Inspection method for press button, production method for electronic apparatus, and program

Abstract

To provide an inspection method for a press button in which determination accuracy can be enhanced, a production method for an electronic apparatus, and a program.SOLUTION: An inspection method comprises: a step for acquiring load waveform data which indicates a relationship between a stroke position of a press button and a depression load; a step for determining whether the press button is in a normal state, on the basis of the load waveform data; and a step for notifying a user of abnormality of the press button when the press button is in an abnormal state. The step for determining whether the press button is in the normal state or abnormal state, includes: inputting the load waveform data as input waveform data, to an auto encoder which has been pre-trained using previously acquired pieces of load waveform data from a plurality of normal press buttons as training data, to acquire output waveform data output from the auto encoder; and determining whether the press button is in the normal state or abnormal state, on the basis of a difference between the input waveform data and the output waveform data.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present disclosure relates to a push button inspection method, an electronic device manufacturing method, and a program. [Background technology]

[0002] Keyboards of electronic devices such as PCs and keys of electronic musical instruments are provided with push buttons that users can press with their hands to operate them. When manufacturing electronic devices and electronic musical instruments, the keystroke feel that a user feels when pressing a push button may be inspected to determine whether the keyboard or key is of good quality. Automating the inspection of push buttons is also being considered.

[0003] For example, Patent Document 1 describes a feel testing method in which data judged by a human to be normal is used as a standard, and the degree of deviation from the operating force at a stroke position that has the feel characteristic of the standard is used as a judgment index to determine the degree of agreement. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2006-292454 A Summary of the Invention [Problem to be solved by the invention]

[0005] The tactile inspection method described in Patent Document 1 has room for improvement in terms of improving the accuracy of judgment.

[0006] The present disclosure provides a push button inspection method, an electronic device manufacturing method, and a program with improved determination accuracy. [Means for solving the problem]

[0007] A method for inspecting a push button according to one aspect of the present disclosure includes: A method for inspecting a push button, comprising the steps of: acquiring load waveform data indicating a relationship between a stroke position of the push button and the pressing load; determining whether the push button is normal or abnormal based on the load waveform data; a step of notifying the user that the push button is abnormal if the push button is abnormal; Including, The step of determining whether the push button is normal or abnormal includes: inputting the load waveform data as input waveform data into an autoencoder that has been previously trained using the load waveform data of a plurality of normal push buttons that have been previously acquired as training data, and acquiring output waveform data that is output from the autoencoder; determining whether the push button is normal or abnormal based on a difference between the input waveform data and the output waveform data; Includes.

[0008] A method for manufacturing an electronic device according to one aspect of the present disclosure, A method for manufacturing an electronic device, comprising: providing an input device including a plurality of push buttons; testing the push buttons of each of the input devices using the testing method described above; attaching the input device, the plurality of push buttons of which have been determined to be normal in the inspecting step, to the electronic device; Includes.

[0009] A program according to one aspect of the present disclosure includes: The above-mentioned inspection method is executed by a computer. Effect of the Invention

[0010] According to the present disclosure, it is possible to provide a push button inspection method, an electronic device manufacturing method, and a program with improved determination accuracy. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a block diagram showing an overview of an inspection system for carrying out an inspection method according to a first embodiment. [Diagram 2] Graph showing a sample load waveform [Diagram 3] FIG. 1 is a block diagram showing a schematic configuration of an autoencoder according to an embodiment of the present invention. [Figure 4] 1 is a flowchart for explaining an inspection method according to a first embodiment. [Figure 5A] Graph showing an example of input waveform data and output waveform data [Figure 5B] A graph showing the difference between the input waveform data value and the output waveform data value at each stroke position in the data of FIG. 5A. [Figure 6A] Graph showing another example of input waveform data and output waveform data [Figure 6B] A graph showing the difference between the input waveform data value and the output waveform data value at each stroke position in the data of FIG. 6A. [Figure 7A] Graph showing another example of input waveform data and output waveform data [Figure 7B] A graph showing the difference between the input waveform data value and the output waveform data value at each stroke position in the data of FIG. 7A. [Figure 8A] Graph showing another example of input waveform data and output waveform data [Figure 8B] A graph showing the difference between the values ​​of the input waveform data and the output waveform data at each stroke position in the data of FIG. 8A. [Figure 9A] Graph showing another example of input waveform data and output waveform data [Figure 9B] A graph showing the difference between the input waveform data value and the output waveform data value at each stroke position in the data of FIG. 9A. [Figure 10] FIG. 13 is a diagram showing an example of an electronic device manufactured by the manufacturing method according to the second embodiment. [Figure 11] A flowchart for explaining a method of manufacturing an electronic device according to a second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] (Background to this disclosure) Electronic devices such as PCs use input devices such as keyboards and mice. Keyboards and mice are equipped with push buttons that allow users to input operations to the electronic device by pressing them with their fingers. By inspecting the feel when pressing the push buttons on the keyboard or mouse, it is possible to determine whether the keyboard or mouse is good or bad.

