general characteristic transfer function of sensor type of interest is determined empirically. For that purpose, a baseline response is acquired from each sensor to get an indication of the transfer characteristics of each individual device. The baseline response is determined under some "preloaded" condition or "at rest" condition. Then a specific transfer function is determined for each individual sensor by applying the corresponding baseline response to the general characteristic transfer function. As for the previous document, it results from the above method that detection of the sensor activation does not take into account the environment of the sensor.
It is also known from the document US 2006/007172, a force sensing resistor with a calibration feature. For that purpose, it comprises the steps of measuring a calibration resistance while the FSR is disconnected and determining a correction factor such that it is the ratio between the nominal value and the current measured value. Subsequent measurements of the FSR resistance are then multiplied by the correction factor in order to scale them to appear as if they were measured from a nominal FSR resistance.
In existing solutions, electronic measures, through a microprocessor
Analogic Digital Converter (ADC), a voltage or frequency that will be the image of the FSR resistance/pressure. As the FSR resistance variation is assumed to follow a 1 /F law, F being the force applied, thus the output voltage or frequency is a straight line as shown on Figure 1 B and 1 C.
The relation between the pressure applied on FSR and the resistance variation is given on Figure 1 A. Therefore, theoretically whatever the resistance is, for a constant force AF, there is constant voltage AV as shown in Figure 1 B or a constant frequency Af as shown in Figure 1 C.
Mainly used algorithms are generally based on high pass filter with long time constant (1 6 samples @ 20ms sampling period). Further, the output value of this filter that depends on the velocity and force of the actuation is compared to thresholds for detecting any change on the sensor.
Another document, US 2009/066673, describes a self-calibration method of a pressure sensor. This method consists in periodically calibrating the sensor when not activated, determining an idle tension according to the current and previous values, compensating the measured value by compensation data and activating the sensor according to an updated threshold value. The compensation data may be in particular the relation between the measured voltage and the force applied on the sensor.
It is further known from the document US 5 514 040 calibration methods of FSR sensors. One of these methods is based on a delay value calibration, while another one is based on an adjustment value calibration. However, none of these methods takes into account the state of the push (released or pressed) and the mechanic structure of the sensor to make the calibration. Another document EP 0 535 907 describes a calibration method of a push button based on a measure done while the button is pressed by the user. Another document US 6 456 952 describes a calibration method of a touch screen including key areas and drift areas. Following the detection of contacts in key areas and surrounding drift areas, an adjustment vector is applied to focus detection. Still another document, EP 1 602 907 describes a calibration method based on two extreme measures, one measure without pressure and another measure with a maximum pressure. Finally, another document US 2007/107487 describes a generic calibration method.
Each of the above presented solutions present some of the following several drawbacks. In particular above presented systems will be influenced by electromagnetic interference (EMI) / compatibility (EMC). Indeed, protection stage against automotive EMI/EMC constraint will influence the voltage input value. Consequently the input voltage is no more linear over the whole R F SR variation range. Further, the mechanical environment, temperature and humidity have an influence on the mechanical pre-load system and FSR intrinsic characteristics. Indeed, FSR resistance variation is not really in 1 /x, but in z/x (with 0,5 < z < 1 ,5) due to the mechanical structure of the sensor which is not a straight line over the whole resistance variation range. Moreover, it appears that high dynamic pressure detection is not reliable since with a high pass filter, the low velocity activation cannot be detected (actuation duration has to be less than the time constant, which may be not compliant with sensor requirements). Known algorithms do not take into account dynamic variation of preload detection. Indeed, due to mechanical warping on mechanical parts or on the FSR sensor itself, the preload applied on the sensor can dynamically change. For instance, between two consecutive presses, the FSR resistance level may change. Furthermore, current solutions do not take into account fast variations of the system such as dynamic variation of preload detection due to mechanical warping on the sensor that may change the for instance the sensor resistance level between two presses.
