Control device, control method and control program
Through the slip detection unit and sensor system, the grip force of the robot is controlled, and the problem of grip force adjustment in the interaction between humans and robots is solved, achieving a natural and robust interactive effect.
Patent Information
- Application Number
- CN202080051472.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2019-07-23
- Filing Date
- 2020-07-10
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to achieve natural interaction between humans and robots, especially in the process of reducing and transmitting grip force, it is difficult to detect slight changes in external forces and distinguish between external forces related to robots and objects, resulting in unnatural interactions.
A slip detection unit is used to detect the slip of the grip part and the object, and control the grip force based on the detected slip, and estimate the applied external force and torque through the tactile sensor and the acceleration sensor to achieve dynamic adjustment of the grip force.
It realizes the natural interaction between humans and robots, can detect slight changes in external forces, distinguish the external forces between the robot and the object, and ensures the robustness and naturalness of the interaction.
Smart Images

Figure CN114206557B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to a control device, a control method, and a control program. Background Art
[0002] In order for a robot to coexist with a human, it is necessary to realize interaction with a human, and for this purpose, in addition to functions that can be realized by software (such as human recognition and voice interaction), functions that can be realized in hardware are also required. As one of the hardware interaction functions, there is a function of exchanging a held object with a human. As a robot control technology, it is necessary not only to increase the holding force in order to receive an object from another person, but also to reduce the holding force in order to transfer the held object to another person.
[0003] Citation List
[0004] Patent Literature
[0005] Patent Document 1: JP 4683073 B2
[0006] Patent Document 2: JP 4600445 B2 Summary of the invention
[0007] Technical issues
[0008] However, conventionally, although many technologies for increasing the gripping force when transferring an object have been studied, technologies for reducing the gripping force have not been sufficiently studied. Therefore, it is difficult to achieve natural interaction between a person and a robot like between a person.
[0009] An object of the present disclosure is to provide a control device, a control method, and a control program that can realize natural interaction between a human and a robot like that between humans.
[0010] Solution to the problem
[0011] In order to solve the above-mentioned problem, a control device according to one aspect of the present disclosure has: a slip detection unit, which detects the slip of a target object held by a holding part; and a control unit, which controls the holding force used by the holding part to hold the target object based on the slip detected by the slip detection unit, wherein the control unit estimates the external force applied to the target object held by the holding part based on the slip detected by the slip detection unit, and controls the holding force based on the estimated external force. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Figure 1 is a perspective view schematically showing the appearance of a robot applicable to the embodiment.
[0013] Figure 2 is a diagram schematically showing force generated in each part when human force acts on an object held by the robot hand.
[0014] Figure 3 This is a diagram showing a model for explaining the Hertz contact theory.
[0015] Figure 4 It shows that the external force Δf t The situation acting on an object is applied to a diagram of an example of a model for explaining the Hertz contact theory.
[0016] Figure 5 is a block diagram schematically showing a configuration of a robot apparatus to which the manipulator according to the embodiment is applicable.
[0017] Figure 6 is a functional block diagram for explaining an example of the functions of the manipulator according to the embodiment.
[0018] Figure 7 is a block diagram showing an example of a hardware configuration of a manipulator according to an embodiment.
[0019] Figure 8 is a flowchart schematically showing an example of control of a transfer operation according to the embodiment.
[0020] Fig. 9 is a flowchart illustrating an example of a trigger detection operation of a transfer operation according to the embodiment.
[0021] Fig.10 is a flowchart illustrating an example of an operation of controlling a transfer grip force according to an embodiment.
[0022] Fig.11 is a flowchart illustrating an example of an operation of stability confirmation of a transfer operation according to an embodiment.
[0023] Fig. 12A is a schematic diagram showing the shear force f of the manipulator t A diagram showing an example of a case where the direction of is consistent with the direction of gravity Mg.
[0024] Fig. 12B is a schematic diagram showing the shear force f of the manipulator t A diagram showing an example of a case where the direction of is consistent with the direction of gravity Mg.
[0025] Fig.13A is a schematic diagram showing the shear force f of the manipulator t A diagram showing an example of a case where the direction of is inconsistent with the direction of gravity Mg.
[0026] Fig. 13B is a schematic diagram showing the shear force f of the manipulator t A diagram showing an example of a case where the direction of is inconsistent with the direction of gravity Mg.
[0027] Fig.14 It is a diagram for explaining the deviation.
[0028] Fig.15 is a flowchart schematically showing an example of control of a receiving operation according to the first modification of the embodiment.
[0029] Fig.16 is a block diagram schematically showing a configuration of a robot apparatus to which a manipulator according to a second modification of the embodiment is applied. DETAILED DESCRIPTION
[0030] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Note that in the following embodiments, the same parts are denoted by the same reference numerals, and redundant descriptions will be omitted.
[0031] Hereinafter, embodiments of the present disclosure will be described in the following order.
[0032] 1. Overview of Embodiments
[0033] 1-1. Examples of manipulators applicable to the embodiments
[0034] 1-2. Prior Art
[0035] 2. More Specific Description of Transfer According to Embodiments
[0036] 2-1. Definition of each part
[0037] 2-2. External force estimation method using slip sensing
[0038] 2-3. Configuration of the robot according to the embodiment
[0039] 2-4. Control of Transfer Operation According to Embodiment
[0040] 2-5. Operation in any transfer direction according to the embodiment
[0041] 2-6. Application Examples of Control According to Embodiment
[0042] 3. First Modification of Embodiment
[0043] 4. Second Modification of Embodiment
[0044] 5. Third Modification of Embodiment
[0045] 6. Other configurations
[0046] 1. Overview of Embodiments
[0047] Embodiments according to the present disclosure relate to control of a robot capable of holding an object, and more particularly to a technology for appropriately controlling a holding force when a person transfers an object to the robot and when a person receives an object from the robot.
[0048] 1-1. Examples of manipulators applicable to the embodiments
[0049] Figure 1 : is a perspective view schematically showing the appearance of a robot applicable to the embodiment. Figure 1 In the figure, for the convenience of explanation, the manipulator 11 includes two fingers, namely, finger portions 13a and 13b, and the finger portions 13a and 13b are connected to the wrist portion 12 and arranged to face each other. That is, the finger portions 13a and 13b are held by using the wrist portion 12 as a holding portion. In addition, the finger portions 13a and 13b are provided with a parallel link mechanism, and are driven to open and close in such a manner that the links on the distal end side (opposite side of the wrist portion 12) of the finger portions 13a and 13b maintain a parallel relationship.
[0050] Hereinafter, the links on the distal end sides of the finger portions 13a and 13b are respectively referred to as fingertip portions. In addition, for example, the finger portions 13a and 13b are driven by an actuator to open and close according to a control signal provided from the outside.
[0051] exist Figure 1 In the robot 11, the touch sensor 15 is provided at the fingertip portion of the finger portion 13a, and the elastic body 14 is also provided for the touch sensor 15. An elastic resin such as rubber is used for the elastic body 14, and it deforms according to a force applied to a side facing a surface in contact with the touch sensor 15. Figure 1 In the example of , the elastic body 14 has a hemispherical shape, but it is not limited to this example. For example, the elastic body 14 may have a plate shape, or the surface in contact with the object held by the robot hand 11 may be subjected to processing such as unevenness.
[0052] The tactile sensor 15 outputs a detection signal corresponding to the applied pressure. In an embodiment, a pressure distribution sensor capable of detecting pressure distribution on a two-dimensional surface is applied as the tactile sensor 15. The tactile sensor 15 can detect the force applied to the elastic body 14 and the displacement amount and displacement direction in a two-dimensional plane of deformation generated by the force.
[0053] The elastic body 14 and the touch sensor 15 are similarly provided on the tip portion of the finger portion 13 b.
[0054] The manipulator 11 can clamp the target object by using the fingertips of the finger parts 13a and 13b, and hold the object using friction. In other words, the fingertips including the elastic body 14 (and the tactile sensor 15) can be referred to as the holding part that holds the object according to the movement of the finger parts 13a and 13b.
[0055] In addition, for example, an acceleration sensor (not shown) that detects the acceleration of the manipulator 11 is provided on the wrist 12. For example, an acceleration sensor that can detect acceleration in each direction of three axes is provided on the wrist 12. The magnitude of gravity applied to the manipulator 11 and the posture of the manipulator 11 can be detected based on the sensor output of the acceleration sensor. Here, for example, when an axis whose direction coincides with the direction of gravity is defined in the manipulator 11, the posture includes the inclination of the axis relative to the direction of gravity and the rotation about the axis in the direction of gravity.
