JP2009045714A - Method and device for teaching attitude of robot arm - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and a device for teaching the attitude of a robot arm capable of preventing the operating angle of the robot arm from being undesirably increased by reducing the labor and time for teaching by the operator and reducing a necessary power by reducing the potential energy of the attitude of the robot arm. <P>SOLUTION: A rear coordinate origin O<SB>7</SB>at which a hand 8 is temporarily positioned is calculated to determine whether interference is present or not. When interference is absent, the angles of joints J1, J2 are specified, and the coordinate origin O<SB>4</SB>of an elbow part 4 is calculated. Based on the coordinate origin O<SB>4</SB>, the presence or absence of interference is determined, and a wrist angle is calculated. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a robot arm posture teaching method and posture teaching device applied to, for example, industrial robots, robots that support human life, and the like.

In the field of industrial robots and robots that support human life, for example, articulated robots having joints with seven or more axes have been proposed (Patent Document 1). This articulated robot has posture redundancy. This "redundancy" is a state in which the possibility is increased by adding an extra thing. In the system, etc., the configuration is duplicated, multiplexed, or a spare means is prepared. Redundancy is ensured.
In addition, as a method for avoiding interference between the articulated robot and the obstacle, the object, etc., the distance from the comparison reference point provided on the object to the comparison reference point provided on the articulated robot is maximized. There was a method to avoid the obstacles by determining the posture of the robot and to perform the target work.
JP-A-2005-14108

A multi-joint robot with seven or more axes has posture redundancy, but the control method is somewhat complicated. For this reason, there is a method of directing and operating the obstacle avoidance and object gripping postures by the teaching device so that the target posture is taken. However, in this method, a person must instruct and operate using a teaching device, in other words, the articulated robot must be operated manually to instruct and store position data and mounting order. It takes time and effort.
Further, when an articulated robot avoids an obstacle and grips an object with a hand attached to the tip, there are many gripping postures of the robot. Some of them are not suitable as robot postures because the so-called “elbow height” is high, so that the potential energy increases and the bending angle of the so-called “wrist” increases.

Specifically, as a method for avoiding interference between the multi-joint robot having the above-described redundancy and the object to be grasped and the obstacle by automatic operation, a comparison reference point is provided for each of the articulated robot and the object to be grasped and the obstacle, There are the following problems in the method of performing the objective work by determining the posture of the robot so as to maximize the distance calculated based on these comparison reference points and avoiding the obstacle.
The robot may avoid the obstacles too much, and the movement angle may become undesirably large in order to realize the posture.
-Since the "elbow height" is not taken into account, the positional energy of the determined robot posture becomes large, which may increase the required power.

  An object of the present invention is to reduce the labor and time of teaching by an operator, prevent an operation angle of the robot arm from becoming undesirably large, and reduce the positional energy of the posture of the robot arm to reduce the necessary power. It is an object to provide a posture teaching method and posture teaching device for a robot arm that can be made small.

In the robot arm posture teaching method according to the present invention, the movable part sequentially connected through the joint is connected to the hand 8 holding the object W, and the hand 8 is connected to the hand 8 via the wrist joints J5 to J7 so that the posture can be changed. The forearm 5 and the upper arm 3 connected to the elbow 4 of the forearm 5 via elbow joints J3 and J4 have redundancy in the posture of the hand 8 when the hand 8 grips the object W. A posture teaching method of the robot arm RA that determines the posture of the multi-joint when the hand 8 grips the object W under the condition of avoiding interference with the obstacle OB with respect to the multi-joint robot arm RA. An attitude for determining the interference between the robot arm RA and the obstacle OB when the attitude of the hand 8 is variously changed and the attitude of each joint is variously changed within a range in which the hand 8 can grip the object W. Change Evaluation value for calculating a predetermined evaluation value in the case where the posture of the hand 8 and each joint, which is considered not to cause interference in the interference determination process (step a1) and the posture change / interference determination process (step a1), is taken. A calculation process (step a2), and an evaluation value comparison / teaching posture determination process (step a3) for teaching the posture of each joint to the posture of each joint having the smallest evaluation value by comparing the evaluation values,
For the evaluation value, a reference coordinate origin 0 ₀ is arbitrarily determined as a fixed coordinate system for the fixed portion of the base end of the robot arm RA, and two orthogonal axes in the horizontal direction (x ₀ axis, y ₀ axis) and vertical direction of one axis to (z ₀ axis) consider the reference coordinate system of the orthogonal three-axis directions of the coordinate axes, and Hijidaka of E _Z the vertical coordinates in the reference coordinate system of a predetermined point in the elbow joint J3, J4 The angle formed by the hand 8 with respect to the forearm 5 in the horizontal plane (x ₀ , y ₀ plane) is the wrist angle θ, and the elbow height E _Z and the wrist angle θ are respectively weighted and the product thereof. The evaluation value is (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ).

