JPH0523982A - Robot control method and device - Google Patents

Abstract

PURPOSE:To draw out the capacity of a robot to the maximum and improve the working quality by controlling each freedom degree displacement of the robot to make a control specification to each direction in a space easily realizable according to the contents of robot operation. CONSTITUTION:Specifications for robot operation are described with a coordinate system fixed to a working space (101), an operation control target in each coordinate axial direction and a coordinate axis round direction is clarified (102), and a robot operation control capacity in each of these directions is found out as a function of attitude states of the robot (103). In succession, a control performance evaluation index as to working specifications in each direction corresponding to each of the robot attitude states is calculated (104), while the robot attitude state, where these weighted overall evaluation index values come to the largest value, is set down to an optimum attitude state (107), with which working is executed (108). With this constitution, since a proper robot attitude conformed to the work specifications is adoptable, working time is shortened and, what is more, working quality is improved.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control method and apparatus for a robot system such as a robot arm and a robot hand, and in particular, when the work is performed using a robot, the control ability required for the work is comprehensively obtained. The present invention relates to a method and apparatus for controlling the displacement of each degree of freedom of the robot so that it can be most effectively exhibited.

[0002]

2. Description of the Related Art Robots are used for various kinds of work, and the position of a work object to be placed on the robot depends on the contents of the work performed by the robot. It is a very important task to arrange the work object so as to maximize the ability of the robot, paying attention to the kinematic characteristics of the robot. Several indicators have been proposed for the purpose of evaluating such things.

One of them is the following two documents, i) T. Y
oshikawa: Analysis and control of robot manipu
lators with redundancy, Robotics Research, th
e First International Symposium, MIT Pr
ess (1984) pp735-747 and ii) Yoshikawa: The operability of the robot arm, which is a concept of operability proposed in the Journal of the Robotics Society of Japan, Volume 2, No. 1, pp63-67 (1984). This considers a set consisting of all the spatial velocities that the hand of the robot can give when the upper limit of the velocity that each degree of freedom of the robot can give is given, and this set, that is, the dimension of the velocity to be controlled by the robot. The manipulability is defined as the volume of an ellipsoid in the Euclidean space with. In other words, the speed that each of the robot's degrees of freedom can set is set, and the speed of the hand can be calculated by considering how freely the robot's hand position / orientation can be manipulated in all directions. It can be said that it is expressed as the sum total.

On the other hand, as another index, the following documents, iii) J.
K. Salisbury and JJ. Craig Articulated
Hands, Force Control and Kinematic Issues,
Internatonal Journal on Robotics Researc
h, Vol. 1, No. 1, pp. 4 to 17 (1982), it is stated that the error with respect to the force generated by the hand of the robot based on the torque error generated in each degree of freedom of the robot when performing force control is as uniform as possible. As a goal, an index showing this uniformity is proposed. That is, the index indicates the directional uniformity of the ellipsoid by using the ratio of the maximum radius and the minimum radius of the ellipsoid.

Comparing these two indices, the manipulability of the former is irrespective of the direction of movement of the robot, and the larger the total sum of easiness of movement of the hand, assuming any movement in any direction, that is, the arbitrary degree. The index has a larger value as the expected value of the operation speed with respect to the direction is larger, and is an evaluation index in which it is in a desirable state. On the other hand, the latter index using the ellipsoidal radius ratio takes a larger value as the error in the force generated at the robot's hand is as equal as possible in any direction, that is, the force isotropic and uniform. It is an evaluation index that indicates a desirable state.

Therefore, from the viewpoint of controlling the robot, control is performed so as to take a posture in which any one of these evaluation indexes is maximized or maximized, or the relative relationship between the work object and the robot is controlled. That was being done.

[0007]

Controlling the robot on the basis of the above-mentioned indexes means that the robot takes an optimum posture in a general sense regardless of the contents of the work performed by the robot. The robot's state and posture are excellent in either the expected value for the ease of movement in that direction or the independence of direction, even if the movement is specified in any direction. The control will be performed so that the robot takes. However, when the individual work performed by the robot is considered concretely, it is clear that it is not always necessary to consider the performance of motions in all directions as a problem.

