JP2006297542A - Slip detection device on finger surface of robot hand - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a slip detection device capable of determining a slip of a gripped object on a finger surface of a robot hand adopting a soft structure on a gripping face. <P>SOLUTION: A tactile sensor 18 comprising a plurality of sensors (for example, pressure sensors) is arranged on a finger surface having a soft structure 10s, to measure a normal contact force Fn acting on a finger tip at the time of starting to grip a gripped object 200, and an initial pressure gravity center position CO. When the gripped object 200 is raised, the soft structure 10s is deformed according to an external force (gravity) F acting on the gripped object 200, so that a pressure gravity center position C is moved. The generation of the slip is determined and estimated on the basis of the difference between the positions C and CO, namely a moving distance S of the gravity center position. At the time of slippage, a gripping force is increased. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an apparatus for detecting slippage on a finger surface of a robot hand, and more particularly to an apparatus for detecting slippage between a finger and an object when the object is gripped on the finger surface of a robot hand having a flexible structure.

  Research is being conducted on a robot hand that grips and conveys an object with unknown characteristics, as well as an object with known characteristics such as weight. In particular, when gripping an object with unknown weight, surface friction coefficient, or strength, the minimum is necessary to grip the object reliably without being damaged or deformed by the gripping force and without dropping. Gripping must be performed with a limited gripping force. For this reason, a technique for appropriately grasping an object whose characteristics are unknown has been studied (see Patent Document 1).

In the technique described in Patent Document 1, the object gripping surface of a robot hand is configured by a curved elastic body, and a plurality of sensors are arranged therein, and shear deformation or shearing of the elastic body obtained from the sensor output is performed. By obtaining the size of the fixed area between the elastic body and the object from the stress and controlling the gripping force etc. accordingly, the gripping force is too strong to crush the object, and the gripping force is The object can be gripped with an appropriate gripping force without being too weak to slip the object.
JP 2000-254884 A

  However, the sensor for detecting shear strain and shear stress used in this technology has a complicated structure and can be relatively large. Also, especially when the gripping surface is made of a flexible structure, the shape of the flexible structure of the gripping surface changes when the object is lifted after gripping, but this strain change detects shear strain and shear stress. This is difficult to distinguish from shear strain and shear stress due to sliding of the object, and may be erroneously detected.

  Therefore, an object of the present invention is to provide a slip detection device that can determine the slip of a gripped object on a finger surface in a robot hand that employs a flexible structure on a grip surface.

  In order to solve the above problems, a slip detection device according to the present invention is a device for detecting a fingertip slip in a robot hand having a finger surface with a flexible structure, wherein the fingertip curved surface is divided into a plurality of regions. Distributed pressure sensors that detect the normal force in each direction, means for calculating the output center of gravity position and normal contact force from the output of these distributed pressure sensors, the relationship between the movement of the output center of gravity position and the normal contact force And means for determining occurrence of slippage.

  Compared to the strain sensor, the pressure sensor has a simple configuration and can be easily downsized. When a stationary object is gripped by a finger having a flexible structure on the surface, the flexible structure is pressed against the object surface by the gripping force and deforms. Therefore, compared with the case where the contact is made on the solid surface, the object comes in contact with a wider range and the reaction force corresponding to the gripping force is detected by the distributed pressure sensor. When the grasped object is lifted from this state, the flexible structure grasping the object surface is displaced downward and deformed to support the load of the object. For this reason, the pressure distribution applied to the flexible structure changes from when it is gripped. Further, when the object cannot be grasped completely and the object slips on the finger surface, the pressure distribution changes according to the slip. Therefore, in the present invention, it is determined whether the deformation of the flexible structure that does not cause slipping or the occurrence of slipping from the moving state of the output center of gravity position and the normal contact force (corresponding to the gripping force) in the change state of the pressure distribution. judge.

  Preferably, the means for determining the occurrence of slip predicts the occurrence of slip in advance from a preset critical movement amount of the output center of gravity position. As described above, in the flexible structure, the shift of the output center of gravity occurs due to the deformation of the flexible structure before the slip occurs. This movement amount is experimentally grasped and set as a critical movement amount, and the occurrence of slip is predicted in advance.

  According to the present invention, it is possible to distinguish and determine the deformation of the flexible structure and the slipping of the object on the surface of the flexible structure. For this reason, only the slip of the object can be accurately determined. In addition, since the determination is performed by a pressure sensor that has a simple structure and can be easily miniaturized, it is easy to improve accuracy. In this way, by detecting the slip of the gripped object on the finger surface, it is possible to grip an object whose characteristics are unknown with a minimum gripping force that does not slip. In particular, by predicting slip in advance, it is possible to handle easily lossy products.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the accompanying drawings. In order to facilitate the understanding of the description, the same reference numerals are given to the same components in the drawings as much as possible, and duplicate descriptions are omitted.