According to this configuration, in the posture change / interference determination process (step a1), the robot arm when the posture of the hand 8 is variously changed and the posture of the multi-joint is variously changed within a range in which the hand 8 can grip the object W. Interference between RA and obstacle OB is determined. In the evaluation value calculation process (step a2), a predetermined evaluation value is calculated when the posture of the hand 8 and each joint that is assumed not to cause interference is taken. This evaluation value is the above (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ), and then the evaluation value in the evaluation value comparison / teach posture determination process (step a3). The posture of each joint is taught to the posture of each joint having the smallest evaluation value. Thus, the posture in which the elbow height _EZ, that is, the position energy and the wrist angle θ is optimized at the same time can be determined without the robot arm RA greatly avoiding the obstacle OB.

The robot arm RA has three degrees of freedom in which the wrist joints J5 to J7 can change the inclination angle of the hand 8 in two orthogonal axes with respect to the forearm 5, and the forearm 5 can rotate around the axis 5i. In addition, the elbow joints J3 and J4 may be joints having two degrees of freedom that can change the inclination angle in the two orthogonal axes. In this case, the robot arm RA has posture redundancy due to such a multi-joint, and can realize a large number of gripping postures. At this time, the posture in which the elbow height _EZ and the wrist angle θ are optimized simultaneously can be determined without the robot arm RA greatly avoiding the obstacle OB.

  In the posture change / interference judgment process (step a1), the provisional positioning process (step a11) for temporarily positioning the hand 8 in the gripping posture with respect to the object W was actually performed. In the state, the temporary positioning in-position determination / interference determination process (step a12) for determining the interference between the robot arm RA and the obstacle OB when the posture of each joint is variously changed by calculation within a possible range, After the positioning in-position change / interference determination process (step a12), the posture of the hand 8 is changed to perform the temporary positioning process (step a11) and the temporary positioning in-position change / interference determination process (step a12). It may include an iterative process (step a13) that is repeated in a range of all postures that the hand 8 can hold.

  In this way, after the hand 8 is temporarily positioned in the gripping posture with respect to the object W, the interference between the robot arm RA and the obstacle OB is determined. In addition, since the temporary positioning process (step a11) and the temporary positioning in-position change / interference determination process (step a12) are repeated only in the range of all the postures that the hand 8 can grip, Time can be reduced.

In the posture change, the interference determination step (step a1), in the wrist coordinate system set the coordinate origin O ₇ on the wrist joint J5～J7, set the object contact point H on the surface 8a of the hand 8, the provisional positioning In the process (step a11), the hand 8 is temporarily positioned in a state where the object contact point H is in contact with the object W, and the temporarily positioned hand is temporarily positioned in the temporary positioning in-position change / interference determination process (step a12). 8, the origin of the wrist coordinate system in the reference coordinate system is obtained as a wrist coordinate from the positional relationship between the object contact point H and the center of the gripping object W, and the obtained wrist coordinate and the center of the obstacle are The distance and the distance between the wrist coordinates and the center of the object W are respectively calculated, and whether the robot arm RA does not interfere with the obstacle OB and the grasped object W is determined according to a predetermined rule. You may make it judge.
Thus, in the state where the hand 8 is temporarily positioned in contact with the object W, the wrist coordinates are obtained, the distance between the wrist coordinates and the center of the obstacle OB, and the wrist coordinates and the center of the object W. And the interference between the robot arm RA and the obstacle OB can be easily and accurately determined.

Both the elbow height weight and the wrist angle weight may be 1.
In this case, even if there is no difference in weighting coefficients, an appropriate posture may be obtained. In such a case, the posture in which the elbow height _EZ and the wrist angle θ are optimized simultaneously can be determined more easily.

