Object hardness recognition system and method and touch sensor preparation method

Abstract

The invention discloses an object hardness identification system and method and a touch sensor preparation method. According to the technical scheme, the machine learning network model is preset in the machine learning calculation unit in advance, the touch sensor outputs the voltage signal when making contact with the object during object hardness recognition, the voltage signal is collected through the signal collection unit, and the detection signal is output according to the voltage signal; and the machine learning calculation unit receives the detection signal, trains the detection signal based on a machine learning network model and then outputs recognized object hardness data, recognition of the object hardness is completed through the touch sensor, and the new application of the touch sensor is further explored.

Description

Technical Field

[0001] The present invention relates to the field of sensor technology, and in particular to an object hardness recognition system and method, and a tactile sensor preparation method. Background Art

[0002] The rapid development of artificial intelligence has further stimulated the rapid growth of demand for various types of sensors, for example, in identifying the hardness of objects.

[0003] Currently, when identifying the hardness of an object, the deformation of the object is generally assessed by measuring parameters such as pressure, indentation displacement, and tactile images when interacting with the object. Although the above method can be used to identify the hardness of an object, identifying the hardness of an object through the response time of a tactile sensor has not yet been discovered.

[0004] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention

[0005] In view of the above-mentioned deficiencies in the prior art, the object of the present invention is to provide an object hardness identification system, method and tactile sensor preparation method to solve the problem that the prior art can identify the hardness of an object, but the hardness of an object identified by a tactile sensor has not yet been discovered.

[0006] The technical solution adopted by the present invention to solve the technical problem is to provide an object hardness recognition system, comprising:

[0007] a tactile sensor for outputting a voltage signal when in contact with an object whose hardness needs to be identified;

[0008] a signal acquisition unit connected to the tactile sensor, configured to receive the voltage signal and output a detection signal according to the voltage signal; and

[0009] a machine learning computing unit connected to the signal acquisition unit, wherein the machine learning computing unit is preset with a machine learning network model, and is configured to receive the detection signal and output recognized object hardness data after training the detection signal in a machine learning network based on the machine learning network model;

[0010] The machine learning network model is a data set consisting of the detection signal and the corresponding object hardness when objects of different hardness are applied to the tactile sensor.

[0011] The present invention further provides that the tactile sensor comprises: a second conductive layer, a second friction layer, a first friction layer, a first conductive layer and a protrusion which are sequentially stacked;

[0012] The material of the second friction layer and the material of the first friction layer are triboelectric materials with opposite electrical properties, and the first friction layer and the second friction layer are bonded to each other;

[0013] The protrusion is used to contact an object whose hardness needs to be identified and transmit force to the first friction layer, so that the first friction layer and the second friction layer generate a triboelectric effect under the action of an external force.

[0014] The present invention further provides that the protrusion is a hemispherical protrusion structure.

[0015] The present invention further provides that the material of the protrusion is made of flame-retardant glass fiber reinforced epoxy resin material.

[0016] The present invention further provides that the upper surface of the first conductive layer and the lower surface of the first conductive layer are respectively covered with copper glue.

[0017] The present invention further provides that the upper surface of the second conductive layer is covered with copper glue.

[0018] The present invention further provides that the first friction layer includes: a polydimethylsiloxane film with a sandpaper surface microstructure.

[0019] The present invention further provides that the second friction layer includes: a polytetrafluoroethylene film with a smooth surface.

[0020] The present invention also provides a method for using the object hardness identification system as described above, wherein the object hardness identification system includes a tactile sensor, wherein the tactile sensor is configured to output a voltage signal when in contact with an object, and the object hardness identification system method includes:

[0021] receiving the voltage signal and outputting a detection signal according to the voltage signal;

[0022] The detection signal is received and the object hardness data is output after training the detection signal based on a machine learning network model, wherein the machine learning network model is a data set consisting of the detection signal and the corresponding object hardness when objects with different hardness are sensed by the tactile sensor.