[0056] The wrist 12 is also provided with a force sensor (not shown) for detecting the force applied to the entire manipulator 11. The force sensor may be constituted by, for example, an acceleration sensor. In an embodiment, the force sensor may be omitted.
[0057] 1-2. Prior Art
[0058] Here, the transfer of an object between a manipulator and a person according to the prior art will be described. According to the prior art, a method is proposed in which a force sensor provided at a wrist is used in a manipulator, and a finger portion is released at a moment when the force sensor detects a change amount equal to or greater than a predetermined threshold (for example, Patent Document 1 and Patent Document 2). However, in order to realize a natural interaction between a person and a manipulator by this method, it is considered that there are three main problems.
[0059] The first problem is that it is difficult to detect slight changes in external force associated with the fingertips due to the use of a force sensor set in the wrist. That is, in order to achieve a natural transfer of an object, it is necessary to be able to detect slight changes in external force. The second problem is that it is difficult to distinguish whether the force detected by the force sensor is an external force associated with the manipulator or an external force associated with the object held by the manipulator. In the force sensor set on the wrist, there is a possibility that the external force associated with the manipulator is mistakenly detected as an external force associated with the object, and it is difficult to achieve the transfer of an object that is robust to external forces. The third problem is two-state control based on a certain threshold. In the threshold-based two-state control, the manipulator keeps holding the object while maintaining a constant holding force until the person applies a certain amount of force, and the manipulator immediately releases the force under a certain threshold as a boundary. Therefore, it is expected that the person cannot predict the movement of the manipulator and will lose his / her balance.
[0060] As described above, in the existing methods, it is difficult to realize the hardware interaction function between the robot and the human.
[0061] In the present disclosure, when the gripping part of the manipulator grips an object, the slippage of the gripping part relative to the object is detected, and the gripping force of the gripping part used to grip the object is controlled according to the detected slippage. As described above, by controlling the gripping force of the gripping part when transferring the object according to the slippage of the object in the gripping part, it is possible to perform interaction between a person and the manipulator as if they were a person.
[0062] 2. More Specific Description of Transfer According to Embodiments
[0063] 2-1. Definition of each part
[0064] Next, transfer of an object between a robot and a person according to the embodiment will be described in detail. Figure 2 is a diagram schematically showing the force generated in each part when a human force acts on the object 300 held by the robot hand 11. Figure 2 In the embodiment, the manipulator 11 is in the following posture: the finger portions 13a and 13b are opened and closed along a plane horizontal to the direction of gravity Mg, and the object 300 as the holding target object is held by each fingertip portion of the finger portions 13a and 13b. For example, a person pulls up the object 300 in the direction opposite to gravity Mg with the hand 301 to receive the object from the manipulator 11.
[0065] Here, each force acting on the object 300 by the fingertip portion (elastic body 14) of the finger portion 13b is defined. By holding the object 300 with the fingertip portion, the normal force f nRi , shear force f tRi and moment n Ri Each force in acts on object 300.
[0066] Normal force f nRi is the force acting on the object 300 in the direction in which each fingertip of the finger portions 13a and 13b clamps the object 300. This normal force f nRi is regarded as the holding force used by the robot 11 to hold the object 300. The shear force f tRi is the force f acting perpendicular to the normal nRi The force in the direction of Figure 2 In the example, the shear force f tRi is the force acting on the object 300 in the vertical direction (direction of gravity Mg). Ri is with the normal force f nRi The force acting in the direction of rotation on a plane perpendicular to the direction of rotation.
[0067] Note that the normal force f nRi , shear force f tRi and moment n RiIn the example, the subscript "i" indicates each finger. That is, in Figure 2 In the example, the normal force f nR1 , shear force f tR1 and moment n R1 Each force in acts on the object 300 through the fingertip portion of the finger portion 13b. In addition, the normal force f nR2 , shear force f tR2 and moment n R2 Each force in acts on the object 300 through the fingertip portion of the finger portion 13a.
[0068] In addition, when the hand 301 holds the object 300, the normal force f nHi , shear force f tHi and moment n Hi Each of the forces f acts on the object 300. nHi It is the force acting on the object 300 in the direction in which each fingertip of the hand 301 (e.g., the thumb and each finger except the thumb) grips the object 300. The shear force f tHi is perpendicular to the normal force f nHi The force acting on the object 300 in the direction of Figure 2 In the example, the shear force f tHi is the force acting on the object 300 in the vertical direction (direction of gravity Mg). Hi is with the normal force f nHi Force acting on object 300 in a rotational direction in a vertical plane.
[0069] As mentioned above, the subscript "i" indicates each finger. That is, in Figure 2 In the example, the normal force (holding force) f nH1 , shear force f tH1 and torque nH1 Each of the forces in acts on the object 300 through the first fingertip (e.g., thumb) of the hand 301. In addition, the normal force f nH2 , shear force f tH2 and moment n H2 The various forces in act on the object 300 through the second fingertips of the hand 301 (eg, each finger except the thumb).
[0070] Here, the friction coefficient between the manipulator 11 (elastic body 14) and the object 300 is defined as the friction coefficient μ R , and define the friction coefficient between the fingers of the hand 301 and the object 300 as the friction coefficient μ H In this case, in order to hold the object 300 without slipping by the hand 301 and the robot arm 11, it is assumed that the relationship of the following equations (1) and (2) is satisfied according to the Amonton-Coulomb friction law.
[0071] μ R f nRi ≥f tRi …(1)
[0072] μ H f nHi ≥f tHi …(2)
[0073] In the embodiment, in a state where equations (1) and (2) are satisfied, the external force and the external torque (torque generated by the external force) applied to the object 300 are estimated using the initial slip detected by the tactile sensor 15 provided in each of the finger portions 13a and 13b. Then, the gripping force of the manipulator 11 is controlled so that the balance between the estimated external force and the force (gripping force) generated by the manipulator 11 and the balance between the estimated external torque and the torque generated by the manipulator 11 are always maintained.
[0074] Formula (3) represents the balance between the estimated external force and the force generated by the manipulator 11. Formula (4) represents the balance between the estimated external torque and the torque generated by the manipulator 11. That is, formula (3) represents the balance in the position direction, and formula (4) represents the balance in the rotation direction. In formulas (3) and (4), the value "i" indicates each finger as described above. The value "N" indicates the number of fingers.
[0075]
[0076]
[0077] 2-2. External force estimation method using slip sensing
[0078] Next, a method of estimating external force using slip sensing according to an embodiment will be described. Figure 3 is a diagram showing a model for describing the Hertz contact theory.
[0079] exist Figure 3 In Figures 320 and 321, the horizontal right direction in the figure is the direction of gravity Mg, and the elastic sphere 311 is Figure 1 and Figure 2 The elastic body 14 corresponds to the rigid plate 310. Figure 2 For example, the elastic sphere 311 is a sphere which faces the Figure 2 The surface of the rigid plate 310 corresponding to the object 300 in FIG. 1 is bonded to the hemisphere of the tactile sensor 15 . Figure 3 FIG. 320 schematically shows a state in which the rigid plate 310 is in contact with the elastic sphere 311 with the direction parallel to the direction of gravity Mg as a tangent.
[0080] Figure 3 FIG. 321 schematically shows the following state: the normal force f n The shear force f is applied to the rigid plate 310 in a direction perpendicular to the tangent line. t Applied in the tangential direction. Shear force f t is obtained by dividing the gravity Mg and the shear force f of hand 301 tH The combined force is expressed as f t =Mg+f tH Note that in FIG. 321 , it is assumed that the shear force f of hand 301 tH is 0.
[0081] exist Figure 3 In Figure 321, the shear force f t The elastic sphere 311 is made to t Therefore, the position of the point where the rigid plate 310 contacts the elastic sphere 311 is relative to Figure 3 The position in the left figure moves, and a displacement u is generated at the position of this point r The displacement u r For example, it occurs at the fingertips of the finger portions 13a and 13b of the manipulator 11, and is hereinafter referred to as the fingertip shear displacement u r .
[0082] Although the details will be described later, in the embodiment, the fingertip shear displacement amount u is obtained based on the detection output of the tactile sensor 15. r That is, the touch sensor 15 is used to detect (sens) the shear force f t A slip sensing sensor is provided to sense the slip between the elastic sphere 311 and the rigid plate 310 induced therefrom.