[0013] Conventionally, to determine whether a keyboard or mouse is a good product or not, a sensory inspection is performed in which an inspector presses a push button to check for the presence or absence of abnormalities. Also, as in the inspection method of Patent Document 1, a tactile inspection method has been proposed in which data obtained when a switch is pressed and judged to be normal by a person is used as a standard, and the degree of deviation from the operating force at a stroke position with a characteristic tactile feel of the standard is used as a judgment index as a degree of coincidence.

[0014] However, conventional sensory inspection by inspectors depends on the skills of the inspectors, and there is a problem that the judgment results vary widely. In addition, the tactile inspection method described in Patent Document 1 has a problem that although the measurement result data falls within the range of the reference data values, it is difficult to detect data including anomalies that would result in defective products. As a result, there is a problem that it is not possible to improve the accuracy of judgment as to whether a product is good or not.

[0015] Therefore, the present inventor(s) devised a push button inspection method, an electronic device manufacturing method, and a program with improved judgment accuracy, and arrived at the following invention.

[0016] (Embodiment 1) [Overall configuration] FIG. 1 is a block diagram showing an outline of an inspection system 1 for carrying out an inspection method according to a first embodiment.

[0017] The inspection system 1 is a system for inspecting keyboard keys (hereinafter referred to as push buttons) used in a PC or the like, mouse buttons used in a PC or the like, or keys of an electronic musical instrument, etc. In this embodiment, inspecting a push button or the like means obtaining the load relative to the stroke position when the push button is pressed, and judging whether the push button is normal or abnormal based on the obtained load.

[0018] 1, the inspection system 1 includes, for example, an actuator 2, a sensor 3, a control unit 4, and a notification unit 5. The configuration of the inspection system 1 is not limited to these, and some components may be omitted, or additional components may be included.

[0019] The actuator 2 is used to press the push button during testing.

[0020] The sensor 3 detects the load when the push button is pressed by the actuator 2. In this embodiment, the load refers to the magnitude of the force required to press the push button. The magnitude of the force required to press the push button is hereinafter referred to as the "pressing load." The sensor 3 outputs the detected pressing load to the control unit 4 as analog data, or converts it into digital data by an A / D converter (not shown).

[0021] The control unit 4 acquires load waveform data based on the pressing load detected by the sensor 3, and determines whether the push button is normal or abnormal based on the load waveform data.

[0022] The control unit 4 can be realized by semiconductor elements, etc., and can be configured by various processors such as a microcomputer, a CPU, an MPU, a GPU, a DSP, an FPGA, an ASIC, etc. The control unit 4 can also realize predetermined functions by reading data and programs stored in a recording medium (not shown) and performing various arithmetic processing.

[0023] The control unit 4 includes, for example, a calculation unit 6 that acquires load waveform data, and an autoencoder 7 that receives the load waveform data acquired by the calculation unit 6 as input waveform data and outputs output waveform data.

[0024] The calculation unit 6 acquires load waveform data generated based on the pressing load detected by the sensor 3. When the pressing load from the sensor 3 is analog data, the calculation unit 6 acquires load waveform data by accepting the analog input and converting it to digital data by an A / D converter (not shown). FIG. 2 is a graph showing a sample of a load waveform. When a push button of a keyboard is pressed, the magnitude of the pressing load changes as the key sinks, i.e., as the stroke position becomes deeper. The load waveform shown in FIG. 2 is a graph showing the relationship between the stroke of the push button and the pressing load. The calculation unit 6 calculates load waveform data for generating a load waveform based on the load detected by the sensor 3.