  FIG. 1 is a diagram showing a configuration of a robot hand 1 having a slip detection device according to the present invention, and FIG. 2 is a block configuration diagram showing a control system thereof. 3 is a schematic configuration diagram of the robot 100 including the robot hand 1, and FIG. 4 is an explanatory diagram for explaining a slip detection device arranged at the fingertip of the robot hand of FIG.

  The robot hand 1 according to the present embodiment includes four fingers: a thumb 10, an index finger 11, a middle finger 12, and a ring finger 13. The thumb 10 is a four-degree-of-freedom link system in which a motor 14 is arranged for each of four joints. The other three fingers 11 to 13 are configured such that the motor 14 is arranged for each of the three joints excluding the most advanced joint, and the most advanced joint is configured to interlock with an adjacent joint (interlocking joint 17). Therefore, each of the four joints constitutes a link system having three degrees of freedom. That is, the degree of freedom of the entire robot hand 1 is 13. The tips 10t to 13t of these fingers are made of a curved elastic body and have a flexible structure on the finger surface. Further, an encoder potentiometer 16 for detecting the bending angle of the joint is disposed at each joint.

  As shown in FIG. 3, the robot hand 1 is attached to the body 8 of the robot 100 by an arm 7. The body 8 has arms 7 on the left and right, and a camera eye 3 as a visual sensor on the head. The arm 7 is also provided with a motor and an encoder potentiometer so that the robot hand 1 can be moved to a desired position.

  The control system is mainly configured by a control ECU 2 including a RAM, a ROM, a CPU, and the like, and recognizes the shape and position of an object by image recognition from an output image of a camera eye 3 that captures an image of the grasped object. An image recognizing unit 20, a hand position / posture calculating unit 21 that calculates a gripping position / griping posture based on the recognition result, and a robot hand control unit 22 that controls the movement of the robot hand 1. The robot hand control unit 22 receives an output signal from an encoder potentiometer 16 and a tactile sensor 18 described later, and controls the motor driver 4 to control the movement of each motor 14.

  The control system is not limited to such a configuration, and the control unit that controls the movement of the robot hand 1 so as to have the instructed posture, and the image recognition device and the gripping posture calculation unit may be separated. . In addition, these calculation units / control units may be divided by hardware or software.

  4A to 4C show the configuration of the fingertip 10 t of the thumb 10. The fingertip 10t has a flexible structure with a curved elastic body 10s on the surface side, and a plurality of tactile sensors 18 are embedded in the curved elastic body 10s. Here, FIG. 4C is a perspective view of the curved elastic body 10s in which the tactile sensor 18 is embedded. As shown in FIG. 4A, the direction toward the tip of the finger 10 on the surface of the fingertip 10t is set as the x-axis, the direction orthogonal to this is set as the y-axis, and the fingertip curved surface is developed in a plane. Is FIG. 4 (b). Here, the size of each tactile sensor 18 is 4 × 4 mm, and a total of 27 tactile sensors 18 are arranged on the fingertip 10t.

  The principle of operation of each tactile sensor 18 is shown in FIG. In each tactile sensor 18, electrodes 18a and 18b are arranged with a gap therebetween to form a capacitor. When pressure is applied, the gap between the electrodes 18a-18b changes, and the capacitance of the capacitor changes. The pressure is obtained from this change in capacitance (output voltage). The tactile sensor 18 is configured as a distributed pressure sensor. As the individual tactile sensors 18 are downsized and arranged in large numbers, the detection accuracy of the contact position is increased, but the output is reduced. Therefore, the number of sensors may be determined as appropriate according to the position / pressure detection accuracy. The structure of the tactile sensor 18 itself is simple, and it is easy to reduce the size thereof. There is an advantage that the cost of the sensor and the cost of the device for processing the sensor output can be relatively low. The same configuration may be applied to the other fingers 11 to 13, and slip detection may be performed only with the thumb 10, and the tactile sensor 18 may not be disposed on the other fingers 11 to 13.

  Since the fingertip 10t is configured by the curved elastic body 10s, as shown in FIG. 4A, when the object 200 is to be gripped with a certain gripping force, the fingertip 10t is subjected to pressure received according to the reaction force received from the object 200. The curved elastic body 10s is deformed. For this reason, the object 200 and the fingertip 10t come into contact with a surface having a certain area rather than a point, and pressure is applied to the tactile sensor 18 applied to the contact area, and the pressure is detected.

FIG. 6 is a graph showing an example of the distribution of pressure detected by the tactile sensor 18 when the object 200 is held. When the tactile sensor position is represented by coordinates (i, j) and the contact position is represented by the center of gravity of the output, the contact position (X _c , Y _c ) can be represented by the following equation (1).