[0083] The fingertip shear displacement u r indicates the displacement in the position direction and is expressed by the following equation (5) according to the Hertz contact theory. In equation (5), R represents the radius of curvature of the elastic sphere 311, and G represents * The combined lateral elastic coefficient of the elastic sphere 311 and the rigid plate 310 is represented by E * represents the combined longitudinal elastic coefficient of the elastic sphere 311 and the rigid plate 310 (see Figure 3 left image).
[0084]
[0085] Here, the shear force f t is unknown. In addition, for example, when controlling the gripping force of the manipulator 11 (i.e., the normal force f n ) in the control unit, the normal force fn On the other hand, the radius of curvature R, the combined lateral elastic coefficient G * and the combined longitudinal elastic coefficient E * is the physical information of the elastic sphere 311 and the rigid plate 310, and is a constant. Therefore, it can be understood that the fingertip shear displacement u r Depends on the normal force f n and shear force f t Therefore, the fingertip shear displacement u r To estimate the shear force f t .
[0086] Based on the above fingertip shear displacement u r The same concept, indicating the fingertip rotation displacement u in the rotation direction theta It is expressed by the following equation (6). In this case, as described above, the fingertip rotation displacement u can be observed. theta To estimate the torque n. Similar to the fingertip shear displacement u r The fingertip rotation displacement u can also be obtained based on the detection result of the tactile sensor 15 as a slip sensing sensor. theta .
[0087]
[0088] Next, refer to Figure 2 , the external force Δf will be considered t The manipulator 11 acts on the object 300 with a normal force (gripping force) f nR When holding the object 300. The external force Δf t is, for example, the shear force Δf of the hand 301 tH .
[0089] Figure 4 It shows that the external force Δf t The situation acting on object 300 applies to the reference Figure 3 Figure 1 shows an example of a model described. Figure 4 Figure 322 schematically shows the external force Δf t (i.e., shear force f tH ) From the above Figure 3 The state shown in FIG. 321 is a state in which the rigid plate 310 is acted on in the direction opposite to the gravity Mg. Figure 2 In the example, Figure 4 The state of FIG. 322 corresponds to a state in which the hand 301, for example, pulls up the object 300 held by the robot 11. In this case, the shear force f tH Has negative value.
[0090] Relative to Figure 3 In the state shown in FIG. 321 , the elastic sphere 311 is deformed according to the shear force f tH (<0) and the state is restored. In other words, Figure 4 The fingertip shear displacement u in the left figure r is obtained by dividing the fingertip shear displacement Δu with a negative value r Add to Figure 3 The fingertip shear displacement u in the state shown in FIG. 321 r When the shear force f tH When (<0)+Mg=0, the elastic sphere 311 is in Figure 4 The state shown in FIG322, and the fingertip shear displacement u r =0.
[0091] Figure 4 Figure 323 schematically shows the shear force f tH From the above Figure 3 The state shown in FIG. 321 is a state in which the gravity Mg acts on the rigid plate 310 in the same direction. Figure 2 In the example, Figure 4 The state shown in FIG. 323 corresponds to a state in which, for example, the hand 301 pulls down the object 300 held by the robot 11. In this case, the shear force f tH Has a positive value.
[0092] Relative to Figure 3 321, the elastic sphere 311 is subjected to the applied shear force f tH (>0) and further deformation (see Figure 4 323). In this case, by dividing the shear force f tH The fingertip shear displacement Δu with a positive value r Add to Figure 3 The fingertip shear displacement u in the state shown in FIG. 321 r To obtain Figure 4 The fingertip shear displacement u in the right figure r .
[0093] In an embodiment, Figure 4 In the case shown in Figures 322 and 323, the fingertip shear displacement Δu is observed r and the fingertip rotation displacement Δu theta , and use the above equations (5) and (6) to estimate the external force Δf applied to the object 300 tH and external torque Δn H Then, based on the estimated external force Δf tH and external torque Δn H, control the gripping force of the manipulator 11 so as to always maintain a balance between various forces and moments.
[0094] Here, the fingertip shear displacement Δu r and the fingertip rotation displacement Δu theta Each of Δu corresponds to a change in the magnitude of the slip of the elastic sphere 311 relative to the object 300 (eg, a change per unit time). In other words, the fingertip shear displacement Δu r and the fingertip rotation displacement Δu theta Each of φ is a value indicating the degree of slippage of the elastic sphere 311 relative to the object 300 .
[0095] exist Figure 3 and Figure 4 In the illustrated Figures 321 to 323 , the absolute contact position moves while maintaining the relative positional relationship between the contact positions of the elastic sphere 311 and the rigid plate 310 , which is the so-called initial slip state. The slip discussed in this specification is the state of this initial slip.
[0096] 2-3. Configuration of the robot according to the embodiment
[0097] Next, the configuration of the manipulator 11 according to the embodiment will be described more specifically. Figure 5 1 is a block diagram schematically showing a configuration of a robot device to which the manipulator 11 according to the embodiment can be applied. Figure 5 In the embodiment, the robot device 1 includes a manipulator 11 and an overall control unit 20 .
[0098] The overall control unit 20 includes, for example, a central processing unit (CPU), a read-only memory (ROM), a random access memory (RAM), various interfaces, etc., and controls the overall operation of the robot device 1 using the RAM as a working memory according to a program pre-stored in the ROM. The interface transmits and receives signals between the overall control unit 20 and the manipulator 11, for example.
[0099] The manipulator 11 includes a gripping force control unit 10, a device 30, and various sensors 31. The device 30 includes, for example, an actuator for a movable part of the manipulator 11 and a drive circuit for driving the actuator, and the device performs opening / closing operations of the finger parts 13a and 13b, gripping operations, etc. according to a drive control signal.
[0100] The various sensors 31 include one or more sensors that detect a force acting on the device 30. For example, the various sensors 31 include the above-mentioned tactile sensor 15, an acceleration sensor (described later) that detects acceleration of the manipulator 11, and the like.
[0101] The grip force control unit 10 generates a drive control signal for driving the device 30 based on the control information supplied from the overall control unit 20 and the sensor information supplied from the various sensors 31 .
[0102] Figure 6 1 is a functional block diagram for explaining an example of the functions of the robot 11 according to the embodiment. Figure 6 In the figure, the manipulator 11 includes a tactile sensor 15, a device 30, a gripping force generating unit 100, a reference force generating unit 101, an adder-subtractor 102, a driving control unit 103, a calculating unit 104, a subtractor 105 and an acceleration sensor 110.
[0103] Reference Figure 5 The described holding force control unit 10 includes, for example, a holding force generation unit 100, a reference force generation unit 101, an adder-subtractor 102, a drive control unit 103, a calculation unit 104, and a subtractor 105. The holding force generation unit 100, the reference force generation unit 101, the adder-subtractor 102, the drive control unit 103, the calculation unit 104, and the subtractor 105 are implemented by a program operating on a CPU (described later) included in the holding force control unit 10. Without being limited thereto, part or all of the following may be configured by hardware circuits that operate in cooperation with each other: the holding force generation unit 100, the reference force generation unit 101, the adder-subtractor 102, the drive control unit 103, the calculation unit 104, and the subtractor 105.
[0104] Furthermore, the various sensors 31 include, for example, a tactile sensor 15 and an acceleration sensor 110. The acceleration sensor 110 is, for example, a sensor capable of detecting acceleration in each direction of three axes, and can detect the posture of the manipulator 11 and gravity Mg.
[0105] The calculation unit 104 calculates the fingertip shear displacement amount u based on the detection output of the tactile sensor 15. r and the fingertip rotation displacement u theta That is, the calculation unit 104 and the tactile sensor 15 constitute a slip detection unit that detects slip with respect to the object 300 held by the manipulator 11 .
[0106] Here, as the tactile sensor 15, for example, a pressure distribution sensor capable of detecting pressure distribution can be applied. For example, in the tactile sensor 15, a plurality of pressure sensors each detecting pressure are arranged in a matrix, and information indicating the pressure detected by each pressure sensor can be output in association with coordinate information in the matrix.
[0107] The calculation unit 104 calculates the movement amount of the center of pressure (CoP) based on the detection output provided from the tactile sensor 15. The center of pressure CoP can generally be calculated by the following equation (7). In equation (7), the variable x i represents the position of the pressure sensor in the X direction (x coordinate), the variable p(x i ) represents the position x i is the pressure at , and the variable N represents the number of distributions.