[0023] The present invention also provides a method for preparing the tactile sensor in the object hardness recognition system as described above, comprising:

[0024] Providing a protrusion, a first conductive layer, a first friction layer, a second friction layer, and a second conductive layer;

[0025] Adhere the first friction layer to the lower surface of the first conductive layer;

[0026] adhering the second friction layer to the upper surface of the second conductive layer;

[0027] Adhere the protrusion to the upper surface of the first conductive layer;

[0028] The first friction layer and the second friction layer are bonded to each other.

[0029] The beneficial effects of the present invention are:

[0030] The present invention discloses an object hardness identification system, method and tactile sensor preparation method. The object hardness identification system includes: a tactile sensor, which is used to output a voltage signal when in contact with an object whose hardness needs to be identified; a signal acquisition unit connected to the tactile sensor, which is used to receive the voltage signal and output a detection signal based on the voltage signal; and a machine learning calculation unit connected to the signal acquisition unit, in which a machine learning network model is preset. The machine learning calculation unit is used to receive the detection signal and output the identified object hardness data after training the detection signal in the machine learning network based on the machine learning network model; wherein the machine learning network model is a data set consisting of the detection signal and the corresponding object hardness when force is applied to the tactile sensor of objects of different hardness. In the technical solution of the present invention, a machine learning network model is preset in a machine learning calculation unit. When identifying the hardness of an object, the sensitivity of the tactile sensor to dynamic excitation is utilized to make the tactile sensor output a voltage signal when in contact with the object. The voltage signal is further collected by the signal acquisition unit and a detection signal is output based on the voltage signal. The machine learning calculation unit receives the detection signal and outputs the identified object hardness data after training the detection signal based on the machine learning network model. When the tactile sensor in the object hardness identification system contacts objects of different hardness, there is a significant difference in the response time of the voltage signal output by the tactile sensor under the same mechanical stimulation. Therefore, the hardness of the object can be distinguished by establishing a quantitative relationship between the response time of the tactile sensor and the hardness of the object. The identification of the object hardness is completed through the tactile sensor, and a new use of the tactile sensor is further explored. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary personnel in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0032] Figure 1 It is a principle block diagram of the object hardness recognition system of the present invention.

[0033] Figure 2 It is the machine learning network structure in the present invention.

[0034] Figure 3 It is a flow chart of the object hardness recognition method based on machine learning algorithm of the present invention.

[0035] Figure 4 The invention relates to a method for preparing a tactile sensor.

[0036] Figure 5 This is the waveform output by the tactile sensor when pressing samples of different hardness in one embodiment of the present invention.

[0037] Figure 6 This is the relationship between the response time, peak voltage and hardness of the tactile sensor in one embodiment of the present invention.

[0038] The marks in the accompanying drawings are: 1. protrusion; 2. first conductive layer; 3. first friction layer; 4. second friction layer; 5. second conductive layer; 6. wire; 7. signal acquisition unit; 8. machine learning calculation unit. DETAILED DESCRIPTION

[0039] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, the specific embodiments of the present invention are now described in detail with reference to the accompanying drawings. In the following description, it should be understood that the directions or positional relationships indicated by "front", "back", "up", "down", "left", "right", "longitudinal", "horizontal", "vertical", "horizontal", "top", "bottom", "inside", "outside", "head", "tail", etc. are based on the directions or positional relationships shown in the accompanying drawings and are constructed and operated in specific directions. They are only for the convenience of describing the technical solution and do not indicate that the devices or components referred to must have specific directions. Therefore, they should not be understood as limiting the present invention.

[0040] The vigorous development of artificial intelligence has led to a rapid increase in demand for various types of sensors. It is particularly noteworthy that tactile sensors are becoming increasingly prominent in artificial intelligence systems. They can perceive the properties of objects with high precision and high sensitivity. They can not only perceive the shape and surface features of objects, but also penetrate deep into the interior of objects to capture rich information such as the hardness and temperature of objects, thereby enabling robots and other intelligent devices to interact with the surrounding environment more flexibly, thereby achieving more sophisticated and intelligent human-computer interaction. Among them, hardness measurement is an important direction for the development of tactile sensors. It represents the exploration of deeper interactions between intelligent systems and the surrounding environment and is an important step in technological progress. At present, when identifying the hardness of an object, the deformation of the object is generally evaluated by measuring parameters such as pressure, indentation displacement, and tactile images when interacting with the object. There is still room for discovery of new parameters to identify the hardness of an object.