For example, the robot 1 shown in FIG. 2 moves the work tool 19 along the surface of the work object 2 in the vertical direction 3 (Z
When moving in the left-right direction 4 (X-axis direction) while keeping the force (axial direction) constant, it is necessary to control the position of the robot in the X-axis direction, while in the Z-axis direction. It is necessary to control the force of the robot. in this way,
In the case where the specifications required for the control for executing the work differ depending on the direction, it is impossible to derive the optimum posture state by the above-mentioned method that has been conventionally proposed. That is, the expected value of the position controllability in all directions or the uniformity of force control is unsuitable as an evaluation index. For example, in the example of FIG. 1, the greater the linear movement speed with respect to the speed that each degree of freedom of the robot can produce in the left-right direction, that is, the X-axis direction, and the higher the linear movement speed in the up-down direction, that is, the Z-axis direction. The smaller the error in the linear direction generated force with respect to the output torque error in the degree of freedom, the more appropriate it is, and it is desirable to perform control so as to realize such a posture of the robot.

In this sense, the present invention determines the proper posture state of the robot, which is not in a general form as in the conventional method, but in relation to the specification of each work performed by the robot. In order to maximize the performance capability of the robot. Further, it is intended to provide a method of determining and controlling an optimum robot posture state in a system including a robot arm and a hand or a robot having so-called redundant degrees of freedom by using such an optimization method. Is. By performing such control, it is possible to perform the work while maximizing the performance capability of the robot, so that the energy consumption of the robot in performing the same work is reduced, or the work operation is accelerated. It is possible to improve the accuracy of motion locus and force control.

[0010]

In order to achieve the above object, the present invention first focuses on the work to be performed by the robot, clarifies the specifications and purpose of the work itself, and sufficiently improves the performance of the robot accordingly. The feature is that the posture of the robot is determined and controlled so that it can be demonstrated. That is, the work specifications are described using a coordinate system fixed to the work space, the motion control target in each coordinate axis direction and the coordinate axis direction is clarified, and the motion control ability of the robot in each relevant direction is defined by the robot. Of the robot, the control performance evaluation index for the work specification in each direction corresponding to each of the robot attitude states is calculated, and the weighted overall evaluation index value of the robot attitude state becomes the largest value. Is determined as the optimum posture state. Rather than determining the "optimum" posture state for the robot independently of the work performed by the robot as in the past, it depends on the content of the work performed by the robot and the relative positional relationship between the robot and the work target. Then, the posture state of the robot that best matches the work specification is determined as "optimum", and the robot is controlled based on the result.

The degree of conformity with the work specifications mentioned here will be described below. In the example shown in FIG. 1 described above, the trajectory is controlled by performing position control in the left-right direction, that is, the X-axis direction. Therefore, the greater the speed that can be taken out in the X-axis direction with respect to the speed of each degree of freedom of the robot 1, the easier the robot will move. Further, force control is performed in the vertical direction, that is, the Z-axis direction, and the control purpose is to press the work tool against the work target 2 with a constant force. Therefore, the smaller the force error in the Z-axis direction with respect to the torque error of each degree of freedom of the robot 1, the more accurately the robot can control the force. Considering the above things in combination, the posture state of the robot is as shown in FIG. 1 in which it is easy to increase the speed in the X-axis direction and to increase the accuracy of the force generated in the Z-axis direction. It is desirable when making work. 2 for the X-axis direction and Z-axis direction described here
It is clear that the optimal robot states for each of the two control goals do not necessarily match. Therefore, as the work specifications, the importance of the work specifications in each axial direction (in the example of FIG. 2, X-axis and Z-axis directions) is considered, and by using this, the “optimal” for the work is comprehensively achieved. The robot posture is determined and the robot is controlled based on the result.

The above-described method for determining and controlling the posture of the robot according to the work is useful for optimizing the placement of the robot and the work object, but when considered as a robot system, , In the case of a system having so-called redundancy, in which the state of the robot satisfying the work specifications can be selected from at least a plurality of states independently of the relative relationship between the work tool and the work object. Especially effective.
For example, this corresponds to a case where work is performed by a combination of a robot arm and a robot hand attached to the tip thereof, or a case where a robot arm having seven or more degrees of freedom is used. For example, in a system in which a robot arm and a robot hand are combined, the posture of the robot hand is determined using the method according to the present invention, and the position of the robot hand is corrected on the robot arm side so as to be a desired position. It may be handled separately in the form of alignment, or by the method according to the present invention while keeping the hand position and posture constant with respect to the robot or robot system having multiple degrees of freedom. It is also possible to collectively determine the posture state of the robot as a whole.