[0108]
[0109] The calculation unit 104 acquires the detection output of the tactile sensor 15 at a predetermined sampling period, performs calculation based on the above-mentioned equation (7), and obtains the center X of the acquired pressure. cop The center X of the pressure obtained in the previous calculation cop The calculation unit 104 outputs the difference as the fingertip shear displacement Δu r Note that the fingertip rotation displacement Δu can be calculated based on the change in the angle direction of the center of pressure CoP theta .
[0110] The fingertip shear displacement Δu output from the calculation unit 104 is r and the fingertip rotation displacement Δu theta Each is input to one input terminal of the subtractor 105. The fingertip shear displacement Δu r and the fingertip rotation displacement Δu theta The target value u r_ref and u theta_ref is input to the other input terminal of the subtractor 105. Here, as the target value u r_ref and u theta_ref , the value "0" is input to the other input terminal of the subtractor 105. For example, these target values u r_ref and u theta_ref The target values u are provided as information included in the control information from the overall control unit 20. Without limitation thereto, these target values u may be r_ref and u theta_ref Stored in advance in a memory or the like included in the grip force control unit 10 .
[0111] The subtractor 105 provides the gripping force generation unit 100 with the fingertip shear displacement amount Δu input to one input terminal. r From the input to the target value u of another input terminal r_ref The value obtained by subtracting .
[0112] The gripping force generation unit 100 generates a gripping force for the manipulator 11 to grip the object 300. More specifically, the gripping force generation unit 100 generates a gripping force for the manipulator 11 to grip the object 300 by applying the value supplied from the subtractor 105 to the fingertip shear displacement u of the above equation (5): r To calculate the shear force f t In addition, the gripping force generation unit 100 generates a value based on the value provided from the subtractor 105 and the shear force f t , according to the above formula (5) to calculate the normal force f n In addition, the gripping force generation unit 100 similarly uses equation (6) to calculate the fingertip rotation displacement u theta Calculate the moment n.
[0113] The normal force f calculated here is n Shear displacement Δu with fingertip r corresponds to, and is hereinafter referred to as the normal force Δf n Similarly, the torque n calculated here is related to the fingertip rotation displacement Δu theta For example, the gripping force generating unit 100 generates a torque according to the normal force Δf n More specifically, as the gripping force, the gripping force generating unit 100 calculates torque information τ0 indicating a torque for driving an actuator included in the device 30, the torque corresponding to the normal force Δf n The torque information τ0 generated by the gripping force generation unit 100 is input to one input terminal of the adder-subtractor 102 .
[0114] The reference force generating unit 101 generates a normal force f n The initial value of the force and moment n is used as the reference force. The normal force f n The initial value of the torque n is, for example, the normal force f when the robot 11 first grasps the object 300 that is the current processing target. n and moment n. In this case, the initial value is the shear force f on hand 301 t The normal force f in a state where it is not applied to the object 300 held by the robot hand 11 n And torque n. The reference force generating unit 101 is based on the normal force f n The reference force is generated by combining the torque information τ1 and the torque n. More specifically, the reference force generation unit 101 generates the torque information τ1 as the reference force.
[0115] As an example, the reference force generation unit 101 is a memory, and stores the gripping force (torque information τ0 in this example) generated by the gripping force generation unit 100 in this state as a reference force (torque information τ1 in this example). Without being limited thereto, the reference force generation unit 101 may store torque information τ, which is an actual torque used by the drive control unit 103 described later to drive the device 30, as a reference force. The reference force generated (read) by the reference force generation unit 101 is input to another input terminal of the adder-subtractor 102.
[0116] The adder-subtractor 102 subtracts or adds the reference force input from the reference force generating unit 101 and the holding force input from the holding force generating unit 100. The result of the addition or subtraction based on the reference force and the holding force performed by the adder-subtractor 102 is output as information indicating the holding force for actually driving the device 30.
[0117] Here, when the object 300 held by the manipulator 11 is transferred to the hand 301, the adder-subtractor 102 subtracts the holding force from the reference force. On the other hand, when the manipulator 11 receives the object 300 from the hand 301, the adder-subtractor 102 adds the holding force to the reference force.
[0118] For example, the sensor output of the acceleration sensor 110 may be used to determine whether the manipulator 11 transfers or receives the object 300. As an example, the gripping force generation unit 100 detects the direction of the gravity Mg based on the sensor output of the acceleration sensor 110. In addition, the gripping force generation unit 100 generates the fingertip shear displacement Δu based on the fingertip shear displacement Δu. r The positive or negative direction of the slippage occurring in the manipulator 11 is detected with respect to the object 300. As an example, it can be imagined that the gripping force generation unit 100 determines that it is reception when the slippage direction is consistent with the direction of gravity Mg, and determines that it is transfer when the slippage direction is inconsistent with the direction of gravity Mg (opposite direction).
[0119] Without being limited thereto, it may also be determined whether the manipulator 11 transfers or receives the object 300 according to an instruction from the overall control unit 20 .
[0120] The information indicating the gripping force output from the adder-subtractor 102 is provided to the drive control unit 103. The drive control unit 103 generates a drive control signal for driving the device 30 based on the provided information indicating the gripping force. For example, the drive control unit 103 generates torque information τ indicating a torque when driving an actuator included in the device 30, for example, based on the information indicating the gripping force provided from the adder-subtractor 102. The drive control unit 103 drives the device 30 based on the generated torque information τ.
[0121] Figure 7 1 is a block diagram showing an example of the hardware configuration of the manipulator 11 according to the embodiment. Figure 7 In , the control system configuration is extracted from the hardware configuration of the robot 11. Figure 7 , the manipulator 11 includes a CPU 1000 , a ROM 1001 , a RAM 1002 , a drive interface (I / F) 1010 , a sensor I / F 1011 , and a communication I / F 1012 .
[0122] The CPU 1000 controls the overall operation of the manipulator 11 by using the RAM 1002 as a working memory according to the program stored in the ROM 1001 .
[0123] The drive I / F 1010 is an interface between a drive unit (e.g., various actuators included in the device 30) and the CPU 1000. The CPU 1000 transmits a drive control signal generated according to a program to the device 30 via the drive I / F 1010. In addition, the device 30 transmits each piece of information on the operation and state of the device 30 (e.g., state information indicating the state of the device 30 itself) to the drive I / F 1010. The drive I / F 1010 receives each piece of information transmitted from the device 30, and passes the information to the CPU 1000.
[0124] The sensor I / F 1011 is an interface between the various sensors 31 and the CPU 1000. The sensor I / F receives sensor outputs output from the various sensors 31, and transfers the received sensor outputs to the CPU 1000.
[0125] The communication I / F 1012 controls the communication between the overall control unit 20 and the manipulator 11. For example, control information output from the overall control unit 20 is sent to the manipulator 11 via the communication I / F 1012. Information output from the manipulator 11 may also be sent to the overall control unit 20 via the communication I / F 1012. In addition, data of a program for operating the CPU 1000 may be sent from the overall control unit 20 to the manipulator 11 via the communication I / F 1012. In the manipulator 11, the CPU 1000 may store the program data received from the overall control unit 20 via the communication I / F 1012 in the ROM 1001. Therefore, the program stored in the ROM 1001 may be updated.
[0126] To achieve the above Figure 6The programs of the gripping force generating unit 100, the reference force generating unit 101, the adder-subtractor 102, the drive control unit 103, the calculation unit 104, and the subtractor 105 shown in the figure are provided by being stored on a computer connected to a network such as the Internet and downloaded via the network. In addition, the program can be provided or distributed via a network such as the Internet. In addition, the program can be provided by being recorded in a computer-readable recording medium (such as a compact disk (CD), a floppy disk (FD), or a digital versatile disk (DVD)) as a file in an installable format or an executable format.
[0127] Each program for realizing the gripping force generating unit 100, the reference force generating unit 101, the adder-subtractor 102, the drive control unit 103, the calculation unit 104, and the subtractor 105 has a module configuration including, for example, the gripping force generating unit 100, the reference force generating unit 101, the adder-subtractor 102, the drive control unit 103, the calculation unit 104, and the subtractor 105. As actual hardware, when the CPU 1000 reads each program from a storage medium such as the ROM 1001 and executes the program, each of the above-mentioned units or components is loaded on a main storage device such as the RAM 1002, and each of these units or components is generated on the main storage device.
