JP2010058256A - Arm position regulating method and apparatus and robot system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an arm position regulating method and an apparatus which can accurately eliminate the deviation of the position of an arm caused by deflection or the like, and to provide a robot system. <P>SOLUTION: The robot includes an arm comprising a joint having a rotary driving actuator 20 and a frame member connected to the joint, and a control unit that controls the actuator 20. An actual angle of the joint is acquired, and the level of deviation between a command angle given from the control unit to the joint and the actual angle of the joint is determined. Joint torque which acts on the joint of the arm is determined, and the spring coefficient of the joint is calculated based on the determined joint torque and the determined level of deviation. The command angle given from the control unit to the joint is corrected based on the calculated spring coefficient. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an arm position adjusting method and apparatus for correcting a positional shift caused by deflection of a robot arm having a plurality of joints, and a robot system.

  In recent years, articulated robots are used for machining various industrial workpieces. By using an articulated robot, it is possible to automatically and accurately process a workpiece from various positions and directions.

However, in such an articulated robot, the arm may be bent and deformed by the weight of the arm or the load applied to the tip of the arm. This deformation of the arm is particularly noticeable in the joint portion having the rotation axis. In other words, the member forming the joint part in the movable part may be elastically deformed, so that the joint part may be displaced in the movable direction from the target position.
Then, a deviation occurs between the target position of the displacement arm of the arm and the actual position of the arm, which contributes to a decrease in workpiece machining accuracy by the articulated robot.

Therefore, in order to suppress such a decrease in processing accuracy, a technique for correcting a positional shift caused by arm bending has been proposed.
For example, in Patent Document 1, the torsion (deflection) of the joint part is obtained from the torque applied to the joint part and the spring constant of the joint part obtained in advance, and the target position of the arm is corrected based on this. A technique for reducing the displacement of the arm due to bending has been disclosed.
JP 2002-307344 A

  However, since the spring constant (rigidity) of each joint is affected by the construction tolerance of the joint, the manufacturing tolerance of the speed reducer built in the joint, and the like, there are individual differences for each arm.

  For this reason, when the amount of flexure of the joint portion is obtained using the spring constant obtained in advance as in the technique of Patent Document 1, the corrected arm is caused by an error in the spring constant due to individual differences of the arms. It is conceivable that the error between the actual position of the arm and the target position of the arm increases.

  In particular, in the case of an articulated robot having a plurality of joints, the above-described error may occur as many as the number of joints provided on the arm, which may hinder the improvement of workpiece machining accuracy by the articulated robot. Conceivable.

  The present invention has been made in view of such problems, and an object of the present invention is to provide an arm position adjusting method and apparatus, and a robot system, which can eliminate a positional shift due to arm deflection or the like more accurately. And

In order to solve the above problems, the present invention is configured as follows.
The arm position adjustment method according to claim 1, wherein an arm having a joint portion having an actuator for rotationally driving and a frame member coupled to the joint portion, and the joint portion of the joint portion so that the arm assumes a target posture. An arm position adjustment method in a robot system having a control device for controlling the actuator, wherein a deviation amount between a command angle given from the control device to the joint portion and a measured value of an actual angle of the joint portion is obtained. A deviation amount obtaining step; a joint torque obtaining step for obtaining a joint torque acting on the joint portion of the arm; the joint torque obtained in the joint torque obtaining step; and the deviation amount obtained in the deviation amount detecting step. And a spring constant calculating step for calculating a spring constant of the joint portion based on Based on the spring constant of the joint calculated in, it is characterized by having a command angle correction step for correcting the command angle given to the joint portion from the control device.

In the deviation amount acquisition step, the angle formed between the axial direction of the frame member adjacent to both sides of the joint portion and the reference direction is measured by an angle sensor that detects the angle of the frame member in the axial direction. (Claim 2).
The actual angle of the joint can be accurately measured with a simple configuration in which the angles of the frame members on both sides of the joint are detected by the angle sensor and the angle formed by the two frame members sandwiching the joint is obtained.
In addition, as a method for obtaining the actual angle of the joint part, a method of obtaining the angle of the joint part from image data obtained by photographing an arm on which a light source is arranged with a camera is conceivable.

Further, the shift amount acquisition step includes a first drive step for driving the joint portion so that an axial direction of a frame member positioned on the distal end side with respect to the joint portion is parallel to a gravity direction, and the first step execution. Then, after executing the first angle obtaining step for obtaining a first angle formed by the axial direction of the frame member and the reference direction, and after executing the first angle obtaining step, the joint portion is provided with the command angle corresponding to a predetermined angle. A second drive step for changing the angle of the frame member; a second angle acquisition step for acquiring a second angle formed by the axial direction of the frame member and the reference direction after the execution of the second step; It is preferable to include a shift amount calculating step of calculating a shift amount of the joint portion based on a difference between the first angle and the second angle.
That is, in the first driving step, the center of gravity of the object supported by the frame member or the frame member on the distal end side relative to the joint portion is set by making the axial direction of the frame member on the distal end side of the arm relative to the joint portion the gravity direction. Therefore, the joint torque (moment) can be made small enough to be ignored. Therefore, the first angle is an angle when the joint portion is in an unloaded state.
Since the amount of deviation is calculated based on the difference between the second angle (when loaded) and the first angle (when no load) (that is, the actual angle of the joint) and the command angle, the spring constant of the joint is accurately determined. You can ask well.