[0041] In view of the problems of the prior art, the present invention provides an object hardness recognition system based on the sensitivity of the tactile sensor to dynamic excitation. Figure 1 、 Figure 2 As shown, the object hardness identification system may include a tactile sensor, a signal acquisition unit 7, and a machine learning calculation unit 8; the tactile sensor is configured to output a voltage signal when in contact with an object whose hardness needs to be identified; the signal acquisition unit 7 is connected to the tactile sensor, configured to receive the voltage signal and output a detection signal based on the voltage signal; the machine learning calculation unit 8 is connected to the signal acquisition unit 7, and the machine learning calculation unit 8 is preset with a machine learning network model. The machine learning calculation unit is configured to receive the detection signal and output the identified object hardness data after training the detection signal in the machine learning network based on the machine learning network model; wherein the machine learning network model is a data set consisting of the detection signal and the corresponding object hardness when objects of different hardness apply force to the tactile sensor, that is, the detection signal output by the signal acquisition unit 7 and the corresponding object hardness when objects of different hardness apply force to the tactile sensor. The data set consisting of the detection signal and the corresponding object hardness means that when an object of a certain hardness applies force to the tactile sensor, the tactile sensor outputs a voltage signal, and the signal acquisition unit 7 collects the voltage signal and outputs the detection signal. At this time, the hardness of the force-applying object is known, and the hardness forms a corresponding relationship with the detection signal.

[0042] In this embodiment, a tactile sensor is used to measure objects of various hardnesses, and the voltage signal output by the tactile sensor is collected by the signal acquisition unit 7 to output a detection signal. The detection signals corresponding to objects of different hardnesses are collected and formed into a data set to form a machine learning network model. When identifying the hardness of an object, the sensitivity of the tactile sensor to dynamic excitation is utilized. When the tactile sensor contacts an object whose hardness needs to be identified, a voltage signal is output. The voltage signal is further collected by the signal acquisition unit 7 and a detection signal is output based on the voltage signal. The machine learning calculation unit 8 receives the detection signal, and outputs the identified object hardness data after training the detection signal in the machine learning network based on the machine learning network model, thereby completing the identification of the object hardness.

[0043] Based on the above discussion, when the tactile sensor in the object hardness identification system contacts objects of different hardness, there is a significant difference in the response time of the voltage signal output by the tactile sensor under the same mechanical stimulation. Therefore, the hardness of the object can be distinguished by establishing a quantitative relationship between the response time of the tactile sensor and the hardness of the object. The hardness of the object is identified through the tactile sensor, further exploring new uses for the tactile sensor.

[0044] Here, it should be noted that the signal output by the tactile sensor is an analog signal, which is collected by the signal acquisition unit 7 and converted into a digital signal for output. The signal acquisition unit 7 can be any digital-to-analog conversion circuit that can collect analog signals and convert analog signals into digital signals for output. Those skilled in the art can determine the specific circuit structure of the signal acquisition unit 7 according to actual needs, and will not go into details here.

[0045] In some embodiments, the tactile sensor may include a second conductive layer 5, a second friction layer 4, a first friction layer 3, a first conductive layer 2 and a protrusion 1 that are stacked in sequence; the material of the second friction layer 4 and the material of the first friction layer 3 are triboelectric materials with opposite electrical properties, and the first friction layer 3 and the second friction layer 4 are in contact with each other; the protrusion 1 is used to contact the force-applying object and transmit force to the first friction layer 3, so that the first friction layer 3 and the second friction layer 4 generate a triboelectric effect under the action of external force.

[0046] Specifically, the upper surface of the first friction layer 3 can be pasted to the lower surface of the first conductive layer 2. Similarly, the lower surface of the second friction layer 4 can be pasted to the upper surface of the second conductive layer 5. Thereafter, the first friction layer 3 and the second friction layer 4 are directly bonded to each other. Further, the protrusion 1 is pasted to the upper surface of the first conductive layer 2.