SUMMARY OF THE INVENTION
One goal of the present invention is to overcome the aforecited drawbacks by providing an algorithm for reliably detecting activation of a tactile pressure sensor with a strength activation management for which over a strength threshold a sensor activation has to be detected regardless the stroke velocity and possible fast variations of the system.
For that purpose, according to a first aspect, the invention concerns an algorithm for detecting activation of a tactile pressure sensor with a mechanic structure comprising the steps consisting of:
a) measuring periodically an input quantity of the sensor, when it is not pressed, the period being set according to sensor dynamic requirements; b) computing a current idle quantity depending on the input quantity measured at step a);
c) computing an activation threshold based on a quantity characterization of the sensor mechanic structure and depending on the idle quantity defined at step b);
d) comparing the sensor input quantity with the last defined idle quantity increased by the activation threshold computed in step c) in order to determine whether the sensor is pressed or not.
Such algorithm has the advantage of ensuring reliable detection of activation of the sensor while with a very shirt overall processing delay and extended resistance value range monitoring. Further, period setting according to sensor dynamic requirements allows to discriminate between slow user actuations and fast environment changes. According to another embodiment, the quantity measured at step a) is the input frequency defining a current idle frequency at step b) and computing an activation threshold based on a frequency characterization of the sensor mechanic structure and depending on the idle frequency defined at step b). Such algorithm provides the ability to use sensors far from the processing unit and to increase the acceptable dynamic range.
According to another embodiment, the quantity measured at step a) is the input voltage defining a current idle voltage at step b) and computing an activation threshold based on a voltage characterization of the sensor mechanic structure and depending on the idle voltage defined at step b).
According to another embodiment, the period at step a) is dynamically set upon detection of environmental constraint changes. Such dynamic setting of the measurement period in step a) ensures auto- adaptation to fast variations of the system constraints as well as environmental changes such as mechanical preload, temperature or humidity. Advantageously, the period is dynamically set to a fast period mode defined by a shorter period, when either a push has been detected as released or the idle quantity has dropped under a predetermined threshold.
According to another embodiment, the idle quantity in step b) is defined as the average value between the previous defined value of the idle quantity and the last measured input voltage in step a).
According to another embodiment, the quantity characterization of the sensor mechanic structure is made all over the range of strength used by sensor.
According to another embodiment, the quantity characterization is made by way of an end of line tester or by a measurement done on several parts during the development.
According to another embodiment, computation in step c) is based on an equation of the sensor behaviour determined by the quantity characterization setting the activation threshold to be near the strength request for such activation.
According to another embodiment, computation in step c) is based on a correspondence table between the idle quantity determined at step b) and a corresponding activation threshold determined by the quantity characterization of the sensor mechanic structure.
According to another embodiment, step c) further consists in computing a non-activation threshold to determine when the sensor is released after having been pressed, said non-activation threshold being strictly lower than the activation threshold and preferably also under another strength threshold non-activation has to be detected. Advantageously, the non-activation threshold is either directly based on the activation threshold or computed from the quantity characterization of the sensor mechanic structure.
According to a second aspect, the invention concerns an algorithm for detecting activation of a tactile pressure sensor among a plurality of tactile pressure sensors being mechanically dependant on each other's, wherein in addition to the first aspect it further comprises the step of determining which sensor has been pressed by comparing for each sensor the gap between the sensor input quantity with the last defined idle quantity and selecting the sensor with the highest gap.
According to another embodiment, it further comprises the steps of calibrating each sensor and storing for each sensor a quantity deviation for a given force applied, said quantity deviation being used for computing the activation threshold at step b).
According to another embodiment, it further comprises the step of adjusting the actuation threshold based on a temperature characterization of each sensor mechanic structure. According to another embodiment, each gap computation is adjusted by a predetermined coefficient for each sensor based on their stored quantity deviation and/or their temperature characterization. According to an advantageous embodiment, the periodicity in step a) is set according to sensor dynamic requirements in order to discriminate between slow user actuations and fast environment sensor changes.
According to another aspect, the invention concerns a push button comprising a tactile pressure sensor controlled by the algorithm according to the first aspect.