  Further, the shift amount acquisition step includes a no-load angle acquisition step for acquiring a no-load angle formed by an axial direction and a reference direction of the frame members adjacent to both sides of the joint portion, and the joint portion. A load applying step for applying a preset load, and a load with which a load-on-time angle formed between the axial direction of the frame member adjacent to both sides of the joint and the reference direction is obtained after the load applying step is executed. It is preferable to calculate the shift amount based on a time angle acquisition step and a difference between the unloaded angle and the loaded angle.

  The arm position adjustment device according to claim 5, wherein an arm having a joint portion having an actuator for rotationally driving and a frame member coupled to the joint portion, and the joint portion of the joint portion so that the arm assumes a target posture. A control device for sending a command angle to the actuator, an angle acquisition means for acquiring an actual angle of the joint, and the command angle given to the joint from the control device and the angle acquisition means. Deviation amount calculating means for calculating a deviation amount from the actual angle of the joint portion, joint torque obtaining means for obtaining a joint torque acting on the joint portion of the arm, and the joint torque obtained by the joint torque obtaining means And a spring constant calculation means for calculating a spring constant of the joint based on the deviation amount obtained by the deviation amount calculation means, and the spring Based on the spring constant determined by the number calculating means it is characterized by having a a command angle correction means for correcting the command angle given to the joint portion from the control device.

  In addition, it is preferable that the angle acquisition unit includes an angle sensor that measures an angle formed between an axial direction of the frame member adjacent to both sides of the joint portion and a reference direction.

  Further, the shift amount acquisition means drives the joint portion so that the axial direction of the frame member positioned on the tip side of the arm with respect to the joint portion to be adjusted is parallel to the gravity direction, and the frame A first angle formed by the axial direction of the member and a reference direction is acquired, the command angle corresponding to a predetermined angle is given to the joint portion to change the angle of the frame member, and the axial direction of the frame member and the Preferably, a second angle formed with a reference direction is acquired, and the shift amount is calculated based on a difference between the first angle and the second angle.

Further, the deviation amount acquisition means acquires an unloaded angle formed by the axial direction of the frame member adjacent to both sides of the joint portion and a reference direction, respectively, and gives a preset load to the joint portion. , Obtaining a loaded angle formed by an axial direction and a reference direction of the frame members adjacent to both sides of the joint portion in a state where the load is applied to the joint portion, and the unloaded angle and the It is preferable to calculate the amount of deviation based on the difference from the loaded angle.
The robot system according to claim 9, wherein the arm of the joint unit includes an arm having an actuator that rotates and a frame member coupled to the joint, and the arm has a target posture. It has the control apparatus which controls this, and the arm position adjustment apparatus of any one of Claims 5-8, It is characterized by the above-mentioned.

According to the present invention (Claims 1 and 5), the actual angle of the joint is obtained for each arm, and the difference between the obtained actual angle and the command angle for control from the control device (that is, the deviation amount) Since the command angle given to the joint from the control device is corrected based on this, the spring constant can be obtained individually for each arm. Then, by correcting the command angle from the control device based on this spring constant, it is possible to eliminate the position shift due to the deflection of the arm and to position the arm with higher accuracy.
As a result, it is possible to optimally correct individual differences in the reduction gear spring constant caused by the manufacturing tolerance of the reduction gear in the joint portion of the arm, and to improve the accuracy of processing performed using the robot. it can. In particular, since the relative positional shift of each arm can be effectively eliminated, the processing accuracy can be further improved when a plurality of arms are cooperated.

Further, according to the present invention (Claims 2 and 6), by providing the angle sensors on the frame members adjacent to the joint part, the actual angle of the joint part including the deflection can be reliably obtained.
Further, according to the present invention (Claims 3 and 7), from the angle (first angle) at the position where the deflection of the joint due to gravity can be substantially ignored (first angle), at the position where the deflection of the joint occurs due to the weight of the frame. Therefore, the deflection (deviation amount) due to the weight of the joint portion can be detected with high accuracy.
Further, according to the present invention (claims 4 and 8), since the angle before and after the load is applied to the joint portion is measured, the deflection (deviation amount) of the joint portion caused by the load can be accurately detected.