[0047] In this embodiment, if Figure 1 As shown, when the tactile sensor is connected to the signal acquisition module, the first conductive layer 2 can be led out through a wire 6 and connected to the signal acquisition unit 7, and the second conductive layer 5 can also be led out through a wire 6 and connected to the signal acquisition unit 7.

[0048] Here, it should be noted that when the first friction layer 3 and the second friction layer 4 are bonded to each other, it should be noted that after the first friction layer 3 is subjected to force, a triboelectric effect can be generated between the bonded first friction layer 3 and the second friction layer 4.

[0049] In some embodiments, as Figure 1 As shown, the protrusion 1 may be a hemispherical protrusion 1 structure.

[0050] Specifically, the first conductive layer 2 may be in a plate-like structure but is not limited thereto. The protrusion 1 may be arranged at the center of the first conductive layer 2, such as Figure 1 As shown, that is, the center point of the protrusion 1 and the center point of the first conductive layer 2 are located on the same axis, so as to ensure that when the protrusion 1 contacts the object whose hardness needs to be identified, the force generated by the protrusion 1 when contacting the object whose hardness needs to be identified can be evenly transmitted to the first conductive layer 2.

[0051] In this embodiment, the protrusion 1 adopts a hemispherical design, which can make the contact area smaller and the stress more concentrated when the protrusion 1 contacts the object whose hardness needs to be identified, so that the obtained triboelectric signal is more stable, and there is a significant difference in identifying objects of different hardness.

[0052] In some embodiments, the protrusion 1 may be made of, but is not limited to, flame-retardant glass fiber reinforced epoxy resin material.

[0053] In this embodiment, the flame-retardant glass fiber reinforced epoxy resin material has high hardness and can effectively transmit the force exerted on the protrusion 1 by the object whose hardness needs to be identified, so that the voltage signal output by the tactile sensor is more accurate, that is, the force loss is small during the force transmission process.

[0054] In some embodiments, the upper surface of the first conductive layer 2 and the lower surface of the first conductive layer 2 are respectively covered with copper paste.

[0055] Specifically, when the first conductive layer 2 is set, the upper surface and the lower surface of the first conductive layer 2 are coated with copper glue, so that it has stickiness on both sides while ensuring the conductive performance, so as to facilitate the bonding of the protrusion 1 and the first friction layer 3; wherein, the material of the first conductive layer 2 can be copper, of course, it can also be a conductive metal such as aluminum, silver, etc., preferably copper.

[0056] In some embodiments, the upper surface of the second conductive layer 5 is covered with copper paste.

[0057] Specifically, when the second conductive layer 5 is provided, the upper surface of the second conductive layer 5 can also be coated with copper glue, so that it has stickiness while ensuring the conductive performance, so as to facilitate the bonding of the second friction layer 4; similarly, the material of the second conductive layer 5 can be copper, of course, it can also be a conductive metal such as aluminum, silver, etc., preferably copper.

[0058] In some embodiments, the first friction layer 3 may include a polydimethylsiloxane film with a sandpaper surface microstructure.

[0059] In this embodiment, 800-mesh sandpaper can be used as a template, and the polydimethylsiloxane solution can be coated on the sandpaper. After film formation, the film is separated from the sandpaper to obtain the first friction layer 3 (polydimethylsiloxane film with a surface microstructure); of course, sandpaper with other mesh sizes can also be used as a template to make a polydimethylsiloxane film with a surface microstructure, for example, 600-mesh sandpaper can be used as a template, and for example, 1000-mesh sandpaper can be used as a template.

[0060] In some embodiments, the second friction layer 4 may include a polytetrafluoroethylene film with a smooth surface.

[0061] In this embodiment, the polytetrafluoroethylene film with a smooth surface is a polytetrafluoroethylene film produced using existing technology.