The posture teaching apparatus for a robot arm according to the present invention is connected to a hand 8 that grips an object W as a movable part that is sequentially connected via a joint, and to the hand 8 so that the posture can be changed via a wrist joint J5 to J7. The forearm 5 and the upper arm 3 connected to the elbow 4 of the forearm 5 via elbow joints J3 and J4 have redundancy in the posture of the hand 8 when the hand 8 grips the object W. An attitude teaching device for a robot arm RA that determines the attitude of each joint when the hand 8 grips the object W under a condition that avoids interference with an obstacle OB with respect to a multi-joint robot arm RA. There,
When the hand 8 is temporarily positioned in a gripping posture with respect to the object W and the posture of the temporary positioning is changed variously, the robot arm RA and the obstacle OB when the posture of each joint is changed variously. The posture change / interference determination means 18 for determining the interference with the robot, and a predetermined evaluation value in the case of taking the posture of the hand 8 and each joint at which no interference is caused by the posture change / interference determination means 18 are calculated. Evaluation value calculation means 19 for comparing the evaluation values, and an evaluation value comparison / teaching posture determination means 20 for teaching the posture of each joint to the posture of each joint having the smallest evaluation value by comparing the evaluation values,
In the evaluation value calculation means 19, a reference coordinate origin 0 ₀ is arbitrarily determined as a fixed coordinate system for the fixed portion of the base end of the robot arm RA, and two orthogonal axes in the horizontal direction (x ₀ axis, y ₀ axis). and vertical 1 axis (z ₀ axis) consider the reference coordinate system of the orthogonal three-axis directions of the coordinate axes, and Hijidaka of E _Z the vertical coordinates in the reference coordinate system of a predetermined point in the elbow joint J4 The angle formed by the hand 8 with respect to the forearm 5 in the horizontal plane (x ₀ , y ₀ plane) is the wrist angle θ, and the elbow height E _Z and the wrist angle θ are respectively weighted and the product thereof. The evaluation value is (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ).

According to this configuration, the evaluation value comparison / teaching posture determining means 20 compares the evaluation values and teaches the posture of each joint to the posture of each joint having the smallest evaluation value. Thus, the posture in which the elbow height _EZ, that is, the position energy and the wrist angle θ is optimized at the same time can be determined without the robot arm RA greatly avoiding the obstacle OB.

The robot arm posture teaching method according to the present invention includes a hand that grips an object as a movable portion that is sequentially connected via a joint, a forearm that is connected to the hand via a wrist joint so that the posture can be changed, and the forearm of the forearm The upper arm connected to the elbow via an elbow joint, and avoiding interference with an obstacle for a multi-joint robot arm having redundancy in the posture of the hand when the object is gripped by the hand A robot arm posture teaching method for determining a posture of the articulated joint when the hand grips an object under conditions,
A posture change / interference determination process for determining the interference between the robot arm and the obstacle when the posture of the hand is variously changed and the posture of each joint is variously changed within a range in which the hand can grip the object, An evaluation value calculation process for calculating a predetermined evaluation value in the case of taking the posture of the hand and each joint that is assumed to cause no interference in the posture change / interference determination process, and the evaluation value is the smallest by comparing the evaluation value Including the evaluation value comparison / teach posture determination process for teaching the posture of each joint,
For the evaluation value, as the coordinate system fixed with respect to the fixed portion of the proximal end of the robot arm, defining a reference coordinate origin 0 ₀ optionally, two orthogonal axes (x ₀ axis, y ₀ axis) in the horizontal direction and the vertical direction Considering a reference coordinate system in three orthogonal directions with one axis (z ₀ axis) as a coordinate axis, the vertical coordinate in the reference coordinate system of a predetermined point at the elbow joint is defined as an elbow height E _Z and a horizontal plane (x _The angle formed by the hand with respect to the forearm in the ₀ , y ₀ plane) is the wrist angle θ, and the elbow height _EZ and the wrist angle are weighted respectively, and the product is (elbow height weight × elbow) Since the height E _Z ) × (wrist angle weight × wrist angle θ) is used as an evaluation value, the labor and time of teaching by the operator are reduced, and the robot arm operating angle is prevented from undesirably increasing. Low position energy of robot arm posture It is possible to reduce the power required achieving.

  An embodiment of the present invention will be described with reference to FIGS. The robot arm posture teaching apparatus according to the present embodiment is applied to a multi-joint robot arm having seven-axis joints. However, the number of joints is not limited to 7 axes, and may be, for example, 8 axes or more. Moreover, 6 axes or less may be sufficient. The following description also includes a description of the robot arm posture teaching method. This articulated robot arm has a function of avoiding an obstacle and determining a posture when a target is gripped by a hand attached to the tip.

  As shown in FIGS. 1 and 2, the robot arm RA is attached to the arm mount 12. The robot arm RA includes the hand 8, a forearm 5 connected to the hand 8 through wrist joints J5 to J7 so that the posture can be changed, and an elbow portion of the forearm 5 as movable parts sequentially connected through joints. 4 has an upper arm 3 connected through elbow joints J3 and J4. The upper arm 3 is attached to the arm mount 12 via shoulder joints J1 and J2.

  The wrist joints J5 to J7 are joints having three degrees of freedom. Of the wrist joints J5 to J7, the wrist joint J5 is a joint that allows the forearm 5 to rotate around the axis 5i. The wrist joints J6 and J7 are joints that can change the inclination angle of the hand 8 with respect to the forearm 5 in two orthogonal axes. The wrist joint J6 is realized by the first wrist 6 connected to the forearm 5, and the wrist joint J7 is realized by the second wrist 7 connected to the first wrist 6. The second wrist 7 includes a bracket 7a formed in a concave shape, and is provided integrally with the bracket 7a so that the hand 8 protrudes from the bottom 7aa of the bracket 7a. The hand 8 is configured to be able to hold the object W.