[First Embodiment]
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.
1 to 5 are all for explaining the arm position adjusting device according to the first embodiment of the present invention. FIG. 1 is a diagram schematically showing the entire configuration of the robot system, and FIG. 2 is a control diagram. 3 is a schematic block diagram showing the functional configuration of the device, FIG. 3 is a schematic diagram showing the state of the arm when determining the spring constant of the joint, FIG. 4 is a flowchart showing the operation procedure of the arm position adjusting device, and FIG. It is a schematic diagram which shows the state of the arm at the time of calculating | requiring the spring constant of a part.

(Outline configuration)
As shown in FIG. 1, a robot system (robot) 1 includes an arm 2, a base 3 on which the arm 2 is fixed, and a control device (displacement amount calculation means, joint torque acquisition means, command angle) constituted by a computer. Correction means) 10.
The arm 2 is a so-called vertical articulated manipulator, and a plurality of serial frame members 4ba, 4S, 4L, 4U, 4R, 4B, and 4T, and joint portions 5S and 5L connected between the frame members. , 5U, 5R, 5B, 5T.
The joint portions 5S, 5L, 5U, 5R, 5B, and 5T are each constituted by an actuator 20 (see FIG. 2), and the adjacent frame members are rotated by driving the actuator 20.
Each of the frame members 4ba, 4S, 4L, 4U, 4R, 4B, and 4T is a structural member that supports the load applied to the arm 2, and constitutes an axis that extends from the base end to the tip end of the arm 2, respectively.

Each of the joint portions 5L, 5U, 5B has a rotation axis that is orthogonal to the extending direction (axial direction) of the adjacent frame member, and the frame member is driven by driving the joint portions 5L, 5U, 5B. It swings.
On the other hand, the joint portions 5S, 5R, 5T have their respective rotational axes parallel to the extending direction of the adjacent frame member, and the frame members are rotated by the rotational drive of the joint portions 5S, 5R, 5T. ing.
The frame member 4T is provided with an attachment portion for attaching an end effector (not shown) such as a welding torch or a workpiece.

Reference numerals 6S, 6L, 6R, and 6T denote angle sensors (angle acquisition means) that detect and detect angles formed by the axial direction of the frame members 4S, 4U, and 4B and the reference direction (here, the gravity direction), respectively. The result is input to the control device 10.
Each actuator 20 incorporated in each joint part 5S, 5L, 5U, 5R, 5B, 5T is connected to the control device 10 so as to be able to transmit information.
The control device 10 inputs each joint part 5S, 5L, 5U, 5R, 5B according to the coordinate information of the position where each joint part 5S, 5L, 5U, 5R, 5B, 5T should be located, which is input in advance by teaching or the like. , 5T respectively control the driving of the actuators 20 constituting the joint portions 5S, 5L, 5U, 5R, 5B, 5T so as to reach the respective target positions (command angles).

Here, the configurations of the control device 10 and the actuator 20 will be described in more detail. As shown in FIG. 2, the control device 10 is assigned functions of a deflection correction amount calculation unit 11, a spring constant calculation unit 12, a position / speed control unit 13, a torque control unit 14, a differentiator 15, and a target position setting unit 16. ing.
The actuator 20 includes a servo motor 21, a speed reducer 22, and an encoder (rotary encoder) 23.

The servo motor 21 is driven in accordance with a command from the control device 10 and rotates the joint portions 5S, 5L, 5U, 5R, 5B, and 5T via the speed reducer 22. The encoder 23 detects the rotation amount (rotational position) of the servo motor 21 and transmits it to the control device 10. In the control device 10, the joint is determined from the input from the encoder 23 and the reduction ratio of the speed reducer 22. The rotational positions of the sections 5S, 5L, 5U, 5R, 5B, and 5T are detected.
The target position setting unit 16 of the control device 10 servos from the coordinate information (target position) of the positions where the joints 5S, 5L, 5U, 5R, 5B, and 5T should be located, which are input in advance by teaching or the like. A required angle (command angle) θref to be rotated by the motor 21 is calculated.

The difference between the required angle θref and the detected rotation angle θenc of the servo motor 21 input from the encoder 23 is calculated, and then the deflection correction amount Δθ * comp calculated by the deflection correction amount calculation unit 11 is added. Then, the position / speed control unit 13 calculates the required torque Tref for the servomotor 21 from the input command angle (θref−θenc + Δθcomp). The required torque Tref is obtained from the difference between the current speeds of the axes in the speed control, and is obtained from the product of the acceleration and inertia.
The torque control unit 14 transmits a command to the servo motor 21 after converting the torque command value to an optimum torque command value for the servo motor 21 in accordance with the target torque Tref.
The deflection correction amount calculation unit 11 includes the spring constant K corresponding to the joints 5S, 5L, 5U, 5R, 5B, and 5T obtained by the spring constant calculation unit 12 and the required torque Tref output from the position / speed control unit 13. From this, the amount of deflection Δθ * comp of each joint portion 5S, 5L, 5U, 5R, 5B, 5T is obtained. The deflection amount Δθcomp is calculated by the following equation (1).
Δθcomp = ΔτL / KL (1)