[0062] It should be noted that, when selecting the materials for the first friction layer 3 and the second friction layer 4, it is necessary to consider that the materials for the first friction layer 3 and the second friction layer 4 are triboelectric materials with opposite electrical properties, and that charge transfer is easily generated when the first friction layer 3 and the second friction layer 4 come into contact with each other. Furthermore, the mechanical properties, flexibility, rigidity, and processability of the friction layers (the first friction layer 3 and the second friction layer 4) must also be considered. In this embodiment, based on existing statistical triboelectric sequences, the material for the first friction layer 3 is preferably polydimethylsiloxane film (polydimethylsiloxane film, as a type of polymer organic silicon compound, has the advantages of good elasticity, simple production, durability, and strong chemical inertness. Furthermore, polydimethylsiloxane has good plasticity and can be surface patterned through a relatively simple process). The material for the second friction layer 4 is preferably polytetrafluoroethylene film relative to the material for the first friction layer 3 (polytetrafluoroethylene film contains more fluorine elements and has a stronger ability to capture electrons than other materials. Therefore, it is generally used as the negative friction layer of the triboelectric device, i.e., the second friction layer 4 of the tactile sensor).

[0063] In this embodiment, the first friction layer 3 and the second friction layer 4 have different affinities for electrons. After the first friction layer 3 and the second friction layer 4 come into contact under the action of an external force, due to the triboelectric effect, after the first conductive layer 2 and the second conductive layer 5 are connected via a load, charges are transferred between the first conductive layer 2, the load, and the second conductive layer 5, thereby generating a voltage signal.

[0064] In addition, since the surface of the first friction layer 3 has a sandpaper surface microstructure, the contact area between the first friction layer 3 and the second friction layer 4 can be effectively expanded during the process of identifying the hardness of the object, thereby improving the triboelectric performance and lowering the detection threshold of the tactile sensor.

[0065] In some embodiments, as Figure 3 As shown, the present invention also provides a method for applying the object hardness identification system as described above, wherein the object hardness identification system includes a tactile sensor, which is configured to output a voltage signal when in contact with an object whose hardness needs to be identified. The object hardness identification system method includes the following steps:

[0066] S100, receiving a voltage signal and outputting a detection signal according to the voltage signal;

[0067] S200, receiving a detection signal and outputting object hardness data after training a machine learning network model based on the detection signal, wherein the machine learning network model is a data set consisting of the detection signal and the object hardness when force is applied to the tactile sensor for objects of different hardness.

[0068] In this embodiment, the tactile sensor is connected to the signal acquisition unit through a wire, the voltage signal output by the tactile sensor is collected by the signal acquisition unit, and a detection signal is output according to the voltage signal; the signal acquisition unit is connected to the machine learning calculation unit, wherein, when the machine learning calculation unit is connected to the signal acquisition unit, the machine learning calculation unit can be integrated into a terminal device capable of human-computer interaction, such as a computer device, a tablet, a laptop computer, etc., and then the signal acquisition unit is communicatively connected to the terminal device capable of human-computer interaction, that is, the signal acquisition unit is connected to the machine learning calculation unit.

[0069] When identifying the hardness of an object, the sensitivity of the tactile sensor to dynamic excitation is utilized, and a voltage signal is output when the tactile sensor contacts the object whose hardness needs to be identified. The voltage signal is further acquired by the signal acquisition unit and a detection signal is output based on the voltage signal. The machine learning calculation unit receives the detection signal and outputs the identified object hardness data after training the detection signal in the machine learning network based on the machine learning network model, thereby completing the identification of the object hardness. The machine learning network model is specifically the machine learning network model described in the object hardness identification system, and will not be repeated here.

[0070] When the tactile sensor in this object hardness recognition system contacts objects of different hardness, there is a significant difference in the response time of the voltage signal output by the tactile sensor under the same mechanical stimulation. Therefore, the hardness of the object can be distinguished by establishing a quantitative relationship between the response time of the tactile sensor and the hardness of the object, and the hardness of the object is identified through the tactile sensor.