[0025] As shown in Figure 2, the pressing load increases as the stroke deepens up to stroke position st1, and reaches a peak load. After that, from stroke position st1 to stroke position st2, the pressing load decreases as the stroke deepens, and reaches a return load at stroke position st2. As the stroke deepens further, the pressing load starts to rise again and continues to increase until it reaches the deepest position of the stroke (stroke position st3). The load waveform of a push button such as a keyboard key generally follows a curve like that shown in Figure 2.

[0026] As described above, the calculation unit 6 calculates the load waveform data represented as a graph shown in Fig. 2 based on the pressing load detected by the sensor 3. In this embodiment, the load waveform data refers to data showing the relationship between the stroke position of the push button and the pressing load at each stroke position. For ease of explanation, the load waveform data will be described using the load waveform shown in Fig. 2. Note that the load waveform data may be a set of discrete values ​​represented by a plurality of stroke positions and a load value for each stroke position.

[0027] Fig. 3 is a block diagram showing a schematic configuration of the autoencoder 7 according to the present embodiment. As shown in Fig. 3, when the load waveform data calculated by the calculation unit 6 is input as input waveform data, the autoencoder 7 outputs output waveform data.

[0028] The autoencoder 7 is previously trained with load waveform data of a plurality of normal push buttons acquired in advance as training data. That is, the autoencoder 7 is a machine learning model that is trained without a teacher using load waveform data of normal push buttons. For example, it is preferable to train the autoencoder 7 with load waveform data of about 200 normal push buttons in advance. In general, a large amount of data needs to be trained in order to build a machine learning model. The present inventors have confirmed that by using the autoencoder 7, even when training is performed with a very small amount of load waveform data, normality / abnormality can be accurately determined as described below.

[0029] The input waveform data input to the autoencoder 7 is composed of a multidimensional column vector arranged in a time series. The autoencoder 7 includes an encoder 71 and a decoder 72. The encoder 71 and the decoder 72 are each composed of a neural network having a plurality of intermediate layers. In the encoder 71, the number of output nodes is smaller than the number of input nodes. Therefore, the encoder 71 performs dimensional compression of the input waveform data to generate compressed data. In the decoder 72, the number of output nodes is larger than the number of input nodes, and the number of output nodes is the same as the number of dimensions of the column vector of the input waveform data. Therefore, when the decoder 72 receives the compressed data output from the encoder 71, it performs dimensional restoration to generate output waveform data with the same number of dimensions as the input waveform data. In the neural networks of the encoder 71 and the decoder 72, learning is performed so that the output waveform data outputs the same value as the input waveform data.

[0030] When normal input waveform data is input to the trained autoencoder 7, normal output waveform data is output from the autoencoder 7. That is, the input waveform data and output waveform data for the autoencoder 7 match. On the other hand, even when abnormal (abnormal) input waveform data is input to the autoencoder 7, normal output waveform data is output from the autoencoder 7. Since only normal input waveform data is used when the autoencoder 7 is trained, when abnormal input waveform data is input, output waveform data that has been changed to normal waveform data is output. That is, if there is an abnormality in the input waveform data, an error occurs between the input waveform data and the output waveform data.

[0031] The calculation unit 6 compares the input waveform data to the autoencoder 7 with the output waveform data from the autoencoder 7, and calculates the difference between the value of the input waveform data and the value of the output waveform data at each stroke position included in the waveform data. If the absolute value of the difference between the value of the input waveform data and the value of the output waveform data at any stroke position is greater than a predetermined threshold, the calculation unit 6 can determine that the input load waveform data is abnormal. This is because, if the absolute value of the difference between the value of the input waveform data and the value of the output waveform data is greater than a predetermined threshold, a deviation occurs between the input waveform data and the output waveform data, and therefore it can be estimated that the input load waveform data was abnormal.

[0032] The control unit 4 outputs the determination result of whether the push button is normal or abnormal to the notification unit 5. For example, if the push button is determined to be abnormal, the notification unit 5 can notify the user of the abnormality.