The elbow joints J3 and J4 are joints having two degrees of freedom capable of changing the inclination angle in the two orthogonal axes. The shoulder joints J1 and J2 are joints that can change the inclination angle of the upper arm 3 with respect to the arm mounting base 12 in two orthogonal axes. The shoulder joint J1 is realized by the first shoulder 1, and the shoulder joint J2 is realized by the second shoulder 2. The articulated robot arm RA includes a motor (not shown) that operates the above-described units, and a posture teaching device 17 (to be described later) that controls the operation of each unit.
The respective axes of the joints J1 to J7 do not necessarily have to be angularly displaceable, and at least one of the joints J1 to J7 has an operation part such as an expansion / contraction operation or a parallel link operation. You may have.

Next, the posture teaching device 17 for the robot arm RA will be described.
As shown in FIG. 5, the posture teaching device 17 includes, for example, a central processing unit (CPU), a storage unit such as a memory, a computer such as a microcomputer having an input / output interface and the like (none of which are shown), This is realized by a teaching posture control program (not shown) to be executed by a computer. The posture teaching program is stored in the storage means. By this attitude control program, each means shown in a conceptual configuration in FIG. 5 is configured. The posture control program and an operation control program (not shown) for controlling the robot arm operation after teaching constitute a higher-level control device (not shown) of the posture teaching device 17.

  The posture teaching device 17 includes a posture change / interference determination unit 18, an evaluation value calculation unit 19, and an evaluation value comparison / teach posture determination unit 20. The posture change / interference determination means 18 includes a temporary positioning in-position change / interference determination unit 18a and a repetition process unit 18b. Among these, the temporary positioning in-position change / interference determining unit 18a variously changed the posture of each joint by calculation within a possible range in a state where the hand 8 is temporarily positioned in the gripping posture with respect to the object W. In this case, the interference between the robot arm RA and the obstacle W is determined. After determining the interference, the repetition process unit 18b changes the posture of the hand 8 to determine the re-temporary positioning of the hand 8 and the interference between the robot arm RA and the obstacle W. Repeat for all possible postures.

  The evaluation value calculation means 19 calculates a predetermined evaluation value when the posture of the hand 8 and each joint, which is assumed not to cause interference by the posture change / interference determination means 18, is taken. The evaluation value comparison / teaching posture determining means 20 compares the evaluation values and teaches the posture of each joint to the posture of each joint having the smallest evaluation value.

In the evaluation value calculation means 19, a reference coordinate origin 0 ₀ is arbitrarily determined as a fixed coordinate system for the fixed portion of the base end of the robot arm RA, and two orthogonal axes in the horizontal direction (x ₀ axis, y ₀ axis). Assuming a reference coordinate system in three orthogonal directions with a vertical axis (z ₀ axis) as a coordinate axis, the vertical coordinate in the reference coordinate system of a predetermined point at the elbow joint J4 is as shown in FIG. in the Hijidaka of E _Z, horizontal (x _0, y ₀ plane) the angle the hand 8 in the makes with the forearm 5, and the wrist angle θ as shown in FIG. 3, the elbow height E _Z and wrist Each angle is weighted, and the product (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ) is used as an evaluation value. That is, a value obtained by multiplying the value obtained by multiplying the elbow height _EZ by the elbow height weight that is the weighting coefficient and the value obtained by multiplying the wrist angle θ by the wrist angle weight that is the weighting coefficient is used as the evaluation value. .
An output value for executing the posture taught by the evaluation value comparison / teaching posture determination means 20 is transmitted to the motors via a drive circuit (not shown), and is used for posture control by driving these motors.

FIG. 6 is a flowchart representing the robot arm posture teaching method as a superordinate concept.
This posture teaching method determines the postures of the joints J1 to J7 when the hand 8 grips the object W under the condition of avoiding interference with the obstacle OB with respect to the articulated robot arm RA. This is a method of teaching the posture of the robot arm RA. This posture teaching method mainly includes a posture change / interference determination process (step a1), an evaluation value calculation process (step a2), and an evaluation value comparison / teach posture determination process (step a3).

  The posture change / interference determination process (step a1) includes a temporary positioning process (step a11), a temporary positioning in-position change / interference determination process (step a12), and a repetition process (step a13). After the start of this process, the process proceeds to the temporary positioning process (step a11), and the hand 8 is temporarily positioned in the gripping posture with respect to the object W. Next, the robot moves to the temporary positioning in-position change / interference determination process (step a12), and the posture of each joint is variously changed by calculation within a possible range in the state where the hand 8 is temporarily positioned. The interference between the arm RA and the obstacle OB is determined. Next, the process proceeds to an iterative process (step a13), the posture of the hand 8 is changed, and the temporary positioning process (step a11) and the temporary positioning in-position change / interference determination process (step a12) are performed. Repeat for all possible postures.