Here, m1 to mn are obtained from the cell position of the tactile sensor 18 in the x direction, n1 to nn are obtained from the cell position of the tactile sensor 18 in the y direction, and f (i, j) is obtained from the output of the tactile sensor 18. Pressure value.

  Here, when an object placed on a plane is gripped (see FIG. 7A), only the normal contact force Fn acts between the finger and the object. In this case, Normally, the output distribution of the tactile sensors 18 group is substantially symmetric in the tangential direction. The gravity center position C0 at this time can be expressed by the above equation (1), and the normal contact force Fn can be expressed by the following equation (2).

  When the object 200 is held and lifted, an external force F that is gravity is applied to the object 200. As a result, the curved elastic body 10s is deformed along the acting direction of the external force, so that the output gravity center C moves downward by a distance S from C0. If the external force F overcomes the maximum static friction force Ft, slip occurs. The moving distance of the output center of gravity at this moment becomes the maximum value Smax. This maximum value Smax depends on Fn and has the relationship shown in FIG.

  In the gripping operation, it is necessary not to deal with the occurrence of slippage but to control it so as not to slip. Therefore, in order to predict the slip in advance, Smax is multiplied by K (however, a constant of K <1). Sslip is set as the critical movement amount of the output center of gravity, and if the movement amount S of the output center of gravity exceeds the critical movement amount Sslip, control is performed as the occurrence of slipping (see FIG. 8). Here, the correspondence relationship between Fn and Sslip may be obtained by experimentally examining the correspondence relationship between Fn and Smax in advance.

  FIG. 9 is a diagram showing a configuration example of an experimental apparatus for examining the relationship between Fn and Smax. The test finger 300 is attached to the tip of an arm 305 that is rotatably attached to a stand 304. A weight tray 303 is arranged between the rolling center of the arm 305 and the finger 300. By adjusting the weight of the weight placed on the weight tray 303, the stress (gripping) applied to the test object 301 of the finger 300 is adjusted. Can be adjusted.

  The test object 301 is disposed on the electronic balance 302, and the output of each sensor of the finger 300 is sent to the personal computer 308 via the communication cable 306 and the data logger 307 and processed.

  FIG. 10 is a flowchart of correspondence measurement between Fn and Smax. First, the load Fn applied from the fingertip of the finger 300 to the test object 301 is set to a desired value by adjusting the weight placed on the weight receiving tray 303 while viewing the display of the electronic balance 300 (step S101). . Next, the output values of the tactile sensor in a stationary state are collected by the data logger 307 and transferred to the personal computer 308 (step S102). Then, based on the output distribution of the tactile sensor, the initial output center-of-gravity position C0 is detected by equation (1) (step S103).

  Subsequently, while moving the test object 301 on the electronic balance 302 at a predetermined speed in the horizontal direction, the output values of the tactile sensor during that time are collected by the data logger 307 and transferred to the personal computer 308 (step S104). Based on the output distribution of the tactile sensor, the moving output center-of-gravity position C is detected by equation (1) (step S105). The movement amount S of the output center of gravity is calculated as the deviation between the obtained C and C0, and the maximum movement amount Smax of the change amount is obtained (step S106). Thereby, Smax corresponding to Fn is obtained.

  Next, it is determined whether or not the number of data (Fn, Smax) obtained by the test is sufficient (step S107). If the number of data is insufficient, the process proceeds to step S108, the weight is changed, the desired Fn is changed, and the process proceeds to step S102 to obtain the next (Fn, Smax) data. . When the number of data reaches a sufficient number, the process proceeds to step S109, where (Fn, Sslip) = (Fn, K × Smax) is set as the critical center-of-gravity movement amount, and this is databased.

In this database, as shown in FIG. 11, as for the critical movement amount Sslip, Fn-Sslip is captured as discrete points on the coordinates. Since Fn at the time of gripping does not necessarily match this discrete point data, the value of Sslip for any Fn is two points before and after (Fn ⁽ⁱ⁾ , S _slip ⁽ⁱ⁾ ), (Fn ^{(i + 1)} , S _slip ^{(i + 1)} ) must be obtained by interpolation. For example, the following equation (3) may be used as the interpolation equation.

  Next, control at the time of gripping an object by the robot hand 1 will be described with reference to a flowchart of FIG. Here, an operation of holding the object 200 on the desk with the right hand 1R of the robot and moving it to a predetermined location will be described.

  First, the image recognition unit 20 performs predetermined image recognition processing from the stereo image of the gripping object 200 acquired by the camera eye 3 to acquire information such as the position and shape of the gripping object 200 (step S121). . The gripping position / posture of the gripping target 200 by the robot hand 1R is set according to the obtained position / shape of the gripping target 200, and the position / posture including the robot arm 7R is calculated by inverse kinematic analysis ( Step S122).