Next, the procedure for calculating the spring constant K by the spring constant calculator 12 will be described with reference to FIG.
First, the joint portions 5L, 5U, and 5B that are orthogonal to the extending direction (axial direction) of the frame member with which the rotation axis is adjacent are targeted for adjustment, the spring constant KL of the joint portion 5L, and the spring constant of the joint portion 5U. KU, the spring constant KB of the joint 5B is obtained.
The joint portions 5S, 5R, 5T are movable only around the rotation axis parallel to the extending direction of the frame member, and the displacement (deflection) due to the load (load) with respect to the rotation direction of the joint portions 5L, 5U, 5B is relative. Here, the joint portions 5L, 5U, and 5B are regarded as structures (links) integral with the frame member. Therefore, for simplification of illustration, the joint portions 5S, 5R, and 5T are schematically shown in FIG. 3, and here, the links L1 to L4 are connected to the joint portions to be adjusted by 5L, 5U, and 5B connections. It is considered as a frame member.
The extending direction of the frame member is a direction extending from the proximal end to the distal end of the arm 2 and refers to the longitudinal direction of the link when the frame member is regarded as a link.
As shown in FIG. 1, the inclination angle sensors 6S, 6L, 6R, and 6T are installed close to the joint portion (or link) on the distal end side, so that in addition to the deflection of the speed reducer 22 at the joint portion, You can also measure the deflection of the link itself.

As shown in FIG. 3, the angle sensor 6S attached to the frame member 4S is an angle in the axial direction of the link L1 between the base 3 and the joint 5L (ie, the frame members 4Ba and 4S and the joint 5S). Is to measure.
The angle sensor 6L attached to the frame member 4L measures the angle in the axial direction of the link between the joint portions 5L and 5U (that is, the frame member 4L). The angle sensor 6R attached to the frame member 4U measures an axial angle of the frame member (that is, the frame member 4L) between the joint portion 5L and the joint portion 5U.

At this time, the attachment positions of the angle sensors 6S, 6L, 6R, and 6T may be arbitrary positions among the links L1 to L4. However, the attachment angle is adjusted so that the robot coordinate system Y-axis and the angles detected by the angle sensors 6S, 6L, 6R, and 6T are the angles formed by the axial directions of the links L1 to L4 and the gravity direction, respectively. Has been.
The robot coordinate system is an orthogonal coordinate system fixed to a robot (here, an arm) body. Here, as shown in FIG. 3, the front and rear direction is the X axis, the left and right direction is the Y axis, and the vertical direction is the vertical direction. The Z axis is taken. For adjusting the mounting angles of the angle sensors 6S, 6L, 6R, and 6T, for example, the angle sensors 6S, 6L, 6R, and 6T are applied with the sensors capable of measuring the angles about two orthogonal axes and rotating one of them. The mounting posture may be finely adjusted so that the angle does not change.

The position P1 is the position (hand position) of the arm tip when the entire arm 2 stands in a direction parallel to the direction of gravity (vertical), and the position P2 is from the position P1 to the joint 5L as indicated by the solid line. Here, it is a hand position when it is rotationally driven by 90 degrees (predetermined angle).
In FIG. 3, as indicated by a broken line, when the hand position of the arm 2 is at P1, the centers of gravity of the links L1 to L4 of the arm 2 are all located on a single straight line parallel to the direction of gravity. For this reason, since the load (load torque) due to its own weight in the rotation direction of each joint portion 5L, 5U, 5B is extremely small, the displacement amount (deflection) in the rotation direction of each joint portion 5L, 5U, 5B should be ignored. Can do.

On the other hand, when the hand position of the arm 2 is P2, the joints 5L, 5U, and 5B are loaded on the actuator 20 in a direction against gravity in order to keep the arm 2 horizontal with respect to its own weight. Torque is generated. For this reason, the speed reducer 22 (not shown) of the joint portions 5L, 5U, and 5B is elastically deformed according to the load torque, and the joint portions 5L, 5U, and 5B are displaced in the rotational direction accordingly.
In the spring constant calculation unit 12, the angular displacement (deviation amount) of each speed reducer 22 between the state of the position P1 and the state of the position P2 obtained by the angle sensors 6S, 6L, 6R, 6T due to the elastic deformation of each speed reducer 22 is obtained. ) And the load torque applied to each speed reducer 22, the spring constant K is calculated.
Hereinafter, the function of the spring constant calculation unit 12 will be described in more detail.