[0071] In some embodiments, as Figure 4 As shown, the present invention also provides a method for preparing a tactile sensor in the object hardness recognition system as described above, and the method for preparing the tactile sensor comprises the following steps:

[0072] S1. Provide a protrusion, a first conductive layer, a first friction layer, a second friction layer, and a second conductive layer;

[0073] S2, adhering the first friction layer to the lower surface of the first conductive layer;

[0074] S3, sticking the second friction layer to the upper surface of the second conductive layer;

[0075] S4, sticking the protrusions on the upper surface of the first conductive layer;

[0076] S5. Laminating the first friction layer and the second friction layer together.

[0077] In this embodiment, first, required materials (a protrusion, a first friction layer, a first conductive layer, a second friction layer, and a second conductive layer) are provided.

[0078] Specifically, when preparing the protrusion, a flame-retardant glass fiber reinforced epoxy resin material can be used and cut and manufactured according to the specific structure of the protrusion; when preparing the first friction layer, 800-grit sandpaper can be used as a template, and a polydimethylsiloxane solution can be coated on the sandpaper. After the film is formed, the film is separated from the sandpaper to obtain the first friction layer (a polydimethylsiloxane film with a surface microstructure); the second friction layer can be a polytetrafluoroethylene film that can be produced by existing technology. When selecting, those skilled in the art can select the polytetrafluoroethylene film according to the thickness requirements and cut it into the required shape; the first conductive layer can be any conductive metal material, for example, the material of the first conductive layer can be copper; for example, the material of the first conductive layer can be aluminum; preferably, the material of the first conductive layer is copper, wherein the upper surface and the lower surface of the first conductive layer are respectively covered with copper glue; when setting the second conductive layer, the setting method of the first conductive layer can refer to the above-mentioned setting method of the first conductive layer, which will not be repeated here, wherein, when the second conductive layer is coated with copper glue, it is only necessary to coat the copper glue on the upper surface of the second conductive layer.

[0079] Thereafter, the first friction layer is adhered to the lower surface of the first conductive layer, the protrusion is adhered to the upper surface of the first conductive layer, and the second friction layer is adhered to the upper surface of the second conductive layer.

[0080] Finally, the first friction layer and the second friction layer are laminated to each other to form a tactile sensor.

[0081] In order to further illustrate the manufacturing method and specific application of the tactile sensor, the following example is used for illustration.

[0082] In Example 1, 800-grit sandpaper with a thickness of 0.1 mm was provided as a substrate. A polydimethylsiloxane solution was spin-coated on the sandpaper and then cured. After film formation, the polydimethylsiloxane film was separated from the sandpaper to obtain a polydimethylsiloxane film with a sandpaper surface microstructure. Furthermore, the polydimethylsiloxane film and polytetrafluoroethylene film were cut into 25 mm × 25 mm sizes. The cut polydimethylsiloxane film was attached to the lower surface of the first conductive layer (a copper plate with copper glue coated on the upper and lower surfaces, respectively) as a first friction layer, and the protrusions (hemispherical protrusions) were attached to the upper surface of the first conductive layer. The cut polytetrafluoroethylene film was attached to the second conductive layer (a copper plate with copper glue coated on the upper surface) as a second friction layer. Finally, the first friction layer and the second friction layer were bonded to each other. At this point, the tactile sensor was completed.

[0083] Furthermore, when applying Figure 1As shown, the first conductive layer is electrically connected to the signal acquisition unit through a wire. Similarly, the second conductive layer is electrically connected to the signal acquisition unit through another wire, and the signal acquisition unit is connected to the machine learning calculation unit (a machine learning network model is preset in the machine learning calculation unit). At this time, the hardness of the object can be identified.

[0084] Specifically, the tactile sensor is pressed at a speed of nearly 2 cm / s, contacting a standard sample with a Shore hardness of 5HA (Hardness) and a pressing depth of 0.5 mm. The tactile sensor outputs a corresponding voltage signal to the signal acquisition unit. After receiving the voltage signal, the signal acquisition unit outputs a detection signal to the machine learning calculation unit. The machine computing learning unit recognizes the hardness of the object after training based on the machine learning network model.