[0128] 2-4. Control of Transfer Operation According to Embodiment
[0129] Next, control of a transfer operation of delivering the object 300 from the manipulator 11 to the hand 301 according to the embodiment will be described in more detail. Figure 8 is a flowchart schematically showing an example of control of a transfer operation according to the embodiment.
[0130] Assuming that according to Figure 8 Before the processing of the flowchart starts, the manipulator 11 holds the object 300 vertically (in a direction parallel to the gravity Mg). In this state, the following situation will be considered: the hand 301 pulls up the object 300, for example, in a direction opposite to the direction of gravity Mg, and receives the object from the manipulator 11.
[0131] The gripping force control unit 10 continuously observes the slippage of the finger portions 13a and 13b at each fingertip portion (step S10). More specifically, the gripping force control unit 10 continuously observes the fingertip shear displacement amount Δu as information indicating the slippage. r and the fingertip rotation displacement Δu theta With the target value u r_ref and u theta_ref The difference between the two is calculated and the slip is detected.
[0132] In the following, for the convenience of explanation, unless otherwise specified, the fingertip shear displacement Δu r With the target value u r_ref The difference between (=0) is set as the fingertip shear displacement Δu r Similarly, unless otherwise stated, the fingertip is rotated by the displacement Δu theta With the target value u theta_ref The difference between (=0) is set as the fingertip rotation displacement Δu theta .
[0133] The holding force control unit 10 detects the triggering of the transfer operation based on the slip detection result (step S11). In the next step S12, the holding force control unit 10 controls the holding force in the transfer operation of transferring the object 300 to the hand 301. In the next step S13, the holding force control unit 10 checks whether the transfer operation has been stabilized by the control of the holding force in step S12. When the stability of the transfer operation is confirmed, the holding force control unit 10 releases the manipulator 11. Therefore, the object 300 held by the manipulator 11 is transferred to the hand 301, and the transfer operation is completed.
[0134] Fig. 9 is a flowchart showing an example of a trigger detection operation of a transfer operation according to an embodiment. Fig. 9 The process of the flowchart corresponds to Figure 8 The processing of step S11 in .
[0135] exist Fig. 9 In step S110 , the grip force control unit 10 determines whether the detected slip is a slip caused by interference.
[0136] As an example, the grip force control unit 10 may detect the slip speed and determine whether the slip is caused by interference based on the slip speed. In this determination, for example, the fingertip shear displacement amount Δu r When another threshold is exceeded, it can be determined that the detected slip is a slip due to disturbance, wherein the other threshold is set to a value larger as an absolute value with respect to the above threshold in the case of performing slip detection.
[0137] As another example, the gripping force control unit 10 may determine whether the detected slip is a slip due to interference by using machine learning, such as long short-term memory (LSTM). In this case, it is conceivable that the determination is performed by using learning data obtained by inputting time series data of position information obtained based on the output of the tactile sensor 15, and learning whether there is interference as output data.
[0138] When the gripping force control unit 10 determines that the detected slip is a slip caused by interference (step S110, "Yes"), the process returns to step S110. On the other hand, when the gripping force control unit 10 determines that the detected slip is not a slip caused by interference (step S110, "No"), the process proceeds to step S111.
[0139] In step S111, the gripping force control unit 10 detects the direction of gravity Mg. The direction of gravity Mg may be detected based on the sensor output of the acceleration sensor 110. Without limitation thereto, the direction of gravity Mg may be detected by predicting the current posture of the manipulator 11 based on the history of operations such as movement and rotation of the manipulator 11.
[0140] In the next step S112, the gripping force control unit 10 determines whether a slip in a direction different from the direction of the gravity Mg detected in step S111 has been detected. r When the absolute value of exceeds the threshold value, the grip force control unit 10 determines that slippage has been detected. In addition, when slippage has been detected, the grip force control unit 10 determines that the fingertip shear displacement Δu r Whether the direction of slip is different from the direction of gravity Mg is determined by the positive or negative value of .
[0141] When it is determined that no slippage in a direction different from the direction of gravity Mg is detected (step S112, "No"), the gripping force control unit 10 returns the process to step S112. On the other hand, when it is determined that slippage in a direction different from the direction of gravity Mg is detected, the gripping force control unit 10 determines that a trigger for a transfer operation is detected, and ends the process according to step S113. Fig. 9 Then, the process proceeds to Figure 8 Processing of step S12.
[0142] Fig.10 is a flowchart showing an example of an operation of controlling the transfer of grip force according to an embodiment. Fig.10 The process of the flowchart corresponds to Figure 8 The processing of step S12 in .
[0143] In step S120, the gripping force control unit 10 reduces the gripping force based on the slip detection result. r and the fingertip rotation displacement Δu theta When at least one of (each absolute value) has an absolute value exceeding 0, the gripping force control unit 10 makes the gripping force (normal force f nR ) is reduced by a predetermined amount.
[0144] That is, in this example, the hand 301 pulls up the object 300 in the direction opposite to the direction of the gravity Mg to receive the object 300 from the manipulator 11. In this case, it is considered that the slippage generated between the manipulator 11 and the object 300 indicates a state such as the following: when the hand 301 tries to pull up the object 300 held by the manipulator 11, the hand 301 feels a slight resistance to the pulling up of the object 300.
[0145] When the gripping force is reduced by a predetermined amount in step S120, the gripping force control unit 10 ends the Fig.10 A series of processes in the flowchart in the figure, and transfer the process to Figure 8 Step S13 in .
[0146] Fig.11 is a flowchart showing an example of an operation of stability confirmation of a transfer operation according to an embodiment. Fig.11 The process of the flowchart corresponds to Figure 8 The processing of step S13 in .
[0147] In step S130, the gripping force control unit 10 converts the gripping force (normal force f nR ) is reduced by a predetermined amount. For example, it can be imagined that the degree of reduction in the gripping force is less than the above Fig.10 In the next step S131, the gripping force control unit 10 performs the same Fig.10 In a similar manner to step S120 in FIG. 1 , it is determined whether slip has occurred.
[0148] In step S131, in the above Fig.10 After the gripping force is reduced in step S120 in the flowchart of FIG. 1 , the slip is determined in a state where the gripping force is further reduced in step S130. Therefore, since the gripping force of the manipulator 11 cannot resist the shear force f of the hand 301, tH and the shear force f tH This slip occurs because the force in the direction of gravity Mg is obtained by combining the gravity Mg in the opposite directions.
[0149] When it is determined that slippage has occurred (step S131, "Yes"), the gripping force control unit 10 transfers the process to step S132 to make the gripping force (normal force f nR ) is returned to the original value. For example, the gripping force control unit 10 returns the gripping force to the state immediately before step S130. When the gripping force is restored in step S132, the gripping force control unit 10 returns the process to step S130.
[0150] On the other hand, when it is determined in step S131 that no slippage has occurred (step S131, "No"), the gripping force control unit 10 ends the operation according to Fig.11 Then, the gripping force control unit 10 transfers the processing to Figure 8 Step S14 in the flowchart is performed to release the robot 11.
[0151] As described above, in the embodiment, the manipulator 11 detects slippage at the fingertips holding the object 300, and based on the detected slippage, controls the holding force for holding the object 300. At this time, in the embodiment, since the slippage is detected by the tactile sensor 15 provided at the fingertips, a slight change in the external force can be detected, and a more natural transfer of the object 300 from the manipulator 11 to the hand 301 can be achieved.
[0152] Furthermore, in the embodiment, by using the tactile sensor 15 provided at the fingertip portion of the manipulator 11, it is possible to easily distinguish whether the detected external force is an external force applied to the manipulator 11 or an external force applied to the object 300 held by the manipulator 11. Therefore, transfer of an object having high robustness to external force can be achieved.
[0153] Furthermore, in the embodiment, the gripping force (release force) can be appropriately reduced in the case where the external force applied to the object 300 is fed back according to the slippage detected by the tactile sensor 15 provided at the fingertip portion of the manipulator 11. Therefore, the following situations can be avoided: the object 300 to be transferred is damaged due to excessive force applied to the object 300, the object 300 is dropped due to too small a force to hold the object 300, etc.
[0154] 2-5. Operation in any transfer direction according to the embodiment
[0155] In the above description, the transfer control has been described in the case where the direction of the shear force of the manipulator 11 is consistent with the direction of the gravity Mg. Without being limited to this, the transfer control according to the embodiment is established in an arbitrary direction. Hereinafter, the following transfer control will be described: (1) The shear force f of the manipulator 11 is consistent with the direction of the gravity Mg. t The transfer control when the direction is consistent with the direction of gravity Mg, and (2) the shear force f of the robot 11 t Transfer control when the direction is inconsistent with the direction of gravity Mg.