[0033] The notification unit 5 may be, for example, a device capable of displaying an image such as a liquid crystal display or an organic EL display, or may be a speaker capable of notifying the user of an abnormality by voice.

[0034] [Inspection method] Next, a description will be given of a method for inspecting a push button in the inspection system 1. Fig. 4 is a flowchart for explaining the inspection method according to the first embodiment.

[0035] First, in step S11, load waveform data is acquired by the control unit 4. The pressing load when the push button to be inspected is pressed is detected by the sensor 3, and the load waveform data is acquired by the calculation unit 6 of the control unit 4 based on the detected pressing load. The calculation unit 6 may acquire the load waveform data based on the pressing load of the push button measured in advance, or may acquire the load waveform data using the pressing load measured in real time.

[0036] In step S12, the calculation unit 6 inputs the acquired load waveform data as input waveform data to the autoencoder 7. In step S13, the autoencoder 7 generates and outputs output waveform data by compressing and restoring the input waveform data. In step S14, the calculation unit 6 calculates the difference between the input waveform data and the output waveform data at each stroke position included in the waveform data, and calculates the difference between the value of the input waveform data and the value of the output waveform data at each stroke position.

[0037] FIG. 5A is a graph showing an example of input waveform data and output waveform data. In FIG. 5A, the curve shown by the solid line is the input waveform data, and the curve shown by the dashed line is the output waveform data. In the example of FIG. 5A, the curve shown by the input waveform data and the curve shown by the output waveform data are almost the same. FIG. 5B is a graph showing the difference between the value of the input waveform data and the value of the output waveform data at each stroke position in the data of FIG. 5A. The dashed line in FIG. 5B indicates a predetermined threshold value, and in the example of FIG. 5B, the predetermined threshold value is 0.05N.

[0038] In step S14, the calculation unit 6 judges whether the push button to be inspected is normal or abnormal based on the difference between the input waveform data and the output waveform data at each stroke position of the push button. In this embodiment, the difference between the input waveform data and the output waveform data indicates the absolute value of the difference between the value of the input waveform data and the value of the output waveform data. More specifically, in this embodiment, the calculation unit 6 calculates the absolute value of the difference between the value of the input waveform data and the value of the output waveform data at each scroll position of the push button. The calculation unit 6 judges the push button to be abnormal if the absolute value of the difference exceeds a predetermined threshold, and judges the push button to be normal if the absolute value of the difference is equal to or less than the predetermined threshold.

[0039] As described above, in the example of Fig. 5A, the curve indicated by the input waveform data and the curve indicated by the output waveform data are roughly the same. When input waveform data such as that shown in Fig. 5A is shown, the absolute value of the difference is equal to or less than a predetermined threshold value (0.05N) at all stroke positions, as shown in Fig. 5B. Therefore, the calculation unit 6 determines that the push button indicating this waveform data is normal.

[0040] FIG. 6A is a graph showing another example of input waveform data and output waveform data. FIG. 6B is a graph showing the difference between the value of the input waveform data and the value of the output waveform data at each stroke position in the data of FIG. 6A. In FIG. 6A, there is a portion where the input waveform data and the output waveform data are separated from each other between the stroke position of about 1.2 mm and about 1.7 mm. That is, there is a portion where the input waveform data and the output waveform data do not clearly overlap between the stroke position of about 1.2 mm and about 1.7 mm. As shown in FIG. 6B, the absolute value of the difference between the value of the input waveform data and the value of the output waveform data is a maximum of about 0.07 N between the stroke position of about 1.2 mm and about 1.7 mm, which exceeds the threshold value of 0.05 mm. Therefore, the calculation unit 6 determines that the push button showing this waveform data is abnormal.

[0041] 6A, an abnormal sound occurs when the push button is pressed. When inspecting such a defect in a noisy factory or the like, it is difficult for an inspector to determine that the button is abnormal. However, when the push button is inspected using the inspection method of this embodiment, it can be determined that the button is abnormal with high accuracy.

[0042] FIG. 7A is a graph showing another example of input waveform data and output waveform data. FIG. 7B is a graph showing the difference between the value of the input waveform data and the value of the output waveform data at each stroke position in the data of FIG. 7A. In FIG. 7A, the stroke position is about 0.3 mm, and there is a protruding part in the input waveform data. As shown in FIG. 7B, when the stroke position is about 0.3 mm, the absolute value of the difference between the value of the input waveform data and the value of the output waveform data is about 0.06 N, which exceeds the threshold value of 0.05 mm. Therefore, the calculation unit 6 determines that the push button showing this waveform data is abnormal.