[0013]

The specifications of each work to be carried out are described using a coordinate system fixed to the work space, and the motion control target in each coordinate axis direction and around the coordinate axis is clarified. Furthermore, the motion control ability of the robot in each coordinate axis direction and the coordinate axis direction corresponding to the clarified work specification, such as position control ability, rotation angle control ability, force control ability, speed control ability, etc. It is calculated as a function of posture. Then, a control performance evaluation index relating to work specifications in each of the directions corresponding to each of the posture states of the robot is calculated based on the obtained motion control capability of the robot, and the calculated control performance evaluation index is further calculated. Weighting is performed to calculate a weighted comprehensive evaluation index value. This weighted comprehensive evaluation index value represents what kind of posture of the robot is most suitable for the work to be performed, and the posture state of the robot having the largest weighted comprehensive evaluation index value is , The optimum posture state based on the weighting selected for the work to be performed.

By making the robot work in such a posture, the robot can maximize its performance according to the contents of the work. That is, it is possible to control the robot in a posture state in which required motion speed, accuracy, etc. are easily obtained according to the content of the work, in other words, a displacement state of each degree of freedom of the robot.

[0015]

First, one embodiment of the present invention will be described in a generalized form. FIG. 3 shows a robot 1 having 6 degrees of freedom and its work target 2. The position of the tip point 10 of this robot 1 and the robot 1 of the coordinate system 30 fixed to the tip point 10
Rotation angle displacement component P with respect to the coordinate system 3 fixed to the base of
_{1 to} P ₆ are represented as a vector P in the form of a vector. That is,

[0016]

[Equation 1]

Further, the displacements θ _{1 to} θ ₆ of the respective degrees of freedom 11 to 16 of the robot 1 are represented as a vector θ in the form of a vector. That is,

[0018]

[Equation 2]

At this time, the geometrical relationship between the two is given by the following equation (3),

[0020]

[Equation 3]

It is assumed that the relationship between the velocity vector K corresponding to the vector P and the velocity vector M of each of the degrees of freedom 11 to 16 of the robot 1 is given by the following equation (4).

[0022]

[Equation 4]

Here, the first term on the right side of the equation (4) is the Jacobian matrix on the right side of the equation (3) and is defined by the following equation (5).

[0024]

[Equation 5]

On the other hand, each degree of freedom 1 of the robot 1 shown in FIG.
_{1 to} 16 torques τ _{1 to} τ ₆ in the form of a vector τ
Express. That is,

[0026]

[Equation 6]

Further, the force and torque at the hand point 10 of the robot 1 are similarly expressed as a vector F in the form of a vector.
That is,

[0028]

[Equation 7]

At this time, the relationship between the two is given by the following equation (8) by using the concept of virtual work.

[0030]

[Equation 8]

That is, the force / torque relationship is derived from the position / orientation relationship by using the Jacobian matrix of the first term on the right side of expression (4), and its transposed matrix and its inverse matrix shown in expression (9) below. Given by.

[0032]

[Equation 9]

In the above example, the number of degrees of freedom of the robot and the number of orthogonal spatial parameters to be controlled are the same, but in the case where the number of degrees of freedom of the robot has more redundant degrees of freedom than the number of control parameters, , A generalized inverse matrix or a pseudo inverse matrix is used instead of the above inverse matrix to derive the same relation.

Now, assuming the work to be performed by the robot 1, its specifications are, for example, a position or a position in the X and Y axis directions (P ₁ and P ₂ axis directions) and around the X and Y axes (P ₄ and P ₅ directions). It is assumed that it is required to control the angle and the force or torque in the Z axis direction (P ₃ axis direction) and around the Z axis (P ₆ direction), respectively. In addition, it is desired to make the work as easy as possible in the P ₁ , P ₂ , P ₄ , P ₅ axis directions and to control the P ₃ , P ₆ axis directions as accurately as possible. . Furthermore, the importance of the requirement should be considered equally for position and for force. In this case, the following expression (10) is established as a relational expression regarding the position or the angle.

[0035]

[Equation 10]

As the relational expression regarding the force / torque, if the inverse matrix of the Jacobian matrix is the following expression (13),
Since the transposed matrix is the following equation (14), the following equation (15) is established.