[0156] Note that, although the details will be described later, in order to establish the normal force f in any holding posture n In order to balance the forces in the direction of the gripping force, in addition to the above-mentioned control of the gripping force, it is preferable to impose position restrictions.
[0157] First, the shear force f of the robot 11 in (1) will be described.t Transfer control when the direction is consistent with the direction of gravity Mg. Fig. 12A and Fig. 12B is a schematic diagram showing the shear force f of the robot 11 t The figure shows an example of a case where the direction of is consistent with the direction of gravity Mg. In other words, Fig. 12A and Fig. 12B In the figure, the normal force f of the manipulator 11 is nR and the normal force f when the hand 301 holds the object 300 nH Perpendicular to the direction of gravity Mg.
[0158] exist Fig. 12A In the example of FIG. 1 , similar to the above description, the finger portions 13a and 13b of the manipulator 11 hold the object 300 in the vertical direction. When the hand 301 pulls up the object 300 in the upward direction in the figure, the transfer of the object 300 from the manipulator 11 to the hand 301 is performed. Fig. 12B In the example of , the hand 301 pulls the object 300 upward to the upper right in the figure, so that the robot 11 holds the object 300 from the upper left in the figure. When the hand 301 pulls the object 300 upward to the upper right in the figure, the transfer of the object 300 from the robot 11 to the hand 301 is performed.
[0159] Observe the shear force Δf on the hand 301 tH The fingertip shear displacement Δu when an external force is applied to the object 300 r Then, consider preventing the fingertip shear displacement Δu r The normal force Δf required for the manipulator 11 of the robot nR In this case, the shear force f of the robot 11 tR It is expressed by the following formula (8).
[0160]
[0161] Here, it is assumed that the hand 301 generates the following gripping force (normal force) Δf nH , the holding force (normal force) Δf nH Only the external force Δf applied to the object 300 is generated tH And the following formula (9) is satisfied. In this case, it is assumed that a shear force Δf that satisfies the following formula (10) and is equal to the shear force Δf can also be generated in the direction of the gravity Mg tH The force Δf gH .
[0162] μ H Δf nH ≥Δf tH …(9)
[0163] μH Δf nH ≥Δf gH …(10)
[0164] At this time, the shear force f of the hand 301 is expressed by the following formula (11): tH .
[0165]
[0166] Substituting equations (8) and (11) into the above equation (3) and taking the force balance into consideration, it is confirmed that the force balance in the shear direction is maintained as shown in the following equation (12).
[0167]
[0168] Note that the balance of force in the normal force direction is offset by the force applied by each finger, and its description will be omitted here.
[0169] Next, the shear force f of the robot 11 in (2) will be described. t Transfer control when the direction is inconsistent with the direction of gravity Mg. Fig.13A and Fig. 13B is a schematic diagram showing the shear force f of the robot 11 t A diagram showing an example of a case where the direction of is inconsistent with the direction of gravity Mg.
[0170] exist Fig.13A In the figure, the normal force f of the manipulator 11 is nR and the normal force f when the hand 301 holds the object 300 nH The direction of the object 300 is consistent with the direction of gravity Mg. When the hand 301 pulls the object 300 in the horizontal direction in the figure, the transfer of the object 300 from the manipulator 11 to the hand 301 is performed. Fig. 13B In the equation, the normal force f nR and f nH The direction of the object 300 has an angle (≠0°) with respect to the direction of gravity Mg. When the hand 301 pulls the object 300 diagonally upward to the right in the figure, the object 300 is transferred from the robot 11 to the hand 301.
[0171] Here, for the sake of explanation, Fig.13A The status shown is as an example.
[0172] Observe the shear force f on hand 301 tH The fingertip shear displacement Δu when an external force is applied to the object 300 r Then, consider preventing the fingertip shear displacement Δu r The normal force Δf required for the robot's manipulator 11 nRIn this case, the normal force f of the manipulator 11 nR and shear force f tR They are expressed by the following equations (13) and (14) respectively.
[0173]
[0174]
[0175] At this time, the normal force f nH and shear force f tH They are expressed as equations (15) and (16) below respectively.
[0176]
[0177]
[0178] When equations (13) to (16) are substituted into the above equation (3) and the force balance is considered, the following equations (17) and (18) are obtained respectively. Note that "0 (?)" at the end of equation (17) indicates that it is unknown whether the value "0" is obtained as the result of equation (17).
[0179]
[0180]
[0181] In this case, the normal force Δf of the manipulator 11 is nR and the normal force Δf of the hand 301 nH The gripping forces satisfying the following equations (19) and (20) are generated respectively.
[0182] μ R Δf nR ≥f tH …(19)
[0183] μ H Δf nH ≥f tH …(20)
[0184] However, the normal force Δf of the manipulator 11 cannot be guaranteed nR Equal to the normal force Δf of hand 301 nH (Δf nR =Δf nH ). If the friction coefficients are consistent with each other (μ R =μ H), and the manipulator 11 and the hand 301 can hold the object 300 with force just before the object 300 slides out, then the force balance in the normal direction is established. However, it can be understood that unless the consistency of these friction coefficients is satisfied and the manipulator 11 and the hand 301 hold the object 300 with force just before the object 300 slides out, the right side of the formula (17) = 0 is not established.
[0185] Note that it is confirmed that the balance of forces in the shear direction is established, as indicated by the above equation (18).
[0186] As described above, depending on the gripping posture of the robot 11, the shear force f tR The direction of may be inconsistent with the direction of gravity Mg. In this case, it can be seen that the force balance in the shear direction is established, but the force balance in the normal direction cannot be guaranteed.
[0187] Examples of a case where the balance of the force in the normal direction is not established include a case where the friction coefficient μ between the manipulator 11 and the object 300 is R Greater than the friction coefficient μ between the fingers of the hand 301 and the object 300 H (μ R >μ H ). In this case, the gripping force (normal force) Δf of the robot 11 nR becomes smaller than the gripping force Δf of the hand 301 nH (Δf nR <Δf nH ). In this case, the force provided by the manipulator 11 and the hand 301 increases relative to the weight of the object 300, and the object 300 moves in the direction opposite to the gravity Mg. This means that the deviation δ between the positions of the fingers of the manipulator 11 increases, as shown in the following equation (21). Fig.14 As schematically shown, the position deviation δ is related to the degree θ of the opening of the fingers 13a and 13b relative to the symmetry center C. a and θ b The difference between .
[0188]
[0189] Therefore, in addition to the force control of the algorithm, position control is applied so that the position deviation δ between the finger portion 13a and the finger portion 13b of the manipulator 11 becomes 0. This makes it possible to achieve both balance of force in the shear direction and balance of force in the normal direction.
[0190] When the manipulator 11 transfers the object 300 to the hand 301, it is possible to estimate to some extent from which direction the hand 301 receives the object 300 based on the holding position of the manipulator 11 obtained by the tactile sensor 15. However, strictly speaking, it is not known from which direction the hand 301 receives until the hand 301 starts the receiving operation. In the transfer control according to the embodiment, as described above, a transfer operation in an arbitrary direction is possible, and a more flexible transfer operation can be achieved.
[0191] 2-6. Application Examples of Control According to Embodiment
[0192] Note that in the above description, the control according to the embodiment has been described as being applied to the transfer of the object 300, but it is not limited to this example. As an application example of the control according to the embodiment, the control can also be applied to a floor placement operation requiring a release force.
[0193] As a control method for the floor placement operation, unlike the transfer operation, it is conceivable to apply a simple control method by an existing method in which a threshold equivalent to the weight of the object 300 is determined and the gripping force is reduced. However, in the control of the existing method, the object 300 to be held is limited to a rigid body. In the case where the object 300 to be held is a flexible object that is deformed or damaged by applying an excessive force, the object itself can be prevented from being damaged when placed on the floor by applying a gripping force control according to the embodiment so that a balance of force and a balance of moment are established in consideration of the magnitude of the external force.
[0194] 3. First Modification of Embodiment
[0195] Next, a first modification of the embodiment will be described. In the above embodiment, the control according to the present disclosure is applied to a transfer operation of transferring the object 300 held by the manipulator 11 to the hand 301. In the first modification of the embodiment, the control according to the present disclosure is applied to a receiving operation of the manipulator 11 receiving the object 300 from the hand 301.