[0043] When input waveform data such as that shown in FIG. 7A is displayed, there is a sense of incongruity when pressing the push button. It is difficult for an inspector to determine that such a defect is abnormal during continuous work. When the push button is inspected using the inspection method of this embodiment, it can be accurately determined that it is abnormal.

[0044] FIG. 8A is a graph showing another example of input waveform data and output waveform data. FIG. 8B is a graph showing the difference between the value of the input waveform data and the value of the output waveform data at each stroke position in the data of FIG. 8A. In FIG. 8A, when only the input waveform data is viewed, it is close to normal waveform data, but when the input waveform data and the output waveform data are compared, it can be seen that there are many parts where the input waveform data and the output waveform data do not overlap, and the difference is large over the entire stroke. As shown in FIG. 8B, the absolute value of the difference between the value of the input waveform data and the value of the output waveform data exceeds the threshold value of 0.05 N over the entire stroke, and is 0.20 N at the maximum. Therefore, the calculation unit 6 determines that the push button showing this waveform data is abnormal.

[0045] When input waveform data such as that shown in FIG. 8A is displayed, the drop load of the push button (see FIG. 2) is smaller than that of a normal push button, and the push button feels heavier when pressed than a normal push button.

[0046] FIG. 9A is a graph showing another example of input waveform data and output waveform data. FIG. 9B is a graph showing the difference between the value of the input waveform data and the value of the output waveform data at each stroke position in the data of FIG. 9A. In FIG. 9A, the stroke position is about 1.0 mm, and there is a protruding portion in the input waveform data. Also, in FIG. 9A, there is a large deviation between the input waveform data and the output waveform data. As shown in FIG. 9B, there are portions in which the absolute value of the difference between the value of the input waveform data and the value of the output waveform data exceeds the threshold value of 0.05 N over the entire stroke, and in particular, at a stroke position of about 1.0 mm, the absolute value of the difference is about 0.3 N, which is larger than other portions. Therefore, the calculation unit 6 determines that the push button showing this waveform data is abnormal.

[0047] As described above, in step S14 of FIG. 4, the calculation unit 6 determines whether the push button is normal or abnormal, based on the input waveform data and output waveform data shown in FIGS. 5A to 9A.

[0048] If the calculation unit 6 determines in step S14 that the push button is normal, the inspection process ends. If the calculation unit 6 determines in step S14 that the push button is abnormal, the process proceeds to step S15, where the calculation unit 6 outputs a notification of the abnormality to the notification unit 5, and the inspection process ends.

[0049] [effect] According to the above-mentioned embodiment, it is possible to provide a method for inspecting a push button with improved judgment accuracy. In the above-mentioned embodiment, the autoencoder is trained in advance using the load waveform data of a normal push button. When the trained autoencoder receives the load waveform data of a normal push button, it outputs the input waveform data as is, and when the trained autoencoder receives the load waveform data of an abnormal push button, it outputs the complemented waveform data so as to become the normal waveform data. Therefore, when the difference between the input waveform data and the output waveform data to the autoencoder exceeds a predetermined threshold, it is possible to judge that the push button to be inspected is abnormal. Since the autoencoder is trained with the load waveform data of a plurality of normal push buttons, it is possible to make a judgment taking into consideration the variation in the pressing load seen in the load waveform data of a normal push button. Therefore, it is possible to improve the judgment accuracy.

[0050] In the above embodiment, an example of determining whether a push button (key) of a keyboard used in an electronic device or the like is normal or abnormal has been described, but the present invention is not limited to this. For example, the inspection method of the present embodiment can also be applied to mouse push buttons used in an electronic device or keys of a keyboard instrument.

[0051] In the above embodiment, the load waveform data is an example of data showing the relationship between the stroke position of the push button and the pressing load at each stroke position, but the present invention is not limited to this. For example, it is also possible to determine whether the push button is normal or abnormal by using analog waveform data showing the relationship between the pressing load and the stroke position.

