Identification device and robot using the same

Abstract

To provide an identification device which can identify a target object by being in contact with the target object, and a robot having the identification device.SOLUTION: An identification device 10 for identifying a target object includes: electric conductors 11, 11a which are elastically deformed by force applied from a target object, the electric resistance characteristics of the electric conductors being changed; input electrodes 12, 12a for applying a voltage to the electric conductors 11, 11a; a plurality of output electrodes 13, 13a for taking out an electric signal from each of different parts of the electric conductors 11, 11a when a voltage is being applied on the electric conductors; and an identification processing unit 14 for performing identification processing of a target object which provides force to the electric conductors 12, 12a by machine learning on the basis of an electric signal as an input value taken out from the output electrodes 13, 13a.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an identification device for identifying an object and a robot having the same. [Background technology]

[0002] Identification devices that identify objects are often used in fields where different processing is performed on objects or the force with which the object is gripped is adjusted depending on the type of object, etc. Conventional identification devices include those that identify objects by reading barcodes attached to the object (see Patent Document 1) and those that identify objects based on images of the object captured by a camera (see Patent Document 2). When using a barcode, it is necessary to attach the barcode to the object in advance, but when using an image captured by a camera, there is an advantage in that it is not necessary to attach a specific object to the object in advance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2013-010150 A [Patent Document 2] JP 2019-133516 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, when identifying an object based on an image captured by a camera and gripping the object, and then moving the object to a different location or changing the magnitude of the gripping force depending on the identification result, it was necessary to provide a camera separate from the device gripping the object. In this regard, if there was a mechanism that could identify the object by touching it, the device gripping the object could also identify the object, which would be effective. The present invention has been made in view of the above circumstances, and has an object to provide an identification device capable of identifying an object by contacting the object, and a robot having the identification device. [Means for solving the problem]

[0005] An identification device according to a first invention that meets the above-mentioned objective is an identification device for identifying an object, and includes an electrical conductor that elastically deforms and changes its electrical resistance characteristics when a force is applied from the object, an input electrode that applies a voltage to the electrical conductor, a plurality of output electrodes that extract electrical signals from different locations on the electrical conductor when a voltage is applied, and an identification processing unit that uses the electrical signals extracted by each output electrode as input values ​​and performs, through machine learning, an identification process for identifying the object that is applying a force to the electrical conductor.

[0006] A robot according to a second invention that meets the above-mentioned objective is a robot having a gripping unit for gripping an object and a control means for controlling the rotation or movement of the gripping unit, wherein the gripping unit is provided with all or part of an identification device for identifying the object being gripped, the identification device comprising: an electrical conductor that elastically deforms and changes its electrical resistance characteristics when force is applied from the object, an input electrode for applying a voltage to the electrical conductor, a plurality of output electrodes for extracting electrical signals from different points on the electrical conductor when a voltage is applied, and an identification processing unit that uses the electrical signals extracted by each output electrode as input values ​​and performs, through machine learning, an identification process for the object that is applying force to the electrical conductor, and the control means controls the operation of the gripping unit based on the identification result of the object by the identification device. Effect of the Invention

[0007] The identification device of the first invention comprises an electrical conductor that elastically deforms and changes its electrical resistance characteristics when a force is applied from an object, an input electrode that applies a voltage to the electrical conductor, a plurality of output electrodes that extract electrical signals from different locations on the electrical conductor when a voltage is applied, and an identification processing unit that uses the electrical signals extracted by each output electrode as input values ​​and performs, through machine learning, an identification process for identifying the object that is applying a force to the electrical conductor, and therefore is able to identify the object by coming into contact with the object. Furthermore, the robot according to the second aspect of the present invention has the identification device according to the first aspect of the present invention, and is therefore capable of identifying an object by the identification device coming into contact with the object. [Brief description of the drawings]