[0196] Fig.15 1 is a flowchart schematically showing an example of control of a receiving operation according to a first modification of the embodiment. Here, for explanation, in the manipulator 11, the positions of the finger portions 13a and 13b are controlled in advance so that the object 300 to be received contacts the elastic body 14 provided at the fingertip portion of each of the finger portions 13a and 13b. In this state, the following case will be considered: the hand 301 slides the object 300 between the fingertip portion of the finger portion 13a and the fingertip portion of the finger portion 13b, for example, to transfer the object 300 to the manipulator 11.
[0197] The grip control unit 10 continuously observes the slippage at each fingertip of the finger portions 13a and 13b (step S20). When a trigger of a transfer operation is detected based on the slippage detection result (step S21), the grip control unit 10 transfers the process to the next step S22 and starts a receiving operation of the object 300.
[0198] In the process of step S21, for example, similar to reference Fig. 9 In the control described in the flowchart, the holding force control unit 10 uses the slip speed and machine learning to determine whether the detected slip is a slip caused by interference. In the case where it is determined that the detected slip is not a slip caused by interference, the holding force control unit 10 detects the direction of gravity Mg. Then, the holding force control unit 10 determines whether a slip in the same direction as the detected direction of gravity Mg has been detected. In the case where it is determined that a slip in the same direction as the direction of gravity Mg has been detected, the holding force control unit 10 determines that a trigger for receiving an operation has been detected, and transfers the processing to step S22.
[0199] In step S23, the gripping force control unit 10 slightly reduces the gripping force to generate slip. In the next step S24, the gripping force control unit 10 determines whether slip has been detected. When the gripping force control unit 10 determines that slip has been detected (step S24, "yes"), the process proceeds to step S25.
[0200] In step S25, the gripping force control unit 10 slightly increases the gripping force. For example, the gripping force control unit 10 increases the gripping force by a degree slightly smaller than the degree by which the gripping force was decreased in step S23. When the gripping force is increased in step S25, the gripping force control unit 10 returns the process to step S23.
[0201] On the other hand, when the grip force control unit 10 determines that no slippage is detected in step S24 (step S24 , “NO”), the process proceeds to step S26 , and reception of the object 300 from the hand 301 is deemed to be completed.
[0202] As described above, in the first modification of the embodiment, the manipulator 11 detects slippage at the fingertips holding the object 300, and based on the detected slippage, controls the holding force for holding the object 300. At this time, in the embodiment, since the slippage is detected by the tactile sensor 15 provided at the fingertips, a slight change in the external force can be detected, and the operation of the manipulator 11 receiving the object 300 from the hand 301 can be achieved by a more natural operation.
[0203] 4. Second Modification of Embodiment
[0204] Next, a second modification of the embodiment will be described. In the second modification of the embodiment, in addition to the tactile sensor 15, an image sensor that detects visual information is used to control the transfer operation.
[0205] Fig.16 is a block diagram schematically showing a configuration of a robot device 1 to which a manipulator 11 ′ according to a second modification of the embodiment is applied. Fig.16 In the embodiment, by adding the image sensor 32 and the image processing unit 33 to the reference Figure 5 The robot 11 of the described embodiment is used to obtain a robot 11 ′.
[0206] The image sensor 32 includes a general image sensor (e.g., a complementary metal oxide semiconductor (CMOS) image sensor or a charge coupled device (CCD) image sensor), and a driving circuit for driving the image sensor. For example, the image sensor 32 is arranged so that at least the fingertips of the finger portions 13a and 13b of the manipulator 11' are included in the imaging range. The captured image data output from the image sensor 32 is provided to the image processing unit 33.
[0207] The image processing unit 33 performs predetermined image processing on the captured image data provided from the image sensor 32. For example, the image processing unit 33 performs image recognition processing for recognizing a predetermined object included in the captured image data. The image processing unit 33 provides the result of performing the image processing on the captured image data to, for example, the grip force control unit 10.
[0208] As an example, the holding force control unit 10 detects the state of transfer of the object 300 and information about the transfer partner based on visual information (captured image data in this example) acquired by the image sensor 32 through image processing by the image processing unit 33. The holding force control unit 10 combines the information detected based on the visual information, and, for example, performs control to multiply the holding force to be reduced in the transfer operation by a variable parameter. For example, in the case where a child receives the object 300 in an unstable situation, a margin for force balance and moment balance can be provided by applying a variable parameter for reducing the holding force to be reduced. Regarding the setting of the variable parameter, for example, the following method can be conceived: in the image processing unit 33, machine learning is used with visual information as input and the safety of the transfer as output.
[0209] The image processing unit 33 supplies the set variable parameter as a result of the image processing to the grip force control unit 10. The variable parameter may be set in the grip force control unit 10 based on the output of the image processing unit 33.
[0210] In addition, by predicting in advance the direction in which the other party receives the object 300 by transferring it based on the visual information acquired by the image sensor 32, and matching the direction with the sliding direction, the gripping force can be controlled by further utilizing the sliding direction. For example, in the case where it is known based on the visual information that the direction in which the hand 301 receives the object 300 by transferring it is a vertical downward direction, the robustness can be improved by reducing the gripping force only with respect to the sliding in the vertical downward direction.
[0211] In addition, the scene or user information that can be obtained based on visual information is combined with Figure 8 The transfer operation trigger detection process in step S11 in the above method can be used in combination to more robustly detect the transfer operation trigger. For example, it can be recognized based on visual information that the hand 301 is approaching the robot 11', and it can be determined that the transfer operation trigger is detected when the hand 301 is approaching within a predetermined distance.
[0212] Without limitation to this, it is also possible to acquire scene or user information using auditory information in addition to visual information, and detect the trigger of the transfer operation based on the acquired information. For example, by further adding a microphone and a voice processing unit that performs voice processing on a voice signal output from the microphone to the Fig.16 The output of the speech processing unit is provided to the gripping force control unit 10. The speech processing unit analyzes the speech signal based on the sound collected by the microphone, obtains the meaning of the speech, and can determine whether the content of the speech is a request to transfer the object 300 from the manipulator 11' to the hand 301.
[0213] In addition, the direction in which the hand 301 receives the object 300 can be predicted based on the visual information and the auditory information. By matching the predicted receiving direction with the sliding direction, the trigger of the receiving operation can be detected more robustly. For example, in the case where it is known that the direction in which the hand 301 performs the reception is the vertical downward direction, even in the case of detecting a sliding in the same direction as the direction of gravity Mg, it can be determined that the receiving operation trigger is detected.
[0214] As described above, by performing trigger detection of the transfer operation and the reception operation by combining the detection result of the tactile sensor 15 with visual information or auditory information, trigger detection with higher robustness can be achieved.
[0215] 5. Third Modification of Embodiment
[0216] Next, a third modification of the embodiment will be described. The third modification of the embodiment is the following example: Figure 5Among the various sensors 31, in addition to the tactile sensor 15, a force sensor is also provided on, for example, the wrist 12 of the manipulator 11, and the sensor outputs of the tactile sensor 15 and the force sensor are used to control the transfer operation.
[0217] The gripping force control unit 10 uses the sensor output of the force sensor provided on the wrist portion 12 and the sensor output of each tactile sensor 15 provided on the fingertip portion of each of the finger portions 13a and 13b. This makes it possible to distinguish between the external force applied to the manipulator 11 and the external force applied to the object 300 gripped by the manipulator 11. As an example, it is conceivable to perform threshold determination on the sensor output of the force sensor, and determine that the external force is applied to the manipulator 11 itself when a force exceeding the threshold is detected in the force sensor.
[0218] Therefore, when an external force is applied to pull the manipulator 11, control can be performed to move the manipulator 11 (or the robot device 1) in the pulling direction without releasing the held object 300. In addition, in the case where the object 300 is pulled, the manipulator 11 (or the robot device 1) can be controlled to transfer the object 300 without moving.
[0219] Note that the effects described in this specification are merely examples and not limitations, and other effects may be provided.
[0220] 6. Other configurations
[0221] Note that the present technology may also have the following configurations.
[0222] (1) A control device comprising:
[0223] a slippage detection unit that detects slippage of a target object held by the holding portion; and
[0224] a control unit configured to control a holding force of the holding portion for holding the target object based on the slip detected by the slip detection unit,
[0225] The control unit estimates an external force applied to the target object held by the holding portion based on the slip detected by the slip detection unit, and controls the holding force based on the estimated external force.