[0052] In the above-described embodiment, the absolute value of the difference between the pressing load of the input waveform data and the pressing load of the output waveform data at each stroke position is calculated, and if the absolute value of the difference exceeds a predetermined threshold value, the push button is determined to be abnormal. However, the present invention is not limited to this. For example, it may be determined whether the push button is normal or abnormal based on the correlation between a function indicating the input waveform data and a function indicating the output waveform data. Also, instead of comparing the difference for each stroke position with a threshold value, it may be determined that the push button is abnormal if the sum of the differences for each stroke position or the sum of the squares of the differences for each stroke position exceeds a predetermined threshold value.

[0053] In the above embodiment, the load waveform data of a normal push button is learned by the autoencoder, but the present invention is not limited to this. A model learned by a machine learning algorithm other than the autoencoder may be used.

[0054] In the above embodiment, an example has been described in which when a push button is determined to be abnormal, the notification unit 5 notifies the abnormality, but the present invention is not limited to this. For example, when the push button is normal, the notification unit 5 may be notified that the push button is normal, and when the push button is abnormal, the notification unit 5 may be notified that the push button is abnormal.

[0055] The above-described embodiment can also be applied to a program, and can also be recorded on a recording medium as a program.

[0056] (Embodiment 2) A second embodiment will be described with reference to Figures 10 and 11. In the second embodiment, the same or equivalent configurations as those in the first embodiment will be denoted by the same reference numerals. In the second embodiment, descriptions that overlap with those in the first embodiment will be omitted.

[0057] In the second embodiment, a method for manufacturing an electronic device including an input device inspected by the inspection method described in the first embodiment will be described.

[0058] Fig. 10 is a diagram showing an example of an electronic device manufactured by the manufacturing method of embodiment 2. Fig. 11 is a flowchart for explaining the manufacturing method of the electronic device according to embodiment 2.

[0059] In the second embodiment, a method for manufacturing an electronic device 10 having an input device shown in FIG. 10 will be described. In the example of FIG. 10, the input device is a keyboard 20 including a plurality of push buttons (keys). The input device is not limited to the keyboard 20, and may be any device having one or a plurality of push buttons. An example of a device having a push button is a keyboard panel of a PC. Alternatively, an example of a device having a push button is a home appliance having a push button such as a television or a robot vacuum cleaner, or a power strip having a push button. Alternatively, the input device may be a pointing device 30 including a push button or a keyboard of an electronic musical instrument. The electronic device 10 is, for example, a notebook PC. In this embodiment, the push buttons included in the input device are judged to be normal or abnormal by the inspection method described in the first embodiment, and an electronic device 10 having an input device in which all the push buttons are normal is manufactured.

[0060] A method for manufacturing the electronic device 10 will be described with reference to Fig. 11. In the following description, an example in which the input device is a keyboard 20 will be described.

[0061] In step S21, the keyboard 20 is prepared. The keyboard 20 includes one or more push buttons (keys).

[0062] Next, in step S22, all the push buttons included in the keyboard 20 prepared in step S21 are inspected using the inspection method described in the first embodiment.

[0063] If all the push buttons included in the keyboard 20 inspected in step S22 are normal, the process proceeds to step S24, where the keyboard 20 is attached to the electronic device 10, and the electronic device is completed.

[0064] If the keyboard 20 inspected in step S22 includes a push button determined to be abnormal, the keyboard 20 is not attached to the electronic device 10 as a defective product, and the process returns to step S21 and another keyboard 20 is prepared.

[0065] [effect] According to the above-described embodiment, it is possible to provide a manufacturing method for electronic devices with improved push button inspection accuracy. In addition, since all push buttons of the keyboard 20 can be inspected in a short time compared to manual inspection by an inspector, the time required for manufacturing can be shortened.

[0066] In the above embodiment, the electronic device 10 is a notebook computer equipped with an input device, but the present invention is not limited to this. The electronic device 10 may be any device that can detect the on and off of a push button and output a signal in response to the on and off of the push button. More specifically, the electronic device 10 may include a device itself, such as a keyboard, a mouse, or an electronic musical instrument, that includes a board on which a microcomputer is mounted.