[0037]

[Equation 11]

Therefore, a partial vector consisting only of elements relating to the axes (P ₁ , P ₂ , P ₄ , P _{5 in} this example) whose position and orientation should be controlled

[0039]

[Equation 12]

And a partial vector consisting only of elements relating to the axes (P ₃ and P ₆ axes in this example, that is, F ₃ and F ₆ ) whose force and torque should be controlled.

[0041]

[Equation 13]

Considering each of these partial vectors, the displacement of each degree of freedom of the robot 1, that is, the property as a function of the posture state vector θ of the robot is examined.

At this time, the respective degrees of freedom 11 to 1 of the robot 1
If the upper limit of the operation speed of 6 is given, the range of values that K can take is determined. Considering the norm of K,

[0044]

[Equation 14]

Thus, if the range of values in which ‖M‖ can move is determined, the set of K can be defined as an ellipsoid in the four-dimensional acreed space. Here, αi is assumed to be an appropriate constant for conversion of linear velocity and rotational velocity.

On the other hand, if the upper limit value of the torque error generated in each of the degrees of freedom 11 to 16 of the robot 1 is given, the range of values that the vector f can take is determined. The norm of vector f is

[0047]

[Equation 15]

Thus, if the range of values in which ‖τ‖ can move is determined, a set of vectors f can be defined as an ellipsoid in a similar acreed space. Here, βi is an appropriate constant similar to αi. However, in this example, the vector f is a two-dimensional vector, and a set of the vectors f can be considered as the area of the ellipse in the plane.

In view of the work specifications in this embodiment, each displacement vector θ of the degree of freedom of the robot 1, that is, the left side of the equation (18) and the left side of the equation (19) are the largest, that is, the robot. The posture of the robot 1 is found, and each degree of freedom 11 of the robot 1 is set so as to satisfy this state as much as possible.
It can be said that controlling ˜16 can be expected to give good results in performing a given task. However, the left side of the equation (18), that is, the vector θ that maximizes ‖K‖ and the vector θ that maximizes the left side of the equation (19), that is, the ‖ vector f‖, do not always match, but in general Are usually inconsistent. Therefore, it is necessary to select and determine the vector θ which is considered to be optimum in consideration of the balance between the above-mentioned ‖K‖ and the ‖ vector f‖. Now, note that ‖K‖ and ‖ vector f‖ have different physical meanings, and normalization is performed so that they can be compared at the same level. The maximum values of ‖K‖ and ‖ vector f‖ for the entire domain of the displacement vector θ of each degree of freedom 11 to 16 of the robot 1 are
Let ‖K‖max and ‖ vector f‖max respectively. However, in the domain of the vector θ, the point where the determinant of the Jacobian matrix shown in the equation (5) is 0, that is,

[0050]

[Equation 16]

In other words, if there is a singular point of this robot, the Jacobian matrix is no longer regular,
Since the inverse matrix shown in the equation (13) is not defined, the neighborhood of such a singular point is excluded from the domain of the vector θ to determine ‖K‖max and ‖vector f‖max. This means that, in the vicinity of the singular point, it is theoretically impossible for the robot 1 to make a linear movement, and it is appropriate to exclude this region from consideration. At this time, the normalized velocity vector and force / torque vector are expressed by the following equations (21) and (22), respectively.

[0052]

[Equation 17]

Here, considering the specifications of the given work, the performance in the position control direction is (2
It can be said that the larger the normalized velocity vector in the equation (1) is, and the smaller the normalized force / torque vector in the equation (22) is, the more preferable the performance in the force control direction is. Therefore, as an evaluation index of position control performance,
The scalar amount Vp is defined by the following formula (23), and the scalar amount Vf is defined by the following formula (24) as an evaluation index of the force control performance.

[0054]

[Equation 18]

The left side of the equation (21) and the left side of the equation (22) are normalized vectors, and the maximum norm is 1. Therefore, the range of the values of Vp and Vf is 0 ≦ Vp, Vf ≦ 1, and it can be said that the larger the values of Vp and Vf are, the more desirable it is for the work target of the present embodiment.