Hereinafter, the arm position adjustment mode according to the first embodiment will be described. As shown in FIG. 4, first, in step S1, the hand position of the arm 2 is moved to P1 by the control device 10 (first driving step).
That is, the axial directions of the links L2 to L4, which are frame members adjacent to the distal ends of the joint portions 5L, 5U, and 5B, are made parallel to the direction of gravity. Thus, the actuator 20 is controlled so that the load torque applied to each joint portion 5L, 5U, 5B becomes substantially zero. At this time, the inclination angles (first angles) θ ^P1 _{Base_S} , θ ^P1 _{L_S} , θ ^P1 _{U_S} , and θ ^P1 _{B_S of the} links L1 to L4 measured by the angle sensors 6S, 6L, 6R, and 6T are _stored in the control device 10. (First angle acquisition step).

Next, in step S2, only the joint portion 5L is rotated by 90 degrees according to a command from the control device 10, and the hand position of the frame member arm 2 on the tip side of the joint portion 5L is moved to P2 (second driving step). ). At this time, the inclination angles (second angles) θ ^P2 _{Base_S} , θ ^P2 _{L_S} , θ ^P2 _{U_S} , and θ ^P2 _{B_S of the} links L1 to L4 transmitted from the angle sensors 6S, 6L, 6R, and 6T are stored in the control device 10. It is stored in the area (second angle acquisition step).
The torque command values τ _L ^P2 , τ _U ^P2 , and τ _B ^P2 to the actuators 20 of the joint portions 5L, 5U, and 5B calculated by the position / speed control of the arm 2 at this time are stored as the joint torque ( Joint torque acquisition step).

In step S3, the inclination angles θ ^P1 _{Base_S} , θ ^P1 _{L_S} , θ ^P1 _{U_S} , θ ^P1 _{B_S} of the links L1 to L4 stored in step S1, and the inclination angles of the links L1 to L4 stored in step S2. Rotational displacement due to bending of the base 3 and the link L1 when the arm 2 is operated from step S1 to step S2 from θ ^P2 _{Base_S} , θ ^P2 _{L_S} , θ ^P2 _{U_S} , and θ ^P2 _{B_S.}

Then, the actual rotation angles (actual angles) Δθ _{Base_S} , Δθ _{L_S} , Δθ _{U_S} , Δθ _{B_S of the} joint portions 5L, 5U, 5B are calculated.
The rotational displacement Δθ _{Base_S} between the base 3 and the link L1 and the actual rotational angles Δθ _{L_S} , Δθ _{U_S} , Δθ _{B_S} of the joint portions 5L, 5U, 5B are expressed by the following equation (2), respectively.

In step S4, the deflection amount of each axis of the robot is calculated from the difference between the command angle to the robot and the actual link movement amount (that is, the actual rotation angle) measured by the tilt angle sensor. Command angles (command values) given from the control device 10 to the actuators 20 of the joints 5L, 5U, 5B of the arm 2 and the joints 5L, 5U measured by the angle sensors 6S, 6L, 6R, 6T , 5B actual rotation angles Δθ _{L_S} , Δθ _{U_S} , Δθ _{B_S,} and deviation amounts (deflection angles) Δθ _{L with} respect to the target angles (ie, command angles) on the commands of the joint portions 5L, 5U, 5B. , Δθ _U , Δθ _B are calculated.
The command angle given from the control device 10 is not limited to the angle, and may be a command value such as the driving amount of each actuator 20. The angle is calculated from the command value and is used as the command angle given from the control device 10. Also good.
That is, Steps S1 to S4 correspond to a deviation amount acquisition step.
The deflection angles Δθ _L , Δθ _U , Δθ _B of the joint portions 5L, 5U, 5B are expressed by the following equation (3).




Here, the value of the command angle Δθ _{L_C} given to the joint 5L is 90 degrees, and the values of the command rotation angles Δθ _{U_C} and Δθ _{B_C} for the joint 5U and the indirect part 5B are 0 degrees.

In step S5, the deflection amounts Δθ _L , Δθ _U , Δθ _B and the joint torques τ ^P2 _L , τ ^P2 _U , τ ^P2 _B calculated by the control device 10 are used, Spring constants (rigidities) KL, KU, KB of the joint portions 5L, 5U, 5B are calculated (identified).
The spring constant is a deviation amount of the angle due to the deflection per unit load applied to each joint part. The spring constant KL of the joint part 5L, the spring constant KU of the joint part 5U, and the spring constant KB of the joint part 5B are expressed by the following equations. It is represented by (4).