  Next, the process proceeds to the evaluation value calculation process (step a2), and the predetermined evaluation value in the case of taking the posture of the hand 8 and each joint that is assumed not to cause interference in the posture change / interference determination process (step a1). calculate. Thereafter, the process proceeds to the evaluation value comparison / teaching posture determination process (step a3), and the postures of the respective joints are taught to the postures of the respective joints having the smallest evaluation value by comparing the evaluation values. Thereafter, this process is terminated.

FIG. 7 is a flowchart showing the robot arm posture teaching method step by step.
The following coordinates are measured based on the switch operation by the operator or the above attitude control program. For example, when a cylindrical obstacle OB is erected on the support surface SF, an “obstacle center coordinate” that is an XY coordinate of the circle center 9 of the obstacle OB in plan view shown in FIG. It is measured in advance at a distance from the origin O _0. At the same time, for example, when the cylindrical object W is erected on the support surface SF, the “grip object center coordinates” that is the XY coordinates of the circle center 10 of the object W in the plan view shown in FIG. Measured at a distance from the reference coordinate origin O ₀ . This reference coordinate origin O ₀ is a coordinate origin arbitrarily determined as a fixed coordinate system for the fixed portion of the base end of the robot arm RA.

Then, the process proceeds to step S2, for obtaining a wrist coordinate to be described later, the hand surface 8a, the coordinates from the coordinate origin _{O 7} set to the wrist joint J7 (hx, hy, hz) sets the point H to be. Next, the process goes to step S3, where the point 8 set on the hand surface 8a is in contact with the object W as shown in FIG. 3, and the end face 16 of the hand 8 is the support surface as shown in FIG. Temporary placement to be horizontal with SF. These steps S1 to S3 correspond to a temporary positioning process.
Next, the process proceeds to step S4, and the coordinates (Wx, Wy) of the coordinate origin O ₇ viewed from the reference coordinate origin O ₀ are determined from the positional relationship between the point H in the temporarily placed hand 8 and the object center coordinates. , Wz) is determined and used as wrist coordinates.

  Next, the distance between the obtained wrist coordinates and the obstacle center coordinates and the distance between the wrist coordinates and the object center coordinates are respectively calculated, and it is confirmed whether the robot arm RA interferes with the obstacle OB and the object W. (Step S5). Thereafter, in step S6, it is determined whether there is interference according to a predetermined rule. For example, it is determined that the robot arm RA and the obstacle OB interfere with each other under a condition that the distance between the wrist coordinates and the obstacle center coordinates is equal to or less than a threshold value defined according to the type and size of the obstacle OB. Further, it is determined that the robot arm RA and the object W interfere with each other under a condition that the distance between the wrist coordinates and the object center coordinates is equal to or less than a threshold value defined according to the type and size of the object W, for example. . These steps S4 to S6 correspond to the temporary positioning in-position change / interference determination process.

  When the interference between the robot arm RA and the obstacle OB or the object W occurs (step S6: No), the process proceeds to step S7, and it is determined whether there is a temporary arrangement of the hand 8 that has not been calculated. . If it is determined that there is an uncalculated temporary arrangement (step S7: Yes), the process of returning to step S3 is repeated. This step S7 corresponds to an iterative process. When there is no interference in step S6 (step S6: Yes), the inclination angle of the shoulder joint J1 is set to a lower limit value within an inclination angle range that the shoulder joint J1 can take (step S8).

Next, the inclination angle of the shoulder joint J2 is calculated from the wrist coordinates and the set lower limit value of the inclination angle of the shoulder joint J1 (step S9). From both the inclination angles of the shoulder joints J1 and J2, so-called elbow coordinates (Ex, Ey, Ez) of the coordinate origin O ₄ of the elbow part 4 with respect to the reference coordinate origin O ₀ are calculated (step S10). Of the elbow coordinates, Ez represents the support surface SF, that is, the elbow height Ez from the ground.
Next, the distance between the obtained elbow coordinates and obstacle center coordinates is calculated. Further, a distance between the elbow coordinates and the object center coordinates is calculated. Thereby, it is confirmed whether or not the robot arm RA does not interfere with the obstacle OB and the object W (step S11).

  Thereafter, in step S12, it is determined whether there is interference according to a predetermined rule. For example, it is determined that the robot arm RA and the obstacle OB interfere with each other under a condition that the distance between the elbow coordinates and the obstacle center coordinates is equal to or less than a threshold value defined according to the type and size of the obstacle OB. Further, it is determined that the robot arm RA and the object W interfere with each other under a condition that the distance between the elbow coordinates and the object center coordinates is equal to or less than a threshold value defined according to the type and size of the object W.