[0085] In Example 2, 800-grit sandpaper with a thickness of 0.1 mm was provided as a substrate. A polydimethylsiloxane solution was spin-coated on the sandpaper and then cured. After film formation, the polydimethylsiloxane film was separated from the sandpaper to obtain a polydimethylsiloxane film with a sandpaper surface microstructure. Furthermore, the polydimethylsiloxane film and polytetrafluoroethylene film were cut into 25 mm × 25 mm sizes. The cut polydimethylsiloxane film was attached to the lower surface of the first conductive layer (a copper plate with copper glue coated on the upper and lower surfaces, respectively) as a first friction layer, and the protrusions (hemispherical protrusions) were attached to the upper surface of the first conductive layer. The cut polytetrafluoroethylene film was attached to the second conductive layer (a copper plate with copper glue coated on the upper surface) as a second friction layer. Finally, the first friction layer and the second friction layer were bonded to each other. At this point, the tactile sensor was completed.

[0086] Specifically, the tactile sensor is pressed at a speed of nearly 2 cm / s, contacting a standard sample with a Shore hardness of 15HA, and the pressing depth is 0.5 mm. The tactile sensor outputs a corresponding voltage signal to the signal acquisition unit. After receiving the voltage signal, the signal acquisition unit outputs a detection signal to the machine learning calculation unit. The machine computing learning unit recognizes the hardness of the object after training based on the machine learning network model.

[0087] Example 3: Provide 800-grit sandpaper with a thickness of 0.1 mm as a substrate, spin-coat a polydimethylsiloxane solution on the sandpaper and solidify it. After film formation, separate the polydimethylsiloxane film from the sandpaper to obtain a polydimethylsiloxane film with a sandpaper surface microstructure. Further, cut the polydimethylsiloxane film and polytetrafluoroethylene film into a size of 25 mm × 25 mm. The cut polydimethylsiloxane film is attached to the lower surface of the first conductive layer (a copper plate with copper glue coated on the upper and lower surfaces respectively) as a first friction layer, and the protrusions (hemispherical protrusions) are attached to the upper surface of the first conductive layer; the cut polytetrafluoroethylene film is attached to the second conductive layer (a copper plate with copper glue coated on the upper surface) as a second friction layer; finally, the first friction layer and the second friction layer are bonded to each other. At this point, the tactile sensor is completed.

[0088] Specifically, the tactile sensor is pressed at a speed of nearly 2 cm / s, contacting a standard sample with a Shore hardness of 30HA, and the pressing depth is 0.5 mm. The tactile sensor outputs a corresponding voltage signal to the signal acquisition unit. After receiving the voltage signal, the signal acquisition unit outputs a detection signal to the machine learning calculation unit. The machine computing learning unit identifies the hardness of the object after training based on the machine learning network model.

[0089] like Figure 5 As shown, Figure 5 The waveform of the tactile sensor output voltage when pressing samples with different Shore hardness is shown. It can be seen that as the hardness of the sample increases, the voltage amplitude increases and the pulse width decreases significantly. The tactile sensor will react faster when contacting a harder elastomer sample. Figure 5 As can be seen from the figure, the tactile sensor can distinguish objects with a Shore hardness of 0HA to 85HA.

[0090] like Figure 6 As shown, Figure 6 The relationship between the response time and peak voltage of the tactile sensor and the Shore hardness of the sample is shown. The response time and peak voltage have an obvious linear relationship with the hardness of the sample. In the range of 0 to 85HA, the response time sensitivity is about -0.107ms / HA.

[0091] In this embodiment, through Young's modulus matching and integrated sensing structure design, the tactile sensor is able to accurately identify the hardness of objects within a wide range of Shore hardness based on the intrinsic response of the tactile sensor. The research on the mechanism of object hardness identification based on response time provides a new way to enrich and enhance the perception ability of tactile sensors.