In this embodiment, four angle sensors are provided. However, the angle sensor 6S measures θ ^P1 _{Base_S} and θ ^P2 _{Base_S} , and the angle sensor 6T measures θ ^P1 _{Base_S} and θ ^P2 _{Base_S} to obtain Δθ _{Base_S.} Then, Δθ _{L_S} can be obtained using the obtained Δθ _{Base_S} and θ ^P1 _{L_S} and θ ^P2 _{L_S} . For this reason, even if the angle sensors 6L and 6R are omitted, the spring constants KL, KU, and KB can be calculated.
That is, the number of sensors can be reduced by obtaining the relative movement amount when the arm 2 is moved from the position P1 to the position P2. That is, it is not necessary to mount sensors for measuring the inclination angle on all the links L1 to L4, and a method of sequentially identifying the spring constant of the joint portion from the base 3 side can be taken.

Next, the joint 5R whose rotation axis is parallel to the extending direction (axial direction) of the frame member is set as an adjustment target.
As shown in FIG. 5, first, the joint portion 5B on the distal end side with respect to the joint portion 5R is set to a predetermined angle (here, 90 °). At this time, since the member on the distal end side relative to the joint portion 5R is not biased with respect to the rotation axis of the joint portion 5R, the load torque (gravity due to the gravity of the member on the distal end side relative to the joint portion 5R is applied to the joint portion 5R. (Also called moment) can be regarded as virtually not applied. Then, the angle sensors 6R and 6T measure the inclination angles (inclinations at no load) θ ^P1 _{R_S} and θ ^P1 _{T_S} of the links in the posture where the gravitational moment is not applied.
Next, the joint portion 5R is rotated by a predetermined angle (here, 90 °). At this time, since the member on the distal end side relative to the joint portion 5R is eccentric with respect to the rotation axis of the joint portion 5R, a load torque due to the gravity of the member on the distal end side relative to the joint portion 5R is applied to the joint portion 5R. In this posture, the angle sensors 6R and 6T measure the inclination angles (load inclination angles) θ ^P2 _{R_S} and θ ^P2 _{T_S} of the links in the posture where no gravitational moment is applied.
Identification can be performed by changing the posture of the robot from a posture in which the moment of gravity is not applied to the joint axis to a posture in which the moment of gravity is applied.
According to the following formulas (5) and (6), the tilt angles θ ^P1 _{R_S} and θ ^P1 _{T_S of the} links of the links in the postures not subjected to the gravitational moment measured by the angle sensors 6R and 6T, and the respective links. The actual rotation angles Δθ _{R_S} and Δθ _{T_S} are obtained from the in-load inclination angles θ ^P2 _{R_S} and θ ^P2 _{T_S,} and the deflection angle Δθ _{R is calculated} from the command angle Δθ _{L_R} given to each joint 5 from the following equation (7). , Δθ _T is obtained. From this, the spring constant KR of the joint 5R is calculated.
The joint 5T is changed from a posture in which a gravitational moment is not applied to the joint shaft 5T to a posture in which the gravitational moment is applied to the joint shaft 5T in a state where a heavy load that can be eccentric with respect to the rotation axis of the joint 5T is attached to the tip side of the 5T. By rotating the joint shaft 5T, the spring constant can be similarly calculated (identified).

Since the arm position adjusting device and the robot system according to the first embodiment of the present invention are configured as described above, the optimum spring constant can be individually obtained for each solid of each arm, and based on this spring constant. By correcting the command angle from the control device, it is possible to eliminate the position shift due to the deflection of the arm and to position the arm with higher accuracy.
In addition, a simpler configuration can be achieved by reducing the number of sensors by using the amount of tilt angle movement necessary for calculating the spring constant as a relative movement amount, or by using a method based only on changes in the weight of the robot due to changes in the posture of the robot. Can do.
Further, the spring constants can be collectively identified for the three joint portions 5L, 5U, and 5B simply by driving the joint portion 5L closer to the base 3 from the position P1 to the position P2.
Further, the joint portions 5R and 5T whose rotation axis is parallel to the axial direction of the arm also take a posture in which the load torque due to gravity can be changed by the rotation of the joint portions 5R and 5T at the tip and the state where the load torque is not applied. The spring constant can be calculated with high accuracy by using the measured values of the angle sensor in such a state and in a state where it is not applied.
In addition, since the spring constant of each joint can be easily calculated, if the spring constant of each joint is calculated regularly, the deflection correction amount calculation unit 11 can respond to changes in the arm over time, etc. Therefore, the optimum deflection correction can be performed each time, and the operation accuracy of the robot system 1 can be further improved.

  Next, a second embodiment of the present invention will be described with reference to the drawings. This embodiment is configured in the same manner as in the first embodiment except for the functional configuration of the control device, and the same parts as those in the first embodiment will not be described and will be described using the same reference numerals.

