DATA ENTRY DEVICE

Abstract

A data input device (10) comprises a functional surface (12) adapted to receive pressure from a user, such as the pressure of a user's finger (14); a piezoelectric film (16) having at least one axis of sensitivity (30, 34) and disposed below the functional surface and adapted to emit an electrical signal as a function of the pressure received by the functional surface; and a computing unit (18) configured to receive the signal generated by the piezoelectric film in response to pressure on the functional surface; and to determine from this signal an area and / or shape of the functional surface pressed by the user using a classification algorithm. Abstract figure: Figure 1

Description

Title of the invention: DATA INPUT DEVICE technical field

[0001] This disclosure falls within the domain of data input devices, and in particular touch surfaces and keyboards. Previous technique

[0002] Keyboards are generally composed of a surface that is sensitive to finger pressure associated with each key. The keyboard may incorporate mechanical elements (for example, an inductive switch), capacitive elements (for example, electrodes), or even ultrasonic elements.

[0003] Regardless of the solution, the pressure-sensitive areas must be arranged in a matrix or in a specific layout unique to a keyboard. Depending on the technology used, various mechanical limitations arise in their integration. For example, mechanical switches can be aesthetically unappealing in addition to requiring a travel distance, while inductive switches suffer from their magnetic sensitivity. As for capacitive matrices, they do not function well when positioned behind metal.

[0004] It would therefore be beneficial to have a keyboard or more generally a data input device that is free from these constraints. Summary

[0005] This disclosure improves the situation.

[0006] A data input device is proposed comprising a functional surface adapted to receive pressure from a user, such as the pressure of a user's finger; a piezoelectric film having at least one axis of sensitivity and being disposed below the functional surface and adapted to emit an electrical signal as a function of the pressure received by the functional surface; and a computing unit configured to: receive the signal generated by the piezoelectric film in response to the pressure on the functional surface; and determine from this signal an area and / or a shape of the functional surface pressed by the user by means of a classification algorithm.

[0007] By "below," we refer to an axis orthogonal to the plane of the two largest dimensions of the functional surface. The underside of the functional surface extends, in use, on the side opposite the face in contact with the user, in particular the user's finger.

[0008] The functional surface can thus form a human-machine interface of the touch surface type. Preferably, the functional surface forms, in use, a surface external surface of a motor vehicle. The invention thus makes it very easy to transform any portion of the bodywork or other external surface of a motor vehicle into a touch surface.

[0009] According to another aspect, a motor vehicle such as a car including the above device is proposed.

[0010] The features described in the following paragraphs may optionally be implemented independently of each other or in combination with each other:

[0011] - the shape is a point, a line or a two-dimensional figure.

[0012] - the computing unit configured to determine from the signal generated by the film piezoelectric timing sequence of pressing of the functional surface using a classification algorithm.

[0013] - the input device is a keyboard, the functional surface includes a plurality of zones, each zone corresponding to a key, and the calculation unit is configured to determine from this signal the key corresponding to the zone of the functional surface pressed by the user.

[0014] - the keys are arranged in a line and the piezoelectric film is a single piezoelectric film having a sensitivity axis arranged parallel to the line of keys.

[0015] - the keys are arranged in a matrix according to rows and columns, and the film pie zoelectric is a unique piezoelectric film having at least two non-parallel axes of sensitivity.

[0016] - the keys are arranged in a matrix according to rows and columns, and the film pie zoelectric consists of a first piezoelectric film having a first axis of sensitivity and a second piezoelectric film having a second axis of sensitivity, the first piezoelectric film being arranged with respect to the second piezoelectric film so as to have the first axis of sensitivity at a non-zero angle, preferably at 90 degrees, from the second axis of sensitivity.

[0017] - the first piezoelectric film is disposed opposite the functional surface of so that the first sensitivity axis is parallel to the lines of the key matrix, and the second piezoelectric film is arranged opposite the functional surface so that the second sensitivity axis is parallel to the columns of the key matrix.

[0018] - the classification algorithm is of set-theoretic type.

[0019] - the classification algorithm is of the boosting type, for example AdaBoost (Adaptive Boosting), Gradient Boosting, or XGBoost (Extreme Gradient Boosting).