[0067] (Outline of the embodiment) (1) A method of inspecting a push button according to one aspect of the present disclosure includes the steps of acquiring load waveform data indicating a relationship between a stroke position of the push button and a pressing load, determining whether the push button is normal or abnormal based on the load waveform data, and notifying the user if the push button is abnormal. The step of determining whether the push button is normal or abnormal includes inputting the load waveform data as input waveform data into an autoencoder that has been trained in advance using load waveform data of a plurality of normal push buttons previously acquired as learning data, acquiring output waveform data output from the autoencoder, and determining whether the push button is normal or abnormal based on a difference between the input waveform data and the output waveform data.

[0068] (2) In the inspection method of (1), the step of determining whether the push button is normal or abnormal may include calculating the absolute value of the difference between the input waveform data and the output waveform data, determining that the push button is abnormal if the absolute value of the difference exceeds a predetermined threshold, and determining that the push button is normal if the absolute value of the difference is equal to or less than the predetermined threshold.

[0069] (3) In the inspection method of (1) or (2), the difference between the input waveform data and the output waveform data is the difference between the value of the input waveform data and the value of the output waveform data at each stroke position of the push button, and the step of determining whether the push button is normal or abnormal may include determining that the push button is abnormal if the absolute value of the difference at at least one stroke position exceeds a predetermined threshold value.

[0070] (4) A method for manufacturing an electronic device according to one aspect of the present disclosure includes the steps of: preparing an input device including one or more push buttons; inspecting each push button of the input device using any one of the inspection methods (1) to (3); and attaching the input device, the one or more push buttons of which are determined to be normal in the inspection step, to the electronic device.

[0071] (5) In the method for manufacturing an electronic device according to (4), the input device may be a keyboard.

[0072] (6) In the method for manufacturing an electronic device according to (4) or (5), the electronic device may be a notebook computer.

[0073] (7) A program according to one aspect of the present disclosure causes a computer to execute any one of the inspection methods (1) to (3). [Industrial Applicability]

[0074] The present disclosure can be applied to testing devices that include push buttons. [Explanation of symbols]

[0075] 1 Inspection system 2 Actuators 3 Sensors 4. Control section 5 Notification section 6 Arithmetic section 7. Autoencoders 10 Electronic equipment 20 Keyboard (input device) 30 Pointing Device 71 Encoder 72 Decoder

Claims

1. A method for inspecting a push button, comprising the steps of: acquiring load waveform data indicating a relationship between a stroke position of the push button and a pressing load; determining whether the push button is normal or abnormal based on the load waveform data; a step of notifying the user that the push button is abnormal if the push button is abnormal; Including, The step of determining whether the push button is normal or abnormal includes: inputting the load waveform data as input waveform data into an autoencoder that has been previously trained using the load waveform data of a plurality of normal push buttons that have been previously acquired as training data, and acquiring output waveform data that is output from the autoencoder; determining whether the push button is normal or abnormal based on a difference between the input waveform data and the output waveform data; Including, Testing method.

2. The step of determining whether the push button is normal or abnormal includes: calculating an absolute value of a difference between the input waveform data and the output waveform data, and judging that the push button is abnormal if the absolute value of the difference exceeds a predetermined threshold, and judging that the push button is normal if the absolute value of the difference is equal to or less than the predetermined threshold; Including, The inspection method according to claim 1 .

3. the difference between the input waveform data and the output waveform data is a difference between a value of the input waveform data and a value of the output waveform data at each stroke position of the push button, the step of determining whether the push button is normal or abnormal includes determining that the push button is abnormal when an absolute value of the difference in at least one stroke position exceeds the predetermined threshold value; The inspection method according to claim 2.

4. A method for manufacturing an electronic device, comprising: Providing an input device including one or more push buttons; testing each of the push buttons of the input device using a testing method according to any one of claims 1 to 3; attaching the input device, the one or more push buttons of which have been determined to be normal in the inspecting step, to the electronic device; Including, Manufacturing method of electronic devices.

5. The input device is configured by a keyboard. The method for manufacturing an electronic device according to claim 4 .

6. The electronic device is a notebook computer. The method for manufacturing an electronic device according to claim 4 .

7. A program for causing a computer to execute the inspection method according to any one of claims 1 to 3.

Images