6 and 7 illustrate the second embodiment of the present invention, FIG. 6 is a diagram schematically showing the state of the arm, and FIG. 7 is a flowchart showing the operation procedure of the arm position adjusting device. .
As shown in FIG. 6, in the present embodiment as well, in the same manner as in the first embodiment, an angle sensor is installed on each of the links L1 to L4, and the inclination angle of each link of the actual robot including the deflection of the shaft is measured. In the figure, W is a load having an accurate weight.
In FIG. 6, P3 indicates the position of the hand when the robot arm is extended in the X-axis direction (dashed line) with no load applied to the hand (tip of the arm 2), and P3 ′ indicates the load on the hand. The hand position when a position command similar to P3 is given (solid line) in a state where is attached. Δθ _L indicates the amount of deflection of the L axis caused by the hand load.

FIG. 7 is a flowchart of the second embodiment. First, the hand of the robot is moved to P3, and each link inclination angle Δθ ^P3 _{Base_S} , Δθ ^P3 _{L_S} , Δθ ^P3 _{U_S} , Δθ ^P3 _{B_S} and joint torques τ ^P3 _L , τ ^P3 _U , τ ^P3 _B are measured. (Step T1).

Next, the load object is gripped by the hand of the robot, and a position command similar to P3 is given. Each link inclination angle Δθ ^{P3 ′} _{Base_S} , Δθ ^{P3 ′} _{L_S} , Δθ ^{P3 ′} _{U_S} , Δθ ^{P3 ′} _{B_S} and joint torques τ ^{P3 ′} _L , τ ^{P3 ′} _U , τ ^{P3 ′} _B are measured (step T2 ).

Load gripping front and rear of the angle sensor value ^{_{Δθ P3 Base_S, Δθ P3 L_S,}} Δθ P3 U_S, Δθ P3 B_S, Δθ P3 'Base_S, Δθ P3' L_S, Δθ P3 'U_S, Δθ P3' from _{B_S,} due to the change of the end load The amount of deflection of each axis is calculated by the following equation (8).

Furthermore, the joint torque that causes the deflection at this time is expressed by the following equation (7) using the joint torques τ ^P3 _L , τ ^P3 _U , τ ^P3 _B , τ ^{P3 ′} _L , τ ^{P3 ′} _U , and τ ^{P3 ′} _B before and after the load gripping. 9) (Step T3).



As described above, the torque command value calculated by the position / speed control may be used as the joint torque of each axis, or the torque value calculated by the Newton Euler method may be used.

And said deflection amount [Delta] [theta] _*, the joint torque .DELTA..tau generating deflection _L, Δτ _U, using .DELTA..tau _B, spring constant KL of each joint from the following equation (10), KU, calculates the KB (step T4).



As described above, the spring constant of each axis can be obtained. In this embodiment, the measurement position is set to one point P3. However, in order to increase the reliability of measurement, several points are measured, and the spring constant is calculated using an approximate expression from the relationship between the amount of deflection of each axis and the joint torque that generates the deflection. You may ask for.
Next, the joint portions 5R and 5T whose rotation axis is parallel to the extending direction (axial direction) of the frame member are to be adjusted. As in the first embodiment, the spring constants of the joint portions 5R and 5T can be identified by changing the posture of the robot from a posture in which no gravity moment is applied to the joint axis to a posture in which gravity is applied.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention.
For example, in the first embodiment, the spring constant is identified from the position in which the frame member on the tip side of the joint part to be the target of the robot is standing in the gravitational direction (vertical) and the two postures in which the L axis is moved 90 degrees. However, in order to further improve the accuracy, a posture (frame member angle) for obtaining a deflection angle and a joint torque in a posture while driving the frame member by 90 degrees is further provided, and an approximate expression such as a least square method is provided. It may be used to calculate the spring constant of each joint.
Further, as described above, the torque command value calculated by the control device may be used for each joint torque, or the torque value calculated by the Newton Euler method may be used.
In the embodiment, the robot system having only one robot arm has been described. However, the robot system may be configured to process various industrial works in cooperation with a plurality of robot arms. . In this case, by adjusting the arm position for a plurality of robot arms in the robot system, a variation in the amount of deflection caused by individual differences of each robot arm is suppressed, and a robot system with higher operation accuracy is configured. Can do.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram for explaining a first embodiment of the present invention and schematically showing an overall configuration of a robot system. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic block diagram illustrating a functional configuration of a control device, illustrating a first embodiment of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram for explaining a first embodiment of the present invention and showing a state of an arm when obtaining a spring constant of a joint portion. 1 is a flowchart illustrating an operation procedure of an arm position adjusting device, illustrating a first embodiment of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram for explaining a first embodiment of the present invention and showing a state of an arm when obtaining a spring constant of a joint portion. It is a figure which illustrates 2nd Embodiment of this invention and shows the state of an arm typically. 9 is a flowchart illustrating an operation procedure of an arm position adjusting device, illustrating a second embodiment of the present invention.