[0020] - the piezoelectric film is directly bonded to the functional surface. Brief description of the drawings

[0021] Other features, details and advantages will become apparent from the detailed description below and from the analysis of the accompanying drawings, in which: Fig. 1

[0022] [Fig-1] schematically shows a data input device according to a method of embodiment, in the form of a keyboard. Fig. 2

[0023] [Fig.2] shows a partial, transparent top view of the input device data from [Fig.1]. Fig. 3

[0024] [Fig.3] shows an example of electrical signals observed when a user presses a part of the data input device of the [Fig.l]. Fig. 4

[0025] [Fig.4] shows an example of an electrical signal (left) and its conversion (right) into time-domain slope graph of said signal. Fig. 5

[0026] [Fig.5] shows examples of graph segmentation into time windows Temporal slope of electrical signals observed when a user presses a part of the data input device. Description of the implementation methods

[0027] Reference is now made to [Fig. 1]. A data input device 10 comprises a functional surface 12 adapted to receive finger pressures 14 from a user, a piezoelectric film 16 disposed below the functional surface 12 and adapted to emit an electrical signal as a function of the finger pressures 14 of the user on the functional surface 12, and a computing unit 18 configured to receive the signal generated by the piezoelectric film 16 in response to the finger pressure 14 on the functional surface 12, and to determine from this signal an area of ​​the functional surface 12 pressed by the user.

[0028] The data input device 10 makes it possible to identify the shape and position of a pressure applied to the functional surface 12 by the user using one or more fingers. The shape of the pressure can be a point (for example, if the data input device 10 is a keyboard), a line, or a two-dimensional figure (for example, a circle, a zigzag, or a triangle).

[0029] The functional surface 12 includes an outer face 20 that receives the pressure of the user's fingers 14, and an inner face 22 connected to the piezoelectric film 16. The outer surface 20 can receive the pressure of the user's fingers 14 directly, for example as a touch surface, or via the interaction of a mechanical device, such as a keyboard key. The functional surface 12 can be a monolithic surface or be made up of a plurality of adjacent surfaces.

[0030] The functional surface 12 can be any surface suitable for transmitting the pressure of the user's fingers to the piezoelectric film 16. Therefore, the functional surface 12 can be relatively rigid as long as the piezoelectric film 16 can detect an atomic displacement in the functional surface 12. The functional surface 12 can be of any material and shape.

[0031] The piezoelectric film 16 can be directly glued to the functional surface 12. In another embodiment, an element is disposed between the piezoelectric film 16 and the functional surface 12. This element allows the user's finger pressure to be transmitted.

[0032] The piezoelectric film 16 includes at least one piezoelectric film having at least one sensitivity axis. Thus, depending on the implementation of the data input device 10, there may be a single film with a single sensitivity axis for detecting pressures arranged in a line (for example, keys arranged in a line or a line-type movement, in which case detection is only required along one dimension / axis), or there may be a single film with two or more sensitivity axes for detecting pressures along two dimensions / axes (for example, keys arranged in rows and columns (matrix) or a two-dimensional figure). Alternatively, the piezoelectric film 16 may include two or more piezoelectric films, each having a sensitivity axis for detecting pressures along two dimensions.

[0033] The processing unit 18 analyzes the finger pressures 14 on the functional surface 12 to deduce their location and shape. To do this, the processing unit uses a classification algorithm. The number of classes in the classification algorithm depends on the number of shapes / locations of finger pressures to be identified. For example, for a 9-digit numeric keypad, the number of classes will be 9.

[0034] The calculation unit 18 can be located near or far from the functional surface 12.

[0035] Now referring to [Fig.2], an example of the implementation of the data input device 10 is shown in the form of a 9-key numeric keypad 24, each key corresponding to a number from 1 to 9. The functional surface 12 is for this purpose divided into 9 zones 26. This division can be virtual or physical.

[0036] Since the data input device 10 is a numeric keypad 24 having rows and columns (i.e., two-dimensional), the piezoelectric film 16 has at least two axes of sensitivity. In the case of [Fig. 2], it was chosen that the piezoelectric film 16 be made up of two piezoelectric films, each having one axis of sensitivity. It would also have been possible to have only one piezoelectric film, but that This one has two or more axes of sensitivity. It would also have been possible to have more than two piezoelectric films.

[0037] In the case of [Fig. 2], therefore, the piezoelectric film 16 consists of a first piezoelectric film 28 having a first sensitivity axis 30, and a second piezoelectric film 32 having a second sensitivity axis 34. The sensitivity axes 30 and 34 are shown as dashed lines in [Fig. 2]. The first piezoelectric film 28 is positioned relative to the second piezoelectric film 32 such that the first sensitivity axis 30 lies at a non-zero angle to the second sensitivity axis 34. In the example of [Fig. 2], the angle between the first sensitivity axis 30 and the second sensitivity axis 34 is 90 degrees. However, it is possible that the first sensitivity axis 30 may be at a non-zero angle other than perpendicular to the second sensitivity axis 34.Having the sensitivity axes 30, 34 of the piezoelectric film(s) allows us to detect a shape that would be either at a two-dimensional position on the functional surface 12, such as a point on the numeric keypad 24, or a two-dimensional shape, such as a circle, triangle or zigzag.