Here, in order to further consider the balance between the position control and the force control, the importance m for weighting is introduced as a parameter. That is, when the position / orientation and the force / torque are controlled at the same time as in the present embodiment, it is considered that one of them is more important than the other. This importance is not the importance with respect to the target specification itself of the work, but is merely for determining the posture state of the robot 1 that easily achieves the target, or for determining the relative positional relationship between the robot 1 and the work target 2. , The position / orientation control or the force / torque control should be prioritized. Now, for example, if the importance level m is defined as the priority ratio of the position control with respect to the force control, then S = (m · Vp + Vf) / (m + 1) (25) as the evaluation index S of the overall control performance. ) Can be defined. Therefore, by selecting the posture state vector θ of the robot 1 which maximizes the evaluation index S and controlling the robot 1 so as to match the posture state vector θ as much as possible, the position of the target work specification It is possible to determine the most suitable posture state, that is, the displacement vector θ of each of the degrees of freedom 11 to 16, in consideration of the balance between the control performance and the force control performance. FIG. 1 is a procedure diagram showing an example of the control flow.

In FIG. 3, for example, assuming that the work object 2 is moved and conveyed to the robot 1 at a constant speed v by a belt conveyor or the like (not shown), the moving position of the work object 2 is While monitoring, the work may be executed when the position comes close to the optimum posture state vector θ of the robot 1. FIG. 4 shows the control device 3 in that case.
A configuration example of is shown.

The control device shown in the figure calculates based on a work specification setting section 31 to which a prespecified work specification is input, a work specification connected to the work specification setting section 31 and a mechanical model of the robot 1. The control performance evaluation index calculation unit 32 that performs the above is analyzed, and the calculation result of the control performance evaluation index calculation unit 32 is analyzed to output the robot posture vector θ and the corresponding robot hand position vector P that are most suitable for the work specifications. Optimal posture state determination unit 33 and moving work object 2
Work object position detection unit 3 for detecting the position of (not shown)
5 is compared with the output of the work target position detection unit 35 and the output of the optimum posture state determination unit 33 (for example, the hand position vector P of the robot 1), and the absolute value of the difference is separately specified as an allowable value. In the following, the comparison unit 34 that generates a start signal to the robot 1 and the robot motion control unit 36 that receives the generated start signal and drives and controls the robot 1 are configured.

In the control device having the above structure, the work object position detecting section 35 detects the position of the moving work object 2, and when the work is performed in that state based on the detected position of the work object 2. The robot posture vector θ and the corresponding robot hand position vector P are output. Comparison unit 34
Is the optimum robot posture vector θ output by the optimum posture state determination unit 33 and the corresponding robot hand position vector P, and the robot posture vector θ output by the work target position detection unit 35 and the corresponding robot hand position vector P. If the absolute values of the differences are compared and are equal to or less than the separately specified allowable value, it is determined that the work may be performed at the detected position of the work target object, and a start signal to the robot 1 is generated. As a result, the robot 1 is
It is possible to perform the work in a posture state close to the posture state suitable for the specifications of the work. The robot motion control unit 36 that drives and controls the robot 1 in response to the generated start signal is not different from the conventional robot control device, and therefore the description thereof is omitted here.

In the above description, the case where the number of parameters of the work specification and the number of degrees of freedom of the robot are both 6 has been described, but these are not necessarily limited to 6 and the work specification parameter The number and the degree of freedom of the robot do not have to match. Also,
The case where the number of axes for which the position control is required is 4 and the number of axes for which the force / torque control is required is 2 has been illustrated, but it goes without saying that the number is not limited to this. In addition, as for the method of obtaining the normalized vector,
It is also possible to use the average value or the minimum value of the K || / f vector f |, and the norm can be appropriately selected according to the method of defining the distance in space. As the control performance evaluation index, Vp, Vf, 1 / (1-Vf), or the like can be used depending on the case. Furthermore, it is also possible to use a geometric mean or a product of appropriate parameters as the comprehensive evaluation index S. As described here, the evaluation index can be set in a modified form of the present embodiment, and is not limited to the contents of the present embodiment.

Next, as an application example of the present invention, a robot
The cooperation system between the arm and the robot hand is described. FIG. 5 is a block diagram of the system, showing a robot arm 1 having 6 degrees of freedom, that is, joints 11 to 16.
The robot hand 5 is attached to the tip of the robot hand, and the robot hand 5 is provided with two fingers 51 and 52. Each finger has two degrees of freedom, and the finger 51 is provided with joints 51a and 51b, and the finger 52 is provided with joints 52a and 52b.
If the work object 6 is grasped by the two fingers 51 and 52 of the hand 5, the position of the work object 6 is
In order to control the posture in space, 6 degrees of freedom are required, whereas the system has 10 degrees of freedom as a whole. Now, the fingers 51 and 52 are the main shaft 50 of the robot hand 5 or the main shaft 60 (5 of the work object 6).
The total number of degrees of freedom is 8 if the restriction is made that the object operates symmetrically with respect to (matching 0). That is, the system has a redundancy degree of freedom.