[0008] [Figure 1] 1 is an explanatory diagram of a robot equipped with an identification device according to an embodiment of the present invention. [Diagram 2] FIG. 4 is an explanatory diagram showing attachment of an electrical conductor to a base body. [Diagram 3] FIG. 2 is an explanatory diagram showing the arrangement of an input terminal section and an output terminal section. [Figure 4] 13A and 13B are explanatory diagrams showing the measurement results of the voltage of the output electrode. [Diagram 5] 13A and 13B are explanatory diagrams showing the measurement results of the voltage of the output electrode. [Figure 6] 13A and 13B are explanatory diagrams showing the determination results of the target sample. [Figure 7] 13A and 13B are explanatory diagrams showing the determination results of the target sample. [Figure 8] FIG. 13 is an explanatory diagram showing the results of the accuracy of determination of a target sample expressed in numerical values. [Figure 9] 13A and 13B are explanatory diagrams showing the measurement results of the voltage of the output electrode. [Figure 10] 13A and 13B are explanatory diagrams showing the measurement results of the voltage of the output electrode. [Figure 11] 13A and 13B are explanatory diagrams showing the determination results of the target sample. [Figure 12] 13A and 13B are explanatory diagrams showing the determination results of the target sample. [Figure 13] FIG. 2 is a block diagram showing connections of a control means. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Next, with reference to the attached drawings, an embodiment of the present invention will be described for understanding the present invention. 1 to 3, an identification device 10 according to an embodiment of the present invention is a device for identifying an object (not shown), and includes electric conductors 11 and 11a that are elastically deformed when a force is applied from the object, an input electrode 12 that applies a voltage to the electric conductor 11, an input electrode 12a that applies a voltage to the electric conductor 11a, a plurality of output electrodes 13 that extract electric signals from different positions of the electric conductor 11, a plurality of output electrodes 13a that extract electric signals from different positions of the electric conductor 11a, and an identification processing unit 14 that performs an identification process of the object applying a force to the electric conductors 11 and 11a by machine learning using the electric signals extracted by each output electrode 13 and each output electrode 13a as input values. The details will be described below.

[0010] In this embodiment, the identification device 10 is provided in a robot 15, as shown in Fig. 1. The robot 15 has two (one example of a plurality) electric conductors 11, 11a, base bodies 16, 16a to which the electric conductors 11, 11a are respectively fixed, and control means 17 that operates a motor (not shown) to move the base bodies 16, 16a together with the electric conductors 11, 11a, respectively, to change the distance between the electric conductors 11, 11a.

[0011] The electric conductors 11, 11a and the base bodies 16, 16a are each plate-shaped, and the electric conductor 11 is attached to one side surface (one example of one side) of the base body 16 as shown in Figs. 1 and 2, and the electric conductor 11a is also attached to one side surface (one example of one side) of the base body 16a. As shown in Fig. 1, the electric conductors 11, 11a and the base bodies 16, 16a are arranged in the order of the base body 16, the electric conductors 11, 11a and the base body 16a. That is, the electric conductors 11, 11a are arranged so that an object placed between the electric conductors 11, 11a can be sandwiched. Thus, in this embodiment, there are a plurality of electric conductors 11, 11a, and the object is sandwiched between the plurality of electric conductors 11, 11a.

[0012] In this embodiment, the electric conductor 11 and the base body 16 are both rectangular plate-shaped, as shown in Fig. 2 and Fig. 3. The electric conductor 11 is smaller than the base body 16, and is attached to the base body 16 at a position where the center of the electric conductor 11 coincides with the center of the base body 16. In this embodiment, the electric conductor 11a, the base body 16a, the input electrode 12a, and the output electrode 13a are designed similarly to the electric conductor 11, the base body 16, the input electrode 12, and the output electrode 13, respectively. Therefore, hereinafter, the electric conductor 11, the base body 16, the input electrode 12, and the output electrode 13 will be described, and detailed descriptions of the electric conductor 11a, the base body 16a, the input electrode 12a, and the output electrode 13a will be omitted.

[0013] As shown in FIG. 2, the electric conductor 11 has a plurality (a large number) of conductive pieces 18 each having electrical conductivity and a rectangular plate-like elastic portion 19 supporting all of the plurality of conductive pieces 18, and the conductive pieces 18 are electrically connected by being in contact with each other. The electric conductor 11 only needs to have a shape that is elastically deformed when a force is applied, and needs to have a certain thickness or more in the direction in which the force is applied. If the thickness of the electric conductor 11 is less than 0.1 mm, the amount of elastic deformation becomes small, and as a result, there are cases in which the object cannot be stably identified. Furthermore, as a result of experimental verification, it was confirmed that the ability to identify the object decreases when the thickness of the electric conductor 11 exceeds 3 mm. Therefore, the thickness of the electric conductor 11 is preferably 0.1 mm or more and 3 mm or less.