[0038] With reference now to [Fig. 3], an example of the signal generated by the first piezoelectric film 28 is shown when the functional surface 12 is touched at different points by the user. It can be seen that for each point of impact of the finger 14 (for example, the user presses an area of ​​the functional surface 12 corresponding to key 1, or to key 3), a different electrical signature E1, E2, ... E1 is generated by the piezoelectric film. Thus, from these signatures, the processing unit 18, using the classification algorithm, will be able to determine which area and / or shape of the functional surface 12 was pressed by the user.

[0039] In order for the classification algorithm to use the electrical signatures, these are first converted into a 1-dimensional N x 1 vector, with the N characteristics of this signal. With reference to Figures 4 and 5, the signal processing part for obtaining the vectors corresponding to the electrical signatures will now be explained.

[0040] Figure 4 shows on the left an example of an electrical signal En as it may have been recorded, for example, in Figure 3. The electrical signal En in the left-hand figure shows the evolution of the intensity of the electrical signal over time. This signal is first converted into a graph (Figure 4, right) showing the evolution of the slope of the signal En over time. This conversion makes it possible to eliminate noise and highlight the variation of the signal. This graph of the evolution of the slope of the signal En over time is then divided into two or three time windows T1, T2, T3, as illustrated in Figure 5.

[0041] Figure 5 shows four graphs of the evolution of the slope of the electrical signal emitted by the piezoelectric films according to the area of ​​the functional surface (i.e., touch in the case of the keyboard). More precisely, graphs (A) and (B) are respectively the electrical signals emitted by the first piezoelectric film 28 when keys 1 and 3 are pressed. Graphs (C) and (D) are respectively the electrical signals emitted by the second piezoelectric film 30 when keys 1 and 3 are pressed.

[0042] Thus, for each electrical signature received by each piezoelectric film, transformed into a time-evolution graph of the slope of the electrical signal over time, the slope graph is divided into two, and possibly three, time windows. These windows correspond to phases of finger 14 pressing on the functional surface 12 (window T1), of finger 14 releasing from the functional surface 12 (window T2), and optionally, if this occurs, of a rebound once the finger is released (as for example in a keyboard) (window T3).

[0043] Figure 5 shows an example of dividing the slope of the electrical signal into three windows (with the bounce present). During the first time window T1, corresponding to the pressing of finger 14 on the functional surface 12, the slope is positive or negative. During the second time window T2, corresponding to the release of finger 14 from the functional surface 12, the slope has the opposite sign to the sign of the slope in the first time window T1. The boundary between the first and second phases is therefore at the moment of sign change. During the third time window T3, corresponding to the optional bounce, the slope has the opposite sign to the sign of the slope in the second time window T2 (respectively, positive or negative). The boundary between the second and third time windows T2, T3 is therefore at the moment of sign change.

[0044] According to one embodiment, a signal processing algorithm allows each time graph of the slope of the electrical signal to be divided into 2 or 3 time windows by looking at the changes in sign of the slope.

[0045] For each time window, features will be extracted. The supervised learning algorithm will use these features to classify the received electrical signature into a corresponding area and / or shape on the functional surface (i.e., for a keyboard, to identify the key pressed by the user, or for a touch surface, to identify the shape traced by the user). There can be a multitude of features, and the algorithm can use one, several, or all of the provided features to determine the corresponding area / shape on the functional surface.

[0046] According to one example, the features extracted from the time-domain slope graph of the electrical signal are: (1) Duration of the time window, for each of the 2 or 3 time windows identified above, determined on the time-domain evolution graph of the slope of the electrical signal, (2) Maximum amplitude of the (3) Time graph of the electrical signal in each time window, (4) Maximum slope amplitude of the time graph of the electrical signal slope in each time window, (5) Ratios of slope amplitudes for each time window and for the different piezoelectric films (if 2 piezos, 3 ratios calculated, if 3 piezos, 9 ratios calculated) calculated on the time evolution graph of the electrical signal slope over time.

[0047] The characteristics thus obtained are put into vector form as input data for the classification algorithm.

[0048] For example, if a single piezoelectric film is used [Feature 1, Feature 2, ..., Feature N]. If two piezoelectric films are used [Feature 1- Piezo 1, Feature 2- Piezo 1, ..., Feature N- Piezo 1, Feature 1- Piezo 2, Feature 2- Piezo 2, ..., Feature N- Piezo 2]. If M piezoelectric films are used [Feature 1- Piezo 1, Feature 2- Piezo 1, ..., Feature N- Piezo 1, Feature 1- Piezo 2, Feature 2- Piezo 2, ..., Feature N- Piezo 2, ..., Feature 1- Piezo M, Feature 2- Piezo M, ..., Feature N- Piezo M].