According to another embodiment, it concerns a plurality of push buttons each comprising a tactile pressure sensor controlled by the algorithm according to the first aspect.
BRIEF DESCRIPTION OF THE DRAWINGS
Other features and advantages of the invention will appear upon reading the following description which refers to the annexed drawings in which:
- Figure 1 A, already described, is a graphic showing the relation between pressure applied on FSR and its resistance;
- Figure 1 B, already described, is a graphic showing the relation between pressure applied on FSR and voltage variation;
- Figure 1 C, already described, is a graphic showing the relation between pressure applied on FSR and frequency variation;
- Figure 2 represents the sensor processing means according to an embodiment of the invention; - Figure 3 represents a diagram of the algorithm for detecting activation of a tactile pressure sensor according to a first embodiment;
- Figure 4 represents a diagram of the algorithm for detecting activation of a tactile pressure sensor according to a second embodiment;
- Figure 5 represents a diagram of the algorithm for detecting activation of a tactile pressure sensor according to a third embodiment;
- Figure 6A represents the voltage characterization of the sensor mechanic structure;
- Figure 6B represents the frequency characterization of the sensor mechanic structure;
- Figure 6C represents the frequency characterization of the sensor against temperature;
- Figure 7 represents a schematic showing the comparisons between the input quantity and the computed thresholds according to any of the three embodiments;
- Figure 8 shows a system comprising four neighbouring push buttons.
DETAILED DESCRIPTION OF THE INVENTION
With reference now to Figure 2 to 8, we will describe in more details several embodiments of algorithms for detecting the activation of a tactile pressure sensor. In order to fulfil system requirements and be independent of product use context, it is proposed to use an algorithm that has to
- be independent of the environment constraints (mechanical preload, temperature and humidity) by computing an idle quantity provided by the electronic stage when the push is not pressed; - adapt the activation / non-activation thresholds to the processed idle quantity, to take into account the non-linearity of the system, over the whole range of the sensor resistance variation; and
- determine the push state (pressed or released) by comparing the input quantity to the idle quantity and to the computed actuation thresholds.
As it can be seen on Figure 2, it represents schematically the sensor processing means. An input quantity such as a voltage (V 0 ) or a frequency (f 0 ) will be periodically measured at the input (V/f_FSR_Linear) of an analogic to digital converter (ADC Driver) for a voltage or of a timer input for a frequency. In order to define a current idle quantity (V id i e ; f id i e ) depending on the measured input quantity when the sensor is not pressed, it is provided with filtering means such as a low-pass filter (LPF) for filtering said input quantity. It further comprises press and release threshold calculation means for computing an activation threshold (AV P ; Af P ) for detection when the sensor is pressed and also preferably a non-activation threshold (AV R ; Af R ) for detection when the sensor is released based on this defined idle quantity and on a corresponding quantity characterization of the sensor mechanic structure. This quantity characterization can be done by design, during development, or directly with or by the product, at the end of line tester (EOLT) calibration services. A look up table (LUT), or an equation of the variation law, may be used to compute the thresholds. Finally it comprises push state calculation means comparing the sensor input quantity (V 0 , fo) with the last defined idle quantity (V id i e ; f id i e ) and the activation / non-activation thresholds (AV P /AV R , Af P /AVf R ) in order to determine whether the sensor is pressed or not. At the output, a push state is delivered.
According to a first embodiment, it is provided to use a FSR driver that generates a periodic square signal which frequency is related to 1 /RFSR- Frequency acquisition has preferably to be performed by a microprocessor thanks to an input capture pin. In this case, FSR driver supplies a digital output, which is much more robust to EMC perturbations and permits to use remote sensors. There is no more limitation due to the operational amplifier output saturation voltage and so the dynamic range is increased. Figure 3 represents a diagram of the algorithm for detecting activation of a tactile pressure sensor based on a frequency quantity acquisition.
A first step a) consists in measuring periodically the input frequency (f 0 ) of the sensor when it is not pressed.