  Next, the motor driver 4 drives the motors 14 at the joints of the robot arm 7R and the robot hand 1R to move the robot hand 1R to the obtained gripping position / posture to grip the object 200 ( Step S123). At the same time, the normal gripping force Fn of the contact point and the output center-of-gravity position C0 are detected from the output of the touch sensor 18 by the above-described equations (1) and (2) (step S124). The critical movement amount Sslip is calculated from the calculated Fn and the stored Fn-Sslip correspondence database using the equation (3) (step S125).

  After gripping, the robot arm 7R is driven to start the lifting operation of the object 200 from the desk (step S126). At the same time, the output center of gravity position C is detected from the output of the tactile sensor 18 by the above-described equations (1) and (2), and the movement amount S (= C−C0) from the initial position is obtained (step S127). Then, the movement amount S and the critical movement amount Sslip are compared (step S128).

  If the movement amount S is less than the critical movement amount Sslip, it is determined that no slip has occurred, and the process moves to step S130 to determine whether or not it is necessary to continue the lifting operation. If it is necessary to continue the lifting operation, the process returns to step S126 to lift the object. On the other hand, if it is determined that it is not necessary to continue the lifting operation, the process moves to step S131, and the robot arm 7R and the hand 1R and, if necessary, the robot body 8 are also driven to move the object 200 to a predetermined position. The hand 1R is opened, the object 200 is released, and the gripping task is completed.

  On the other hand, if it is determined in step S128 that the movement amount S exceeds the critical movement amount Sslip, it is determined that slipping occurs, the process moves to step S129 to increase the gripping force, and then the process proceeds to step S125. It moves, calculates Sslip that matches the normal gripping force Fn after increasing the gripping force, and continues the lifting operation.

  Thus, before slipping actually occurs, stable gripping can be realized by predicting the occurrence of slipping and increasing the insufficient gripping force. In particular, even when the weight of the object to be grasped is unknown or when the surface is slippery, it is possible to predict slipping and set an appropriate gripping force, so that gripping can be performed reliably. In addition, since the initial gripping force can be set to a minimum gripping force, an object that is easily damaged can be safely handled. Furthermore, since it is not necessary to register the characteristics of the gripping object in advance, it is possible to handle various gripping objects, no database for the object is required, and the robot hand control program becomes simple, Controllability is improved.

It is a figure showing composition of robot hand 1 which has a slip detection device concerning the present invention. It is a block block diagram which shows the control system of the robot hand of FIG. It is a schematic block diagram of the robot 100 provided with the robot hand 1 of FIG. It is explanatory drawing explaining the slip detection apparatus arrange | positioned at the fingertip of the robot hand of FIG. It is a figure explaining the operation | movement principle of the tactile sensor of FIG. It is a graph which shows the example of distribution of the pressure detected by tactile sensor 18 at the time of object grasping. It is a figure explaining the force which acts between the finger | toe and an object at the time of an object grip. It is an example which shows the relationship between Smax and Fn. It is a figure which shows the structural example of the experimental apparatus for investigating the relationship between Fn and Smax. It is a flowchart of the corresponding | compatible measurement of Fn and Smax. It is a figure explaining the interpolation of Fn-Sslip. It is a flowchart explaining the holding | grip operation | movement by the robot hand of FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Robot hand, 3 ... Camera eye, 4 ... Motor driver, 7 ... Arm, 8 ... Torso, 10 ... Mother finger, 10s ... Curved elastic body, 10t ... Finger tip, 11 ... Indicating finger, 12 ... Middle finger, 13 ... Ring finger , 14 ... Motor, 16 ... Encoder potentiometer, 17 ... Interlocking joint, 18 ... Tactile sensor, 20 ... Image recognition unit, 21 ... Hand position / posture calculation unit, 22 ... Robot hand control unit, 100 ... Robot, 200 ... Gripping Object, 300 ... finger, 301 ... test object, 303 ... dish, 304 ... stand, 305 ... arm, 306 ... communication cable, 307 ... data logger, 308 ... personal computer.

Claims (2)

In a robot hand having a finger surface with a flexible structure, a device for detecting the slip of the fingertip,
A distributed pressure sensor that divides a fingertip curved surface into a plurality of regions and detects respective forces in the normal direction of the surface;
Means for calculating an output center of gravity position and a normal contact force from the output of the distributed pressure sensor;
Means for determining the occurrence of slipping from the relationship between the movement of the output center of gravity position and the normal contact force;
A slip detection device for a finger surface of a robot hand, comprising:   2. The slip detection device for a finger surface of a robot hand according to claim 1, wherein the means for determining the occurrence of slip predicts the occurrence of slip in advance from a preset critical movement amount of the output center of gravity position.

Images