[0226] (2) The control device according to (1) above, wherein
[0227] The slip detection unit detects a shear displacement amount generated relative to the grip portion as the slip, and
[0228] The control unit estimates the external force based on the shear displacement amount.
[0229] (3) The control device according to (1) above, wherein
[0230] The control unit also estimates a moment applied to the target object held by the holding portion based on the slip detected by the slip detection unit, and also controls the holding force by using the estimated moment.
[0231] (4) The control device according to (3) above, wherein
[0232] The slip detection unit detects a rotation displacement amount generated relative to the grip portion as the slip, and
[0233] The control unit estimates the torque based on the rotational displacement amount.
[0234] (5) The control device according to any one of (1) to (4) above, wherein the grip portion comprises:
[0235] an elastic body that contacts the target object and generates friction force on the target object; and
[0236] a sensor that detects a force generated in the elastic body according to the friction force, and
[0237] The slip detection unit detects the slip based on a detection output of the sensor.
[0238] (6) The control device according to (5) above, wherein
[0239] The sensor is a pressure distribution sensor that detects pressure distribution on a two-dimensional surface, and
[0240] The slip detection unit detects the slip based on a movement of a center position of the pressure distribution detected by the pressure distribution sensor.
[0241] (7) The control device according to any one of (1) to (6) above, wherein
[0242] The control unit controls a gripping force for transferring the target object gripped by the gripping portion, in a case where the slip detected by the slip detection unit is not slip caused by disturbance and is slip in a direction different from a gravity direction.
[0243] (8) The control device according to (7) above, wherein
[0244] The control unit performs control of reducing a holding force for holding the target object according to the slip detected by the slip detection unit.
[0245] (9) The control device according to (8) above, wherein
[0246] The control unit also performs control to reduce a holding force used by the holding portion to hold the target object, and performs control to return the holding force to the holding force immediately before the control when the slip detection unit detects a slip of a predetermined level or more.
[0247] (10) The control device according to any one of (1) to (9) above, wherein
[0248] The control unit performs control of the gripping force based on visual information in a range including at least the grip portion and the slip detected by the slip detection unit.
[0249] (11) The control device according to (10) above, wherein
[0250] The control unit predicts a direction in which the gripping portion transfers the target object based on the visual information, and performs control of the gripping force according to the predicted direction.
[0251] (12) The control device according to any one of (1) to (11) above, wherein
[0252] The control unit also performs control of the holding force by using a detection result of a force detected at a holding portion holding the holding portion.
[0253] (13) A control method comprising:
[0254] a slippage detection step of detecting, by a slippage detection unit, a slippage of the target object held by the holding portion; and
[0255] a control step, wherein the control unit controls the gripping force of the gripping portion for gripping the target object based on the slip detected in the slip detection step, wherein
[0256] The controlling step includes estimating an external force applied to the target object held by the holding portion based on the slip detected in the slip detecting step, and controlling the holding force based on the estimated external force.
[0257] (14) A control program causing a processor to execute the following steps:
[0258] a slippage detection step of detecting a slippage of a target object held by the holding portion; and
[0259] a control step, which controls the gripping force of the gripping portion for gripping the target object based on the slip detected in the slip detection step, wherein
[0260] The controlling step includes estimating an external force applied to the target object held by the holding portion based on the slip detected in the slip detecting step, and controlling the holding force based on the estimated external force.
[0261] Reference numerals list
[0262] 1Robotic Device
[0263] 10 Grip force control unit
[0264] 11 Robot
[0265] 12 Wrist
[0266] 13a, 13b finger part
[0267] 14 Elastomer
[0268] 15Tactile Sensor
[0269] 20 Overall controller
[0270] 30 devices
[0271] 31 Various sensors
[0272] 32 Image Sensor
[0273] 33 Image processing unit
[0274] 100 Grip Force Generator Units
[0275] 101 Reference force generation unit
[0276] 102 Adder-Subtractor
[0277] 103 drive control unit
[0278] 104 computing units
[0279] 105 Subtractor
[0280] 110 Acceleration sensor
[0281] 300 objects
[0282] 301 lots
[0283] 1000 CPU
Claims
1. A control device comprising: a slippage detection unit that detects slippage of a target object held by the holding portion; as well as a control unit configured to control a gripping force of the gripping portion for gripping the target object based on the slip detected by the slip detection unit, wherein the control unit estimates an external force applied to the target object held by the holding portion based on the slip detected by the slip detection unit, and controls the holding force based on the estimated external force, wherein, when the slip detected by the slip detection unit is not a slip caused by interference and is a slip in a direction different from the direction of gravity, the control unit controls the gripping force for transferring the target object gripped by the gripping portion, wherein controlling the gripping force for transferring the target object gripped by the gripping portion includes: performing control to reduce the gripping force for gripping the target object according to the slip detected by the slip detection unit, and performing control to return the gripping force to the gripping force immediately before the control when the slip detection unit detects a slip of a predetermined level or above, and The control unit further predicts a direction in which the gripping portion transfers the target object based on visual information in a range including at least the gripping portion, and controls the gripping force by matching the predicted direction with a direction of slippage.
2. The control device according to claim 1, wherein The slip detection unit detects a shear displacement amount generated relative to the grip portion as the slip, and The control unit estimates the external force based on the shear displacement amount.
3. The control device according to claim 1, wherein The control unit also estimates a torque applied to the target object gripped by the gripping portion based on the slip detected by the slip detection unit, and also controls the gripping force by using the estimated torque.
4. The control device according to claim 3, wherein The slip detection unit detects a rotational displacement amount generated relative to the grip portion as the slip, and The control unit estimates the torque based on the rotational displacement amount.
5. The control device according to claim 1, wherein: The gripping portion comprises: an elastic body that contacts the target object and generates frictional force on the target object; and a sensor that detects a force generated in the elastic body according to the friction force, and The slip detection unit detects the slip based on a detection output of the sensor.
6. The control device according to claim 5, wherein The sensor is a pressure distribution sensor that detects pressure distribution on a two-dimensional surface, and The slip detection unit detects the slip based on a movement of a center position of the pressure distribution detected by the pressure distribution sensor.
7. The control device according to claim 1, wherein The control unit also performs control of the gripping force by using a detection result of a force detected at a holding portion that holds the gripping portion.
8. A control method comprising: a slippage detection step of detecting slippage of the target object held by the holding portion by a slippage detection unit; as well as a control step of controlling the gripping force of the gripping portion for gripping the target object based on the slip detected in the slip detection step by a control unit, wherein The control step comprises: estimating an external force applied to the target object held by the holding portion based on the slip detected in the slip detecting step, and controlling the holding force based on the estimated external force, wherein, in a case where the slip detected in the slip detection step is not a slip caused by interference and is a slip in a direction different from the direction of gravity, controlling the gripping force for transferring the target object gripped by the gripping portion, wherein controlling the gripping force for transferring the target object held by the gripping portion includes: performing control to reduce the gripping force for holding the target object according to the slip detected in the slip detection step, and performing control to return the gripping force to the gripping force immediately before the control when a slip of a predetermined level or above is detected in the slip detection step, and Wherein, in the control step, the direction in which the gripping portion transfers the target object is predicted based on visual information in a range including at least the gripping portion, and the gripping force is controlled by matching the predicted direction with the direction of slippage.
9. A computer-readable recording medium storing a control program, wherein the control program causes a processor to execute the following steps: a slippage detecting step of detecting slippage of a target object held by the holding portion; and a control step of controlling the gripping force of the gripping portion for gripping the target object based on the slip detected in the slip detection step, wherein The control step comprises: estimating an external force applied to the target object held by the holding portion based on the slip detected in the slip detecting step, and controlling the holding force based on the estimated external force, wherein, in a case where the slip detected in the slip detection step is not a slip caused by interference and is a slip in a direction different from the direction of gravity, controlling the gripping force for transferring the target object gripped by the gripping portion, wherein controlling the gripping force for transferring the target object held by the gripping portion includes: performing control to reduce the gripping force for holding the target object according to the slip detected in the slip detection step, and performing control to return the gripping force to the gripping force immediately before the control when a slip of a predetermined level or above is detected in the slip detection step, and Wherein, in the control step, the direction in which the gripping portion transfers the target object is predicted based on visual information in a range including at least the gripping portion, and the gripping force is controlled by matching the predicted direction with the direction of slippage.