[0014] The shape of each conductive piece 18 is not particularly limited, and may be, for example, cylindrical or linear. In this embodiment, carbon nanotubes are used as the conductive pieces 18, but the present invention is not limited to this. For example, organic nanowires other than carbon nanotubes, metallic nanowires made of one or more elements selected from the group consisting of Ag, Au, Ni, Cu, Pd, Pt, Rh, Ir, Ru, Os, Fe, Co, and Sn, IrO 2 , In 2 O 3 , SnO 2The conductive piece 18 can be an oxide nanowire made of one or more oxides selected from the group consisting of ITO, a polymerizable polymer wire, or a composite wire in which the surface of an insulating nanowire such as DNA is covered with a conductive material.

[0015] The elastic portion 19 supports the conductive pieces 18 so that the conductive pieces 18 are distributed in a predetermined manner (in this embodiment, substantially uniformly distributed) when the electrical conductor 11 is not elastically deformed. When a force is applied to the electrical conductor 11, the entire electrical conductor 11 is elastically deformed mainly due to the elastic deformation of the elastic portion 19. The conductive pieces 18 move due to the elastic deformation of the elastic portion 19 (the electrical conductor 11) (naturally, some conductive pieces 18 do not move depending on the degree of elastic deformation of the elastic portion 19), and when the force is released from the electrical conductor 11, the electrical conductor 11 returns to its shape before the elastic deformation, and the distribution (position) of the conductive pieces 18 also substantially returns to the original.

[0016] In the present embodiment, the elastic portion 19 is made of polydimethylsiloxane, but is not limited thereto. The elastic portion 19 may be made of any material as long as it is elastically deformable and has insulating properties. For example, hydrogel may be used for the elastic portion 19. When the conductive pieces 18 are carbon nanotubes and the elastic portions 19 are polydimethylsiloxane, the electrical conductor 11 can be manufactured, for example, by the first to third steps described below.

[0017] First step: Water is poured into a container containing sugar (any water-soluble substance is acceptable) and carbon nanotubes, and the resulting solution is poured into a mold (e.g., a tray) and dried in an oven or the like to obtain an intermediate body in the shape (e.g., a plate) that corresponds to the mold. Second step: After the intermediate body is immersed in liquid polydimethylsiloxane, a vacuum is applied, and the intermediate body is cooled in a refrigerator or the like, and then dried in an oven or the like. Third step: The intermediate that has been dried is immersed in water to dissolve and remove the sugar contained in the intermediate, and then the intermediate is dried to produce an electrical conductor 11.

[0018] Therefore, a large number of cavities are formed inside the electrical conductor 11, and a large number of holes are formed on the surface thereof. These cavities and holes are not shown in the enlarged view of FIG. Alternatively, the electrical conductor 11 may be produced by forming a material mainly composed of carbon nanotubes and polydimethylsiloxane into a wide sheet shape, and then cutting the sheet shape into a desired shape.

[0019] 2 and 3, one input electrode 12 and multiple output electrodes 13 are attached to the base body 16 having the electric conductor 11 attached to one side surface. The one input electrode 12 and multiple output electrodes 13 are each needle-shaped and protrude from the other side surface of the base body 16 (an example of the other side, and in this embodiment, the surface opposite to the surface to which the electric conductor 11 is attached). On one side surface (one side) of the base body 16, one input terminal section 20 electrically connected to the one input electrode 12 and multiple output terminal sections 21 electrically connected to the multiple output electrodes 13 are provided. In this embodiment, the input electrode 12 and the input terminal section 20 are electrically connected to each other, and the output electrode 13 and the output terminal section 21 are electrically connected to each other through conductor wires provided on the base body 15.

[0020] The input terminal 20 and the output terminals 21 are each semispherical and protrude from one side of the base body 16. In this embodiment, as shown in FIG. 3, the input terminal 20 is provided at the center of the base body 16, and the output terminals 21 are provided around the input terminal 20. The electrical conductor 11 is bonded to one side of the base body 16 with a silane coupling agent and is in contact with the input terminal 20 and all of the output terminals 21. In this embodiment, the input terminal 20 and the output terminals 21 are arranged at equal intervals so that the input terminal 20 and the output terminals 21 are distributed uniformly. It goes without saying that the distribution does not have to be uniform.