Now, consider the operation of inserting the work object 6 into the hole 71 provided in another work object 7 by this robot system. At this time, as the specifications of the work, the position is controlled in the direction along the main axis 60 of the hole 71 or the work target 6, and the force is controlled in the direction orthogonal thereto. Therefore, if one pays attention only to the finger 51 having two degrees of freedom, the link mechanism shown in FIG. 6 may be considered. As shown in FIG. 6, the angle θ ₁ ,
If θ ₂ is determined, the X and Y axes are determined, and the length between the joints 51a and 51b is l ₁ and the length between the joint 51b and the finger tip 51c is l ₂ , then X = l ₁ cos θ ₁ + l ₂ cos (θ ₁ + θ ₂ ) (26) Y = l ₁ sin θ ₁ + l ₂ sin (θ ₁ + θ ₂ ) (27) holds, and the Jacobian matrix is

[0063]

[Formula 19]

It becomes Therefore, the above equation (28) is

[0065]

[Equation 20]

When expressed as follows,

[0067]

[Equation 21]

Therefore, the following equations (31) to (34) are obtained.

[0069]

[Equation 22]

In the above case, det (J)
= L ₁ l ₂ sin θ ₂ holds. Here, the given work specification controls the position in the X direction and the force in the Y direction. In this case, the ease of movement in the X direction for realizing the position in the X direction and the force control in the Y direction are controlled. Precision can be considered as an index for control performance evaluation. Therefore, (31)
Focusing on the values of the formula and the formula (34), if the priority and importance between them are assumed to be equal, the comprehensive evaluation index can be calculated in the same manner as described in detail in the above embodiment. . Here, if the maximum values of the differential values of θ ₁ and θ ₂ are equal, and the maximum error values of τ ₁ and τ ₂ are also equal, the optimum finger angles 51a and 51b of the robot hand 5 are As angles, θ ₁ ≈ 65 °, θ ₂ ≈ -1
Get 40 °. However, it is simplified by setting l ₁ = l ₂ . Further, the angles of θ ₁ and θ ₂ are calculated at a pitch of 5 °, and the above is the optimum posture in units of 5 °. Similarly, regarding the peculiar posture, the range of 5 ° from here is regarded as a neighborhood and excluded.

As described above, since the optimum postures of the finger joints of the hand 5 corresponding to the work are obtained, at the stage when this is determined, the robot arm 1 has the remaining degree of freedom of the entire system. There are 6 degrees of freedom. Therefore, there is no redundancy as a system, and it is possible to obtain a solution in which the position and orientation of the work object 6 grasped by the hand 5 are in a desired state in the three-dimensional space. By controlling the entire robot system by using the robot hand 5, it is possible to cause the robot hand 5 for fine motion control to perform the work while keeping an optimal posture state according to the work.

FIG. 7 shows a configuration example of the control device 3 of the system in this case. The control device shown in the figure calculates based on a work specification setting unit 31 to which a prespecified work specification is input, and a work specification connected to the work specification setting unit 31 and the mechanical model of the robot hand 5. A control performance evaluation index calculation unit 32 to perform, an optimum posture state determination unit 33 that analyzes a calculation result of the control performance evaluation index calculation unit 32, and outputs a robot / hand posture vector θh most suitable for work specifications, A coordinate conversion calculation unit 37 that performs coordinate conversion calculation using the output of the optimum posture state determination unit 33 and the mechanism parameter of the robot 1 to obtain each degree-of-freedom displacement vector θ of the robot 1 and outputs the result to the robot 1. It is configured to include.