[0049] The classification algorithm will have as many classes as there are zones / shapes to detect. For example, for a 9-digit keypad, there will be 9 classes.

[0050] According to one embodiment, the classification algorithm is a set-theoretic algorithm. According to another embodiment, the classification algorithm is a boosting-type set-theoretic algorithm.

[0051] A boosting algorithm is a machine learning technique that combines the predictions of several weak or basic models to create a stronger and more accurate predictive model. Boosting belongs to the ensemble learning methods, where several models are trained and combined to make better predictions than the individual models.

[0052] In boosting, the basic models are generally simple and weak, such as decision trees with limited depth or shallow neural networks. Each basic model is trained on a subset of the training data or with different weights assigned to the data points. The models are trained iteratively, with each subsequent model focusing on the examples that were misclassified by the previous models.

[0053] During the training process, reinforcement assigns higher weights to misclassified examples, effectively giving them greater importance. Subsequent models are then trained to rank the misclassified examples, attempting to correct the errors made by the previous models. This iterative process continues until a predefined stopping criterion is met, such as a maximum number of models or a desired level of accuracy.

[0054] The final prediction is made by combining the predictions of all the basic models, generally through a weighted voting system. The weights are determined based on the performance of each basic model during training. The idea behind boosting is that by combining several weak models, each focusing on different aspects of the data, the overall model can capture complex relationships and improve predictive accuracy.

[0055] Popular boosting algorithms include AdaBoost (Adaptive Boosting), Gradient Boosting and XGBoost (Extreme Gradient Boosting).

Claims

Demands

1. A data input device (10) comprising: a. A functional surface (12) adapted to receive pressure from a user, such as the pressure of a user's finger (14), said functional surface (12) forming, in use, an external surface of a motor vehicle; b. A piezoelectric film (16) having at least one axis of sensitivity (30, 34) and being disposed below the functional surface and adapted to emit an electrical signal as a function of the pressure received by the functional surface; and c. A computing unit (18) configured to: i. Receive the signal generated by the piezoelectric film in response to pressure on the functional surface; and ii.Determine from this signal an area and / or a shape of the functional surface pressed by the user using a classification algorithm, the calculation unit (18) being configured to: - determine a time evolution of the slope of said signal generated by the piezoelectric film, - define, on this time evolution of the slope, time windows (T1, T2, T3) corresponding respectively to phases of pressing, releasing and optionally of rebound on the functional surface (12), - for each time window (T1, T2, T3), extract characteristics of said time evolution of the slope, and put these characteristics in vector form as input data for the classification algorithm.

2. Device according to claim 1, wherein the shape is a point, a line or a two-dimensional figure.

3. Device according to the preceding claim, wherein the computing unit is configured to determine from the signal generated by the piezoelectric film the time sequence of pressing the functional surface by means of a classification algorithm.

4. Device according to any one of the preceding claims, wherein the input device is a keyboard (24), the functional surface includes a plurality of zones (26), each zone corresponding to a key, and the processing unit is configured to determine from this signal the key corresponding to the zone of the functional surface pressed by the user.

5. Device according to claim 4, wherein the keys are arranged along a line and the piezoelectric film is a single piezoelectric film having a sensitivity axis arranged parallel to the line of keys.

6. Device according to claim 4, wherein the keys are arranged in a matrix along rows and columns, and the piezoelectric film is a single piezoelectric film having at least two non-parallel sensitivity axes.

7. Device according to claim 4, wherein the keys are arranged in a matrix in rows and columns, and the piezoelectric film consists of a first piezoelectric film (28) having a first sensitivity axis (30) and a second piezoelectric film (32) having a second sensitivity axis (34), the first piezoelectric film being arranged with respect to the second piezoelectric film so as to have the first sensitivity axis at a non-zero angle, preferably at 90 degrees, from the second sensitivity axis.

8. Device according to the preceding claim, wherein the first piezoelectric film is arranged opposite the functional surface so as to have the first axis of sensitivity parallel to the lines of the key matrix, and the second piezoelectric film is arranged opposite the functional surface so as to have the second axis of sensitivity parallel to the columns of the key matrix.

9. Device according to any one of the preceding claims, wherein the classification algorithm is of set-theoretic type.

10. Device according to the preceding claim, wherein the classification algorithm is of the boosting type, for example Adaptive Boosting, Gradient Boosting, or Extreme Gradient Boosting.

11. Device according to any one of the preceding claims, wherein the piezoelectric film is directly bonded to the functional surface.

12. Motor vehicle comprising the data input device according to any one of claims 1 to 11.

Images