[0021] By making the input terminal portion 20 and the multiple output terminal portions 21 semispherical, the input terminal portion 20 and the multiple output terminal portions 21 are brought into contact with or pressed against the electric conductor 11, and this makes it possible to prevent deterioration of characteristics compared to the case where the input terminal portion 20 and the multiple output terminal portions 21 are made into pin-like shapes and inserted into the electric conductor 11. That is, repeated compression and relaxation of the electric conductor 11 over a long period of time changes the contact between the pin and the electric conductor 11, and the variation in characteristics increases. However, by making the input terminal portion 20 and the multiple output terminal portions 21 semispherical and brought into contact with the electric conductor 11, deterioration of characteristics can be reduced even if repeated compression and relaxation of the electric conductor 11 over a long period of time occurs. In addition, by bonding the base body 16 and the electric conductor 11 with a silane coupling agent, a certain degree of electrical conductivity between the input terminal portion 20 and the multiple output terminal portions 21 and the electric conductor 11 is maintained, and the silane coupling agent can ensure a certain degree of flexibility even when hardened, so that changes in characteristics can be suppressed. If the base body 16 and the electrical conductor 11 are joined with an instant adhesive or the like, the instant adhesive will have a considerable hardness when hardened, and this may cause the deformation of the electrical conductor 11 to change partially, which may also lead to deterioration of the characteristics.

[0022] On one side surface, an area where one input terminal portion 20 and multiple output terminal portions 21 are arranged and where the electrical conductor 11 is adhered is defined as a predetermined area, and on the other side surface, an electrode group where one input electrode 12 and multiple output electrodes 13 are arranged in two rows, and another electrode group where multiple output electrodes 13 are arranged in two rows are arranged so as to sandwich the area behind the predetermined area. Note that, although detailed explanation is omitted, a terminal portion for grounding is also provided on one side surface of the base body 16, and a needle-shaped electrode for grounding is also provided on the other side surface of the base body 16.

[0023] As shown in FIG. 1, the control means 17 is accommodated in a housing 22 together with the identification processing unit 14, and is designed to be able to apply an externally applied voltage to the input electrode 12. The voltage applied from the control means 17 to the input electrode 12 is applied to the electrical conductor 11 via the input terminal unit 20. The multiple output electrodes 13 each extract an electrical signal (meaning a DC, AC, or pulsed voltage signal, or a DC, AC, or pulsed current signal) from a point (a plurality of different points) that is in contact with the output terminal unit 21 of the electrical conductor 11 via the output terminal unit 21 and a conductor. The electrical signal that each output electrode 13 obtains from the electrical conductor 11 is sent to the identification processing unit 14 via a separate conductor.

[0024] When a force is applied to the electrical conductor 11, the electrical conductor 11 undergoes elastic deformation around the portion to which the force is applied, causing a change in the state of electrical connection between the conductive pieces 18, which in turn changes the electrical resistance characteristics of the electrical conductor 11. Here, a change in the state of electrical connection between the conductive pieces 18 means that two conductive pieces 18 that were not in contact come into contact, or two conductive pieces 18 that were in contact go out of contact, or the contact position, number of contacts, contact area, etc. of two conductive pieces 18 that were in contact change.

[0025] Since the electrical resistance characteristics of electrical conductor 11 change due to the elastic deformation of electrical conductor 11, the magnitude of the electrical signal extracted to each output electrode 13 from electrical conductor 11 to which a voltage is applied also changes according to the elastic deformation of electrical conductor 11. The same is true for electrical conductor 11a, so when electrical conductors 11, 11a are elastically deformed by sandwiching an object between them, the magnitude of the electrical signal extracted to each output electrode 13 and the magnitude of the electrical signal extracted to each output electrode 13a change according to the elastic deformation of electrical conductors 11, 11a.

[0026] In this embodiment, it is assumed that the electric conductors 11 and 11a directly contact the object to sandwich the object, but it is also possible to attach a predetermined object (e.g., an elastic sheet or film) to the surface of the electric conductor 11 close to the object and to the surface of the electric conductor 11a close to the object, and to sandwich the object between the electric conductors 11 and 11a via the object. In this case, the electric conductors 11 and 11a do not directly contact the object.

[0027] The identification processing unit 14 acquires the analog electric signals provided from each output electrode 13a through an AD conversion unit 31 (see FIG. 13) that converts the analog electric signals provided from each output electrode 13a into digital values. The identification processing unit 14 has a machine learning function that identifies an object sandwiched between the electrical conductors 11, 11a by machine learning based on the acquired digital values. The identification processing unit 14 can be configured by a microcomputer or the like.