A second step b) consists in processing the idle frequency based on the input frequency measured at step a). The level of the idle frequency fidie(t) is periodically updated in order to be auto-adaptive against the environment constraint. Thus, a periodic task with a long period shall apply a low-pass filter, preferably of the first order, to the input frequency (f 0 ), only when the sensor is detected as released. The period of the f id i e calculation shall be set according to the system dynamic requirements, i.e. period value shall be able to discriminate between slow user actuations and fast environment system changes. For that purpose, the output of the 1 st order low pass filter is advantageously given by the following formula:
facile = if" 1 idle + ί"θ) / 2
A third step c) consists in computing actuation thresholds according to the idle frequency defined at step b). For that purpose, it is provided to use a frequency characterization of the sensor mechanic structure, shown at Figure 6A. To let the system less sensitive to EMI filtering, mechanic structure and sensor technology influence, it is proposed to make a frequency characterization of the mechanic structure, all over the range of strength used by the system and implement it on the software. This characterization can be done by design, during development, or directly with or by the product, at an end of line tester calibration services. With such characterization, it is possible to adapt the system non-linearity and to have an accurate value of the strength for the sensor activation or non- activation.
This characterization can be defined with an equation of the system behaviour. The equation of the curve (Figure 6A) is then programmed in a memory of the local user terminal and can define itself the Af P frequency requested depending on f id i e . So, whatever the preload is, the system can manage the Af P and Af R frequency thresholds to be near the strength request for the activation or non-activation. The kinds of the equation used or its order define the precision of the system.
Alternatively, the frequency characterization could be done in the form of a table of Af P and Af R values depending on f id i e . The software will then adapt both threshold values Af P and Af R based on the f id i e voltage of the sensor.
An example of table is given below:
Af P (Hz) for sensor fidie frequency (KHz)
activation
1.4 < fidie < 1.6 910
1.6 < fidie < 2.35 620
2.35 < fidie < 2.87 440
2.87 < fidie < 3.25 330
3.25 < fidie < 3.54 260
3.54 < fidie < 3.77 200
3.77 < fidie < 3.96 170
3.96 < fidie < 4.12 140
4.12 < fidie < 4.24 120
4.24 < fidie < 4.35 100
4.35 < fidie < 4.45 90 Depending on the accuracy needed, a table for Af R can also be done, otherwise, Af R can be calculated from Δί Ρ . The deep of the tables define the precision of the system.
A fourth step d) consists in comparing the sensor input frequency (f 0 ) with the last defined idle frequency (f id i e ) increased by the activation threshold (Δί Ρ ) in order to determine the push state, i.e. whether the sensor is pressed or not. An advantageous way of determining the push state is detailed below in relation with Figure 7.
Figure 4 represents a diagram of the algorithm for detecting activation of a tactile pressure sensor according to a second embodiment based on a voltage quantity acquisition. The algorithm is similar to the one of the first embodiment with the difference that voltages instead of frequencies are considered.
Step a) consists in measuring periodically the input voltage (V 0 ) of the sensor when it is not pressed.
Step b) consists in processing the idle voltage based on the input voltage measured at step a). The level of the idle voltage V id i e (t) is periodically updated in order to be auto-adaptive against the environment constraint. Thus, a periodic task with a long period shall apply a low-pass filter, preferably of the first order, to the input voltage (V 0 ), only when the sensor is detected as released. The period of the V id i e calculation shall be set according to the system dynamic requirements, i.e. period value shall be able to discriminate between slow user actuations and fast environment system changes. For that purpose, the output of the 1 st order low pass filter is advantageously given by the following formula:
V n id |e = (V n - 1 id | e + V n o) / 2 Step c) consists in computing actuation thresholds according to the idle voltage defined at step b). For that purpose, it is provided to use a voltage characterization of the sensor mechanic structure, shown at Figure 6B. To let the system less sensitive to EMI filtering, mechanic structure and sensor technology influence, it is proposed to make a voltage characterization of the mechanic structure, all over the range of strength used by the system and implement it on the software. This characterization can be done by design, during development, or directly with or by the product, at an end of line tester calibration services. With such characterization, it is possible to adapt the system non-linearity and to have an accurate value of the strength for the sensor activation or non-activation. More generally, with the characterization, it is possible to adapt the system non-linearity and to have an accurate value of the strength for the sensor activation or non-activation.