Explanation of symbols

1 Robot system (robot)
2 Arm 3 Base 4ba, 4S, 4L, 4U, 4R, 4B, 4T Frame member 5S, 5L, 5U, 5R, 5B, 5T Joint portion 6S, 6L, 6R, 6T Angle sensor 10 Control device 11 Deflection correction amount calculation Unit 12 Spring constant calculation unit 13 Position speed control unit 14 Torque control unit 15 Differentiator 16 Target position setting unit 20 Actuator 21 Servo motor 22 Decelerator 23 Encoder L1-L4 Link

Claims (9)

A robot system comprising: an arm having a joint portion having an actuator that rotates and a frame member coupled to the joint portion; and a control device that controls the actuator of the joint portion so that the arm assumes a target posture. The arm position adjustment method of
A deviation amount obtaining step for obtaining a deviation amount between a command angle given from the control device to the joint portion and a measured value of an actual angle of the joint portion;
A joint torque obtaining step for obtaining a joint torque acting on the joint portion of the arm;
A spring constant calculating step for calculating a spring constant of the joint part based on the joint torque obtained in the joint torque obtaining step and the deviation amount obtained in the deviation amount detecting step;
A command angle correction step of correcting the command angle given to the joint portion from the control device based on the spring constant of the joint portion calculated in the spring constant calculation step. Position adjustment method. In the deviation amount acquisition step,
The angle formed between the axial direction of the frame member adjacent to both sides of the joint portion and a reference direction is measured by an angle sensor that detects an angle in the axial direction of the frame member, respectively. The arm position adjusting method according to claim 1. The deviation amount acquisition step includes:
A first driving step for driving the joint portion so that the axial direction of the frame member positioned on the distal end side of the arm relative to the joint portion to be adjusted is parallel to the direction of gravity;
A first angle acquisition step of acquiring a first angle formed by the axial direction of the frame member and a reference direction after the first step is performed;
A second driving step for changing the angle of the frame member by giving the command angle corresponding to a predetermined angle to the joint after the first angle obtaining step is performed;
A second angle obtaining step for obtaining a second angle formed by the axial direction of the frame member and the reference direction after the second step is performed;
The arm position adjustment method according to claim 1, further comprising: a displacement amount calculating step of calculating a displacement amount of the joint portion based on a difference between the first angle and the second angle. The deviation amount acquisition step includes:
An unloaded angle acquisition step for acquiring an unloaded angle formed by an axial direction of each of the frame members adjacent to both sides of the joint and a reference direction; and
A load applying step for applying a preset load to the joint part;
After the load application step is executed, an on-load angle acquisition step for acquiring an on-load angle formed by an axial direction and a reference direction of each of the frame members adjacent to both sides of the joint portion, and
The arm position adjusting method according to claim 1 or 2, wherein the shift amount is calculated based on a difference between the unloaded angle and the loaded angle. An arm having a joint part having an actuator for rotationally driving and a frame member connected to the joint part;
A control device for sending a command angle to the actuator of the joint so that the arm assumes a target posture;
Angle acquisition means for acquiring an actual angle of the joint part;
A deviation amount calculating means for calculating a deviation amount between the command angle given to the joint portion from the control device and an actual angle of the joint portion acquired by the angle acquisition means;
Joint torque obtaining means for obtaining joint torque acting on the joint portion of the arm;
A spring constant calculating means for calculating a spring constant of the joint portion based on the joint torque obtained by the joint torque obtaining means and the deviation amount obtained by the deviation amount calculating means;
An arm having command angle correction means for correcting the command angle given to the joint portion from the control device based on the spring constant calculated by the spring constant calculation means; Positioning device. The angle acquisition means includes
6. The arm position adjusting device according to claim 5, further comprising an angle sensor for measuring an angle formed by an axial direction of each of the frame members adjacent to both sides of the joint portion and a reference direction. The deviation amount acquisition means includes
Driving the joint so that the axial direction of the frame member positioned on the tip side of the arm relative to the joint to be adjusted is parallel to the direction of gravity;
Obtaining a first angle between the axial direction of the frame member and a reference direction;
Changing the angle of the frame member by giving the command angle corresponding to a predetermined angle to the joint,
Obtaining a second angle formed by the axial direction of the frame member and the reference direction;
The arm position adjusting device according to claim 5 or 6, wherein the shift amount is calculated based on a difference between the first angle and the second angle. The deviation amount acquisition means includes
Obtaining an unloaded angle between the axial direction of the frame member adjacent to both sides of the joint and the reference direction,
Apply a preset load to the joint,
In the state where the load is applied to the joint part, obtain the loaded angle formed by the axial direction and the reference direction of the frame member adjacent to both sides of the joint part,
7. The robot arm position adjusting device according to claim 5, wherein the shift amount is calculated based on a difference between the unloaded angle and the loaded angle. An arm having a joint portion having an actuator for rotationally driving, and a frame member coupled to the joint portion;
A control device that controls the actuator of the joint so that the arm has a target posture, and the arm position adjustment device according to any one of claims 5 to 8. , Robot system.