[0028] The machine learning system mounted in the identification processing unit 14 is not particularly limited, and linear regression, a support vector machine, or the like can be used. In this embodiment, the identification processing unit 14 finally detects (identifies) information on the object sandwiched between the electrical conductors 11, 11a (for example, the type, size, shape, weight, hardness, and the like of the object) from a plurality of electrical signals given to the identification processing unit 14. The control means 17 determines the magnitude of the force sandwiching the object, as well as the orientation and movement direction of the electrical conductors 11, 11a sandwiching the object, and the like, based on the detection result of the identification processing unit 14, and transmits a predetermined command signal to a motor or the like of the robot 15 according to the determination. EXAMPLES

[0029] Next, an experiment conducted to confirm the effects of the present invention will be described. In the experiment, two experimental models were prepared, each of which had two sets of a rectangular plate-shaped base body with a total of 49 terminals, each of which had seven parallel rows of seven hemispherical terminals arranged linearly at equal intervals, on one side and 50 needle-shaped electrodes on the other side, and a square plate-shaped electric conductor bonded to the center of one side of the base body with a silane coupling agent and in contact with all of the 49 terminals, as shown in Figures 2 and 3. In one experimental model, the thickness of the two electric conductors was 0.3 cm, and in the other experimental model, the thickness of the two electric conductors was 0.5 cm.

[0030] Each experimental model had two rows of 13 electrodes arranged in a straight line with equal pitch on one side of the base body in the longitudinal direction (total of 26 electrodes), and two rows of 12 electrodes arranged in a straight line with equal pitch on the other side of the base body in the longitudinal direction (total of 24 electrodes). Of the 49 terminals, the four terminals at the four corners were ground terminals, the central terminal was an input terminal, and the remaining 44 terminals were output terminals.

[0031] One electrode in each group of 26 electrodes was used as an input electrode, the other as a ground electrode, and the remaining 24 electrodes in each group of 26 electrodes and the 24 electrodes in each group of 24 electrodes were used as output electrodes, totaling 48 electrodes (the remaining 32 electrodes were unused). The system was designed so that the voltage of each output electrode was input as an electrical signal to a computer equipped with machine learning system software. Linear regression was used as the machine learning system, carbon nanotubes as the conductive pieces, and polydimethylsiloxane as the elastic part.

[0032] In the experiment, two electric conductors were placed facing each other, and with a voltage of 5V applied to each input electrode, the two electric conductors were moved together with a base body, etc., by operating a motor to shorten the distance between the two electric conductors, and the target sample was directly sandwiched between the two electric conductors, and the voltage of each output electrode was measured. Furthermore, the machine learning system was made to determine what the target sample sandwiched between the two electric conductors was based on the measured voltage of each output electrode.

[0033] The voltage of the output electrode was measured 100 times for each target sample, and 80 of the 100 voltage measurements were used as training data, and the remaining 20 were used as validation data (data for determining the target sample). Four target samples were prepared: 1) a resin apple toy with a height of 5.2 cm and a width of 5.5 cm, 2) a resin ball (sphere) with a diameter of 5 cm, 3) a resin cubic block with a side length of 2.2 cm, and 4) a stuffed hedgehog with a height of 15.5 cm and a width of 11.5 cm.

[0034] For an experimental model with an electrical conductor having a thickness of 0.3 cm (hereinafter referred to as the "0.3 cm experimental model"), the voltage measurement results of each output electrode when a toy apple, a ball, a cube block, and a stuffed toy were sandwiched are shown in Figures 4(A), (B), 5(A), and (B), respectively. The judgment results of the target sample by the 0.3 cm experimental model (the judgment results of which target sample was sandwiched) are shown in Figures 6(A), (B), 7(A), and (B), respectively, and the numerical results of the accuracy of the judgment of the target sample by the same experimental model are shown in Figure 8.

[0035] For an experimental model with an electrical conductor having a thickness of 0.5 cm (hereinafter referred to as the "0.5 cm experimental model"), the measurement results of the voltage of each output electrode when a toy apple, a ball, a cube block, and a stuffed animal were sandwiched are shown in Figures 9(A) and (B), 10(A) and (B), respectively, and the evaluation results of the target samples using the 0.5 cm experimental model are shown in Figures 11(A) and (B), and 12(A) and (B), respectively.

[0036] "Toy apple" and the like in Figures 4(A), (B) to 7(A), (B) and Figures 9(A), (B) to 12(A), (B) indicate target samples that are actually sandwiched between electrical conductors. "Toy apple" means a toy apple, "Ball" means a ball, "Rigid block" means a cubic block, and "Plush toy" means a stuffed toy.