This characterization can be defined with an equation of the system behaviour. The equation of the curve (Figure 6) is then programmed in a memory of the local user terminal and can define itself the AV P voltage requested depending of V id i e . So, whatever the pre-load is, the system can manage the AV P and AV R voltage thresholds to be near the strength request for the activation or non-activation. The kinds of the equation used or its order define the precision of the system.
Alternatively, the voltage characterization could be done in the form of a table of AV P and AV R values depending of V id i e . The software will then adapt both threshold values AV P and AV R based on the V id i e voltage of the sensor. The deep of the tables define the precision of the system. An example of table is given below:
Depending on the accuracy needed, a table for AV R can also be done, otherwise, AV R can be calculated from AV P . The deep of the tables define the precision of the system.
A fourth step d) consists in comparing the sensor input voltage (V 0 ) with the last defined idle voltage (V id i e ) increased by the activation threshold (AVp) in order to determine the push state, i.e. whether the sensor is pressed or not. An advantageous way of determining the push state is detailed below in relation with Figure 7.
Figure 5 represents a diagram of the algorithm for detecting activation of a tactile pressure sensor according to a third embodiment based on a frequency quantity acquisition and adapted to a system with a plurality of sensors as shown in Figure 8.
A first step a) consists in measuring periodically the input frequency (f 0 ) of the sensor when it is not pressed. A second step b) consists in processing the idle frequency based on the input frequency measured at step a). The level of the idle frequency fidie(t) is periodically updated in order to be auto-adaptive against the environment constraints. For instance, due to mechanical constraint applied on the sensor or to mechanical geometry changes consecutive to a push actuation, the level of the preload seen by the sensor may quickly change and then the idle frequency may become inaccurate in case it is computed by a task with a long and static period. Thus, for preventing from inaccurate computation, this task period would preferably also be dynamically modifiable to a fast period in case of fast environmental change detection. The task is then activated in a "Fast Mode" in order for the system to adapt the idle frequency to the new environmental conditions. This "Fast Mode" is triggered either when an actuation has been detected as released, because the idle frequency may have changed after the press or when the idle frequency has suddenly dropped less than an adjustable threshold.
A third step d ) consists in computing actuation thresholds. The resistance variation is linearized by the electronic stage and provides a periodic signal, which frequency directly stands for the force applied on the film sensor. The slope of the line - Frequency (Hz) against Force (N) - is different from one push to the other one. It is therefore proposed to calibrate each push and store the frequency deviation for a given force applied in a non volatile memory. Thus, it is possible to easily change the sensitivity of the system and make it possible to detect an actuation for a modifiable threshold.
A fourth step c2) consists in adjusting the actuation thresholds against temperature. To let the system less sensitive in temperature due to mechanic and sensor technology, it is proposed to make a characterization of the system structure over the range of temperature and implement it on the software. The resistance variation is linearized by the electronic stage and provides a periodic signal, which frequency directly stands for the force applied on the FSR film sensor. Only the slope of this line depends on the temperature as shown on Figure 6C.
The variation of the slope according to the temperature may be integrated by the algorithm via a curve stored in a non volatile memory, and the standard frequency deviation given for a known force (stands for the slope of the line at a nominal temperature) is also stored in the memory and can be specific for each sensor or plurality of sensors depending on the mechanical structure. Then the actuation threshold is given by the standard frequency deviation multiplied by an adjustment coefficient for the temperature. This coefficient depending on the temperature could be provided by a table of data. This table can also be stored in a non volatile memory and be adapted according the film behaviour.