Here, when interference occurs (step S12: No), the set inclination angle of the shoulder joint J1 is increased by a certain amount (step S13), and whether this inclination angle is within an inclination angle range that the shoulder joint J1 can take. Is determined (step S14). If it is not within the range of inclination angles that the shoulder joint J1 can take (step S14: No), the process of returning to step S7 is repeated.
When there is no interference in step S12 (step S12: Yes), the wrist angle θ is calculated from the obtained coordinates (hx, hy), coordinates (Wx, Wy), and elbow coordinates (Ex, Ey) (step S15). . As shown in FIG. 3, the wrist angle θ is an angle formed by the hand 8 with respect to the forearm 5 in the x ₀ , y ₀ plane, that is, the horizontal plane.

Next, each of the elbow height _EZ and the wrist angle θ is weighted, and the product (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ) is calculated as an evaluation value. (Step S16). This “evaluation value” is expressed as “evaluation function” in step 16 of FIG. This step S16 corresponds to an evaluation value calculation process. In the present embodiment, the weight of the elbow height _EZ is “1”, and the weight of the wrist angle θ is “1”. However, these weightings are not limited to “1”. Further, the weight of the elbow height _{EZ and} the weight of the wrist angle θ may be different.

Next, the obtained evaluation value is compared with the evaluation value previously stored in the storage means or the like (step S17). When the obtained evaluation value is smaller than the previously stored evaluation value (step S18: Yes), the obtained shoulder angle J1, J2 angle, elbow height E _Z , and wrist angle θ are stored in the storage means. (Step S19). Thereafter, a process of increasing the inclination angle of the shoulder joint J1 by a certain amount within a possible range is executed (step S13, step S14: Yes), and the process of returning to step S9 is repeated. If the inclination angle range exceeds the inclination angle range that can be taken by the shoulder joint J1 (step S14: No), the process of returning to step S3 through step S7 is repeated.

When the calculation is completed for all possible postures in which the hand 8 can hold the object W (step S7: No), the stored both inclination angles of the shoulder joints J1 and J2 and the elbow height are stored. E _Z and wrist angle θ are optimum postures of the robot arm RA. From the inclination angles of the shoulder joints J1 and J2, the elbow height E _Z , and the wrist angle θ, the inclination angles of the remaining elbow joints J3 and J4 and wrist joints J5 to J7 are calculated, and the inclination angles of all the joints are calculated. Obtained (step S20). Thereafter, this process is terminated.

According to the posture teaching method and posture teaching device of the robot arm RA described above, after the hand 8 is temporarily positioned, the coordinate origin O ₇ set to the wrist joint J7 is calculated, and the robot arm is based on the coordinate origin O _7. The presence or absence of RA interference is determined. By determining that there is no interference, the inclination angles of the shoulder joints J1 and J2 are defined, and the coordinate origin O ₄ of the elbow ₄ is calculated.
Based on the coordinate origin O 4 of the elbow ₄ or the like, the presence or absence of interference of the robot arm RA is further determined. The wrist angle θ is calculated based on the determination that there is no interference, and then the elbow height E _Z and the wrist angle θ that minimize the evaluation value are obtained. Thus, the posture in which the elbow height _EZ, that is, the position energy and the wrist angle θ is optimized at the same time can be determined without the robot arm RA greatly avoiding the obstacle OB.

The robot arm RA has wrist joints J5 to J7, and these wrist joints J5 to J7 are joints having three degrees of freedom. Among these, the wrist joint J5 is a joint that allows the forearm 5 to rotate around the axis, and the wrist joints J6 and J7 can change the inclination angle of the hand 8 in the two orthogonal axes with respect to the forearm 5, respectively. It is a joint to do. Furthermore, the elbow joints J3 and J4 are joints having two degrees of freedom that can change the inclination angle in the two orthogonal axes. Therefore, the robot arm RA has posture redundancy due to such multi-joints, and can realize a large number of gripping postures. At this time, the posture in which the elbow height _EZ and the wrist angle θ are optimized simultaneously can be determined without the robot arm RA greatly avoiding the obstacle OB.

  The posture change / interference determination process (step a1) includes the temporary positioning process (step a11), the temporary positioning in-position change / interference determination process (step a12), and the repetition process (step a13). Therefore, it is possible to determine the interference between the robot arm RA and the obstacle OB after temporarily positioning the hand 8 in the gripping posture with respect to the object W. In addition, the temporary positioning process (step a11) and the temporary positioning in-position change / interference determination process (step a12) are repeated only in the range of all the postures that can be gripped by the hand 8. Time can be reduced.