[0037] Graph lines corresponding to the four target samples are depicted in Figures 6(A), (B), 7(A), (B), 11(A), (B), and 12(A), (B). Each graph line is a series of values ​​(hereinafter referred to as "likelihood values") that indicate the likelihood that the target sample sandwiched between the electrical conductors is the target sample corresponding to the graph line, and the larger the likelihood value is, the more likely it is that the target sample sandwiched between the electrical conductors is the target sample.

[0038] In the experiment, first, the machine learning system was trained using training data so that the likelihood value corresponding to the target sample sandwiched between an electrical conductor would be 1, and the likelihood values ​​corresponding to the other three target samples would be 0.Then, using the validation data, the likelihood values ​​corresponding to each of the four target samples are shown in Figures 6(A) and (B), Figures 7(A) and (B), Figures 11(A) and (B), and Figures 12(A) and (B).

[0039] From the results shown in Figures 6(A) and (B), Figures 7(A) and (B), Figures 11(A) and (B), and Figures 12(A) and (B), it can be seen that the 0.3 cm experimental model was able to determine the target sample more accurately than the 0.5 cm experimental model.

[0040] The values ​​shown in Figure 8 represent the average value of the likelihood values ​​of each graph line shown in Figures 6(A), (B), 7(A), and (B). For example, the average values ​​of the likelihood values ​​corresponding to the toy apple, ball, cubic block, and stuffed toy when the toy apple was pinched were 0.69, 0.05, 0.1, and 0.17, respectively. From the results shown in Figure 8, it was confirmed that the 0.3 cm experimental model stably made correct judgments for all target samples.

[0041] 1, the gripping section for gripping an object in the above-described robot 15 has the electrical conductors 11, 11a, input electrodes 12, 12a, output electrodes 13, 13a, base bodies 16, 16a, input terminal section 20, and output terminal section 21 of the identification device 10. That is, a part of the identification device 10 is provided in the gripping section, but the entire identification device (entire including the identification processing section) can also be provided in the gripping section.

[0042] Here, in robot 15, the gripper rotates and moves (including gripping of the object by electrical conductors 11, 11a) by the rotational motion of each part based on each of a plurality of robot axes 25 of arm 24, the tip of which is attached with a gripper. In this embodiment, control means 17 controls the operation of the gripper based on the identification result of the object by identification device 10. The multiple robot axes 25 correspond to multiple motors 26 shown in FIG. 13, respectively, and the control means 17 controls the rotation output and rotation direction of each motor 26 via motor drive means 27.

[0043] In this embodiment, an electric motor such as a servo motor is used as each motor 26, but a solenoid or one equipped with an electric motor and a solenoid may also be used. The robot 15 is also provided with a DA conversion unit 28 that passes current through the input electrodes 12, 12a in response to commands from the control means 17, and a display input unit 29 that combines a display means such as a liquid crystal display or an organic EL display with an input means such as a touch panel.

[0044] In this embodiment, the display means and the input means are implemented as one device, but each may be implemented as an individual device. For example, the input means may be in the form of a dial or buttons. Furthermore, the control means 17 may be connected to a communication means 30 for performing data communication with an external device through wired communication or wireless communication.

[0045] An example of the operation of the robot 15 will now be described. When an operation command for the robot 15 is input from the display input unit 29 or the communication unit 30, the control unit 17 outputs a signal to the motor driving unit 27 to drive a predetermined motor 26 and move the gripping unit to a position where it can grip an object at a predetermined position. Next, the control unit 17 outputs a signal to the DA conversion unit 28 to supply a voltage to the input electrodes 12, 12a, and also outputs a signal to the motor driving unit 28 to drive the predetermined motor 26 so as to move the electric conductors 11, 11a together with the base bodies 16, 16a closer to each other, so that the object is gripped between the electric conductors 11, 11a.

[0046] By gripping an object between the electrical conductors 11, 11a, a signal is output from the output electrodes 13, 13a, and the signal is converted into a digital signal by the AD conversion unit 31. The identification processing unit 14 identifies what object is being gripped based on the digital signal. For example, a trained model is stored in advance in the identification processing unit 14. The control means 17 reads data such as the gripping force, orientation, and rotation speed of the gripping unit from a memory or the like provided in the control means 17 based on the signal corresponding to the type of identified object provided from the identification processing unit 14.