In case of a multi push system, it is further provided an intermediary discrimination step c3) prior to determination of a push state, which consists in discriminating which sensor is pressed among the plurality of sensors through a discrimination process. In case of use of multiple pushes that are mechanically dependant on each others, it is proposed to discriminate the sensor (i.e. push) which the maximum force is applied on. To know which sensor is being actuated, computation of the difference between the current frequency and the idle frequency (which stands for the current force applied and the preload applied in idle position) is done. This difference is computed for each push of the system and is adjusted by a coefficient, which stands for the transmission rate of the mechanical structure surrounding the sensor on the film. This coefficient (for each push) is highly linked with the mechanic and is preferably fixed since the mechanical structure is known and characterized.
The calculation follows the given formula, for each push number n:
Af = |f n 0 - F n id ie| x Coef n The discrimination consists in finding the push number n which is maximizing this Δί.
A sixth step d) consists in comparing the sensor input frequency (f 0 ) with the last defined idle frequency (f id i e ) increased by the activation threshold (Δί Ρ ) in order to determine the push state, i.e. whether the sensor is pressed or not. An advantageous way of determining the push state is detailed below in relation with Figure 7. It is further provided to detect any fast change in environmental conditions, such as fast frequency drop and manual handling (consecutive to a push press for instance). In those conditions, the mechanical influence on the sensor can change, for instance modification of preload distribution between all pushes. Then, in case of fast change detection, the algorithm will trigger the "Fast Mode" provided in step a).
Figure 7 represents a schematic showing the comparisons between the input quantity Q 0 (being either V 0 or f 0 ) and the computed thresholds AQR/AQR (either AV P /AV R or Af P /Af R ). The push state is calculated by comparing the input quantity to the idle quantity level Q id i e (either V id i e or f id i e ) +/- the threshold values ΔΟ Ρ and AQ R .
The algorithm prevents from fast state toggling (around the decision threshold) by using a "release" decision threshold (Q 0 >Qi d i e +AQp) different from "pressed" decision threshold (Ο 0 <Ο½ θ +ΔΟρ,). AQ R is defined to be always less than ΔΟ Ρ for generating a hysteresis filter to take into account the signal behaviour when the push is released. Thus, the scheduling of the associated task is compliant with a high reactivity of the system (for instance, an activation detection can be done in less than 20ms).
Figure 8 shows a system comprising four neighbouring push buttons and which algorithm is above described in relation with the third embodiment in Figure 5. Indeed, the present could be extended for several pushes of the tactile pressure sensor kind. Although, the present description was directed to a single push button, it can be extended to a more complex mechanic system with several pushes located near one from the other. The problem in this kind of system with several pushes is not to activate all the pushes when the user presses on only one button, i.e. button 3 in the represented example. The known common solution is to make a mechanic structure with mechanical isolation or with a big distance between each push, so that the force apply on one push has no effect on the other pushes. Such solution is not admissible since it results in a much bigger mechanical volume of the system than needed.
To let the system less sensitive to the mechanic influence between each pushes, it is proposed to make a quantity characterization (either voltage or frequency) of the mechanic structure from one push to the other, all over the range of strength used by the system and implement it on the software. With this characterization the software knows the influence that has the pressure of one push on all the other pushes and can correct it accordingly. Thus, if the user presses the push button 3, with the table of influence given by the quantity characterization, the software can correct the thresholds (AV P , AV R or Δί Ρ , Af R ) or the idle quantity (V id i e or f id i e ) of the other pushes. As the threshold level of the other pushes has changed, the software let them inactive, i.e. as being in a released state. The software knows the push on which the user is pressing, by calculating the higher gap between its input quantity and its idle quantity level. As the system is not linear, this calculation has to be preferably corrected according to its idle quantity.
Having described the invention with regard to certain specific embodiments, it is to be understood that these embodiments are not meant as limitations of the invention. Indeed, various modifications, adaptations and/or combination between embodiments may become apparent to those skilled in the art without departing from the scope of the annexed claims.
Next Patent: BIOMARKER FOR HYPERTENSIVE DISORDERS OF PREGNANCY