Further, wrist coordinates (Wx, Wy, Wz) are obtained in a state where the hand 8 is temporarily positioned in contact with the object W, and the wrist coordinates (Wx, Wy, Wz) and the center of the obstacle OB are obtained. The distance and the distance between the wrist coordinates (Wx, Wy, Wz) and the center of the object W are respectively calculated, and interference between the robot arm RA and the obstacle OB is easily and accurately according to predetermined rules such as threshold judgment. Can be judged.
Since both the elbow height weight and the wrist angle weight are set to 1, an evaluation value with no difference in weighting coefficient can be obtained. Therefore, the posture in which the elbow height Ez and the wrist angle θ are optimized at the same time can be determined more reliably.

  Also, the distance between the coordinates (hx, hy, hz) and the object center coordinates or the distance between the coordinates (hx, hy, hz) and the obstacle center coordinates is calculated, and the robot arm RA, the obstacle OB, and the object. When it is determined that there is interference in the process of determining the presence or absence of interference with W (step S12), the temporary position of the hand 8 is released, and the process returns to the process of temporarily positioning the hand 8 at a new temporary position (step S3). . By repeatedly executing the determination of the presence or absence of interference based on the temporary position of the hand 8, it is possible to prevent the operation angle of the robot arm RA from being undesirably increased.

The evaluation value is stored in the storage means, and the elbow height weight × elbow height E _Z , wrist angle weight × wrist angle θ calculated in the process of obtaining the elbow height E _Z and the wrist angle θ When the calculated evaluation value is smaller than the stored evaluation value, the calculated elbow height weight × elbow height E _Z and wrist angle weight × wrist angle θ are newly stored in the storage means as evaluation values. The required power can be reliably reduced by reducing the positional energy of the posture. Thereby, for example, each motor can be made to have a low rated output, and the manufacturing cost can be reduced.

When it is determined in step S12 that there is interference, the set inclination angle of the shoulder joint J1 is increased by a fixed amount, and the coordinate origin O ₄ of the elbow ₄ is calculated based on the increased inclination angle. As described above, since the inclination angle of the shoulder joint J1 is increased by a certain amount until it is determined in step S12 that there is no interference, it is possible to reliably prevent the robot arm RA from largely avoiding the obstacle OB. It is possible to prevent the operating angle from undesirably increasing in order to realize the posture of the arm RA.

1 is a perspective view schematically illustrating a configuration of a robot arm according to an embodiment of the present invention. It is a front view showing the positional relationship between an object and an obstacle and the robot arm. It is a top view showing the positional relationship between an object and an obstacle and the robot arm. It is an expansion perspective view of the hand of the robot arm. It is a block diagram showing the electric constitution of the attitude | position teaching apparatus of the robot arm. It is a flowchart which expresses the attitude | position teaching method of the robot arm as a high-level concept. It is a flowchart showing the posture teaching method of the robot arm in steps.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... 1st shoulder 2 ... 2nd shoulder 3 ... Upper arm 4 ... Elbow part 5 ... Forearm 6 ... 1st wrist 7 ... 2nd wrist 8 ... Hand 17 ... Posture teaching apparatus 18 ... Posture change and interference judgment means 18a ... Temporary positioning Inner posture change / interference determination unit 18b ... iteration process unit 19 ... evaluation value calculation unit 20 ... evaluation value comparison / teaching posture determination unit J1-J7 ... joint W ... target OB ... obstacle

Claims (6)