[0047] Then, the control means 17 outputs a signal to the motor driving means 27 according to the read data, and the motor driving means 27 drives each motor 26 with a predetermined output and a predetermined rotation direction. The control means 17 may display the type of object on the display input unit 29. By determining the type of object held by the gripping unit, etc., it is possible to determine the transport conditions according to the type of object, so that the object is not damaged and the object can be prevented from falling with an extremely high probability.

[0048] For example, if the object to be grasped is a ball, by controlling each motor 26 so that the base bodies 16, 16a are horizontal while grasping the ball, the direction of gravity acting on the ball becomes perpendicular to the base bodies 16, 16a, and the ball almost never slips off the base bodies 16, 16a. Furthermore, the robot 15 of this embodiment is capable of identifying objects even in places where image recognition by a camera is not possible, such as darkness, and is therefore considered to have improved versatility.

[0049] Although the embodiment of the present invention has been described above, the present invention is not limited to the above-mentioned embodiment, and all changes in conditions that do not depart from the gist of the present invention are within the scope of application of the present invention. For example, the identification device can be provided on something other than a robot (for example, equipment that clamps and fixes objects), or the identification device can be used independently. Furthermore, the electrical conductor does not have to be in the form of a rectangular (including square) plate, but may be, for example, a circular, triangular, or polygonal plate having 5 or more sides, and it does not have to be in the form of a plate, but for example, a spherical electrical conductor or a columnar electrical conductor can also be used.

[0050] The number of electric conductors may be one or three or more, and when there is one electric conductor, the object is identified by pressing the object against the electric conductor, and when there are three electric conductors, the electric conductors can be designed to be held in a state where the three electric conductors are in contact with the object. Furthermore, it is possible to electrically connect a plurality of input electrodes to one electrical conductor, or to electrically connect a single output electrode to one electrical conductor, although from the viewpoint of stably identifying an object, it is preferable that a plurality of output electrodes are electrically connected to one electrical conductor. [Explanation of symbols]

[0051] 10: identification device, 11, 11a: electrical conductor, 12, 12a: input electrode, 13, 13a: output electrode, 14: identification processing unit, 15: robot, 16, 16a: base body, 17: control means, 18: conductive piece, 19: elastic portion, 20: input terminal portion, 21: output terminal portion, 22: housing, 24: arm, 25: robot axis, 26: motor, 27: motor driving means, 28: DA conversion unit, 29: display input unit, 30: communication means, 31: AD conversion unit

Claims

1. An identification device for identifying an object, an electric conductor that is elastically deformed by a force applied from the object and has a change in electrical resistance characteristic; an input electrode for applying a voltage to the electrical conductor; a plurality of output electrodes for extracting electrical signals from different points of the electrical conductor when a voltage is applied; and an identification processing unit that uses the electrical signals extracted by each output electrode as input values ​​and performs an identification process on the object that is applying a force to the electrical conductor through machine learning.

2. 2. The identification device according to claim 1, wherein the electrical conductors are a plurality of electrical conductors, and the object is sandwiched between the plurality of electrical conductors.

3. 3. An identification device according to claim 1, further comprising a base body having one side on which the electrical conductor is attached, and the input electrode and the plurality of output electrodes protruding from the other side of the base body.

4. 4. The identification device according to claim 3, further comprising an input terminal portion electrically connected to the input electrode and a plurality of output terminal portions electrically connected to the plurality of output electrodes, respectively, on one side of the base body, and the electrical conductor is adhered to one side of the base body with a silane coupling agent and is in contact with the input terminal portion and the plurality of output terminal portions.

5. 3. The identification device according to claim 1, wherein the electrical conductor is in the form of a plate having a thickness of 0.1 mm to 3 mm.

6. A robot having a gripping unit that grips an object and a control means that controls rotation or movement of the gripping unit, The gripping unit is provided with an identification device for identifying the object being gripped, in whole or in part; The identification device includes: an electric conductor that is elastically deformed by a force applied from the object and has a change in electrical resistance characteristic; an input electrode for applying a voltage to the electrical conductor; a plurality of output electrodes for extracting electrical signals from different points of the electrical conductor when a voltage is applied; an identification processing unit that performs an identification process of the object exerting a force on the electrical conductor by machine learning using the electrical signals extracted by the output electrodes as input values; The robot is characterized in that the control means controls the operation of the gripping part based on the identification result of the object by the identification device.

Images