[0092] In summary, the present invention provides an object hardness identification system and method, and a tactile sensor preparation method, which have the following beneficial effects:

[0093] When identifying the hardness of an object, the sensitivity of the tactile sensor to dynamic excitation is utilized to make the tactile sensor output a voltage signal when in contact with the object. The voltage signal is further acquired by the signal acquisition unit and a detection signal is output based on the voltage signal. The machine learning calculation unit receives the detection signal and outputs the identified object hardness data after training the detection signal based on the machine learning network model. When the tactile sensor in the object hardness identification system contacts objects of different hardness, the response time of the voltage signal output by the tactile sensor under the same mechanical stimulation is significantly different. Therefore, the hardness of the object can be distinguished by establishing a quantitative relationship between the response time of the tactile sensor and the hardness of the object. The hardness of the object is identified through the tactile sensor, further exploring new uses for the tactile sensor.

[0094] In addition, the object hardness recognition system proposed in this application has great application prospects in the field of intelligent human-computer interaction.

[0095] It is understandable that the above embodiments only express the preferred implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the patent scope of the present invention. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present invention, the above technical features can be freely combined, and several deformations and improvements can be made, all of which fall within the scope of protection of the present invention. Therefore, all equivalent changes and modifications made to the scope of the claims of the present invention should fall within the scope of coverage of the claims of the present invention.

Claims

1. An object hardness recognition system, characterized in that: include: a tactile sensor for outputting a voltage signal when in contact with an object whose hardness needs to be identified; a signal acquisition unit connected to the tactile sensor, configured to receive the voltage signal and output a detection signal according to the voltage signal; as well as a machine learning computing unit connected to the signal acquisition unit, wherein the machine learning computing unit is preset with a machine learning network model, and is configured to receive the detection signal and output recognized object hardness data after training the detection signal in a machine learning network based on the machine learning network model; The machine learning network model is a data set consisting of the detection signal and the corresponding object hardness when objects of different hardness are applied to the tactile sensor.

2. The object hardness identification system according to claim 1, characterized in that: The tactile sensor comprises: a second conductive layer, a second friction layer, a first friction layer, a first conductive layer and a protrusion which are stacked in sequence; The material of the second friction layer and the material of the first friction layer are triboelectric materials with opposite electrical properties, and the first friction layer and the second friction layer are bonded to each other; The protrusion is used to contact an object whose hardness needs to be identified and transmit force to the first friction layer, so that the first friction layer and the second friction layer generate a triboelectric effect under the action of an external force.

3. The object hardness identification system according to claim 2, characterized in that: The protrusion is a hemispherical protrusion structure.

4. The object hardness identification system according to claim 2 or 3, characterized in that: The protrusion is made of flame-retardant glass fiber reinforced epoxy resin material.

5. The object hardness identification system according to claim 2, characterized in that: The upper surface of the first conductive layer and the lower surface of the first conductive layer are respectively covered with copper paste.

6. The object hardness identification system according to claim 5, characterized in that: The upper surface of the second conductive layer is covered with copper paste.

7. The object hardness identification system according to claim 2, characterized in that: The first friction layer includes a polydimethylsiloxane film with a sandpaper surface microstructure.

8. The object hardness identification system according to claim 7, characterized in that: The second friction layer includes a polytetrafluoroethylene film with a smooth surface.

9. A method for applying the object hardness identification system according to any one of claims 1 to 8, characterized in that: The object hardness identification system includes a tactile sensor, which is used to output a voltage signal when in contact with an object. The object hardness identification system method includes: receiving the voltage signal and outputting a detection signal according to the voltage signal; The detection signal is received and the object hardness data is output after training the detection signal based on a machine learning network model, wherein the machine learning network model is a data set consisting of the detection signal and the corresponding object hardness when objects with different hardness are sensed by the tactile sensor.

10. A method for preparing the tactile sensor in the object hardness identification system according to any one of claims 2 to 8, characterized in that: include: Providing a protrusion, a first conductive layer, a first friction layer, a second friction layer, and a second conductive layer; Adhere the first friction layer to the lower surface of the first conductive layer; adhering the second friction layer to the upper surface of the second conductive layer; Adhere the protrusion to the upper surface of the first conductive layer; The first friction layer and the second friction layer are bonded to each other.

Images