As movable parts that are sequentially connected via joints, a hand 8 that grips the object W, a forearm 5 that is connected to the hand 8 via wrist joints J5 to J7 so that the posture can be changed, and an elbow part 4 of the forearm 5 The multi-joint robot arm RA having redundancy in the posture of the hand 8 when the hand 8 grips the object W with the upper arm 3 connected to the arm 8 through elbow joints J3 and J4. The robot arm RA posture teaching method for determining the posture of the multi-joint when the hand 8 grips the object W under the condition of avoiding interference with OB,
Posture change / interference for judging the interference between the robot arm RA and the obstacle OB when the posture of the hand 8 is changed variously within the range in which the hand 8 can grip the object W and the posture of each joint is changed variously. Judgment process (step a1),
An evaluation value calculation process (step a2) for calculating a predetermined evaluation value in the case of taking the posture of the hand 8 and each joint in which no interference occurs in the posture change / interference determination process (step a1);
An evaluation value comparison / teaching posture determination process (step a3) for teaching the posture of each joint to the posture of each joint having the smallest evaluation value by comparing the evaluation values,
Including
About the evaluation value,
As a fixed coordinate system for the fixed portion of the base end of the robot arm RA, a reference coordinate origin 0 ₀ is arbitrarily determined, and two orthogonal axes in the horizontal direction (x ₀ axis, y ₀ axis) and one axis in the vertical direction (z Considering a reference coordinate system in three orthogonal directions with the coordinate axis ( ₀ axis) as the coordinate axis,
The vertical coordinates in the reference coordinate system of a predetermined point in the elbow joint J3, J4 and Hijidaka of E _Z,
An angle formed by the hand 8 with respect to the forearm 5 in a horizontal plane (x ₀ , y ₀ plane) is a wrist angle θ,
Weight each elbow height _EZ and wrist angle θ,
The product is (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ) as an evaluation value.
A robot arm posture teaching method characterized by the above.   2. The robot arm RA according to claim 1, wherein the wrist joints J5 to J7 can change the inclination angle of the hand 8 in the two orthogonal axes with respect to the forearm 5, and the forearm 5 rotates around the axis 5i. A robot arm posture teaching method, which is a joint having three possible degrees of freedom and the elbow joints J3 and J4 are joints having two degrees of freedom in which the inclination angle can be changed in two orthogonal axes. In claim 1 or claim 2, the posture change / interference determination process (step a1) includes:
Actually, a temporary positioning process (step a11) for temporarily positioning the hand 8 in a gripping posture with respect to the object W;
Temporary positioning in-position change / interference determination process for determining interference between the robot arm RA and the obstacle OB when the posture of each joint is variously changed by calculation within the possible range with this temporary positioning performed (step) a12)
After this temporary positioning in-position change / interference determination process (step a12), the posture of the hand 8 is changed to perform the temporary positioning process (step a11) and the temporary positioning in-position change / interference determination process (step a12). A repeating process (step a13) that repeats in a range of all postures that can be held by the hand 8.
Robot arm posture teaching method. In claim 3, in the posture change / interference determination process (step a1),
In the wrist coordinate system set the coordinate origin O ₇ on the wrist joint J5～J7, set the object contact point H on the surface 8a of the hand 8,
In the temporary positioning process (step a11), the hand 8 is temporarily positioned in a state where the object contact point H is in contact with the object W,
In the temporary positioning in-position change / interference determination process (step a12), from the positional relationship between the object contact point H of the temporarily positioned hand 8 and the center of the gripping object W, the wrist coordinate system in the reference coordinate system is determined. The origin is obtained as wrist coordinates, the distance between the obtained wrist coordinates and the center of the obstacle OB, and the distance between the wrist coordinates and the center of the object W are calculated, respectively, and the robot arm RA calculates the obstacle OB and the object to be grasped. Judging whether or not to interfere with the object W according to a predetermined rule,
Robot arm posture teaching method.   The robot arm posture teaching method according to any one of claims 1 to 4, wherein the elbow height weight and the wrist angle weight are both 1. As movable parts that are sequentially connected via joints, a hand 8 that grips the object W, a forearm 5 that is connected to the hand 8 via wrist joints J5 to J7 so that the posture can be changed, and an elbow part 4 of the forearm 5 The multi-joint robot arm RA having redundancy in the posture of the hand 8 when the hand 8 grips the object W with the upper arm 3 connected to the arm 8 through elbow joints J3 and J4. A robot arm posture teaching device that determines the posture of each joint when the hand 8 grips the object W under a condition that avoids interference with OB,
When the hand 8 is temporarily positioned in a gripping posture with respect to the object W and the posture of the temporary positioning is changed variously, the robot arm RA and the obstacle OB when the posture of each joint is changed variously. Posture change / interference determining means 18 for determining interference with
An evaluation value calculation means 19 for calculating a predetermined evaluation value when the posture of each of the hand 8 and each joint that is assumed to cause no interference in the posture change / interference determination means 18;
An evaluation value comparison / teaching posture determining means 20 for teaching the posture of each joint to the posture of each joint having the smallest evaluation value by comparing the evaluation values,
Including
In the evaluation value calculation means 19,
As a fixed coordinate system for the fixed portion of the base end of the robot arm RA, a reference coordinate origin 0 ₀ is arbitrarily determined, and two orthogonal axes in the horizontal direction (x ₀ axis, y ₀ axis) and one axis in the vertical direction (z Considering a reference coordinate system in three orthogonal directions with the coordinate axis ( ₀ axis) as the coordinate axis,
The vertical coordinates in the reference coordinate system of a predetermined point in the elbow joint J3, J4 and Hijidaka of E _Z,
An angle formed by the hand 8 with respect to the forearm 5 in a horizontal plane (x ₀ , y ₀ plane) is a wrist angle θ,
Weight each elbow height _EZ and wrist angle θ,
The product is (elbow height weight × elbow height E _Z ) × (wrist angle weight × wrist angle θ) as an evaluation value.
A posture teaching apparatus for a robot arm.

Images