In the control device having the above configuration, at least one kind of specification input data to the work specification setting unit 31 designated in advance is calculated based on the mechanism model of the robot hand 5 (not shown). It is sent to the performance evaluation index calculation unit 32. The work specification setting section 31 is provided with a function of judging a work state by a sensor (not shown) attached to the hand 5 and selecting the work specification according to the judgment. The result obtained from the control performance evaluation index calculation unit 32 is analyzed by the optimum posture state determination unit 33, and the robot / hand posture vector θh (each of the joints 51a, 51b, 52a, 52b) most suitable for the work specification is analyzed. Displacement) is output. This output is sent to the hand 5 (not shown) for position control and the robot 1
Is given to the coordinate conversion calculation unit 37 of. The calculation unit 37 performs coordinate conversion calculation using the displacement data of the joints 51a, 51b, 52a, 52b of the hand 5 and the mechanism parameter of the robot 1 to obtain a vector θ representing the displacement of each degree of freedom of the robot 1, The result is output to the robot 1,
Position control is performed.

Next, another application example of the present invention will be described. In this application example, the robot arm 1 and the robot hand 5 shown in FIG.
52 is operated under the same conditions), the force control for the direction of the spindle 60 of the work object 6 and the direction orthogonal thereto are performed under the same work specifications as described in the above embodiment. The position control performance index is set for the robot, and the state of the robot that maximizes the overall evaluation index, which is a weighted average thereof, is determined after determining the position and orientation of the work object 6, and the robot is controlled based on this. To do. In other words, as one total system having redundant degrees of freedom, the position and orientation of the target work object 6 are fixed, and the degrees of freedom 11 to 11 of the robot 1 that maximize the above comprehensive evaluation index under the conditions. 16 and the degrees of freedom 51a, 51b (and 52a, 52b) of the hand 5 are obtained, and the robot system is controlled based on the solution. This concept can be applied to the determination of the posture state of a robot having seven degrees of freedom or more, that is, having redundant degrees of freedom, as shown in FIG. That is, after giving values of six independent variables of the position and orientation of the hand 18 of the robot 1 shown in FIG.
This corresponds to determining the degrees of freedom, that is, the displacements of the joints 11 to 17 so that the comprehensive evaluation index S has the largest value, and controlling the robot 1 based on this result.

FIG. 9 shows a structural example of the control device 3 in this case. The control device shown in the drawing retrieves teaching data of the position and orientation of the hand of the robot 1 in a predetermined sequence from a work specification setting unit 31 into which a predesignated work specification is input, and a teaching data storage unit (not shown), and obtains the target position. , An interpolation calculation unit 38 for calculating posture data, and a coordinate conversion calculation unit 39 for performing calculation based on the work specifications connected to the work specification setting unit 31 and the mechanical model of the robot 1 and the output of the interpolation calculation unit 38. A control performance evaluation index calculation unit 32 that calculates the value of the control performance evaluation index based on the output of the coordinate conversion calculation unit 39; The optimum posture state determination unit 33 that determines the most suitable displacement of the robot and outputs the result to the robot 1.

In the control device having the above-mentioned configuration, the specification input data to the work specification setting section 31 designated in advance is sent to the coordinate conversion calculating section 39 which calculates based on the mechanism model of the robot 1 (not shown). . The target position and orientation data of the hand of the robot 1 are also input to the calculation unit 39. This target data is extracted from a teaching data storage unit (not shown) in a predetermined sequence, and the interpolation calculation unit 3
8 required. The coordinate conversion calculation unit 39 calculates the displacement value of the remaining degrees of freedom for each displacement value while scanning the displacement value of one specific one of the degrees of freedom 11 to 17 of the robot 1 over the entire range thereof. The displacements of the degrees of freedom 11 to 17 of the robot 1 thus obtained are sent to the control performance evaluation index calculation unit 32, and the degrees of freedom 1 of the robot 1 are calculated.
The value of the control performance evaluation index corresponding to each combination vector θ of the displacement states 1 to 17 is calculated, and the maximum value among the index values calculated by the control performance evaluation index calculation unit 32 in the optimum posture state determination unit 33 is calculated. The posture state of the robot 1 corresponding to the one that takes a value, that is, the displacement of the degrees of freedom 11 to 17 is determined, and this is sent to the robot 1 and the control is executed.

By performing such control, it becomes possible to control the robot having redundant degrees of freedom in a form suitable for the specifications of the work to be performed.

According to each of the above-described embodiments, according to the contents of the work to be performed by the robot, the displacement of each degree of freedom of the robot, that is, the posture state, is set to the state which is the easiest to achieve the target set as the specification of the work. Can be controlled, so that it becomes easy to achieve the work goal when the work is executed. For this reason, it is possible to contribute to the shortening of the working time of the robot and the improvement of the working quality, and at the same time, it is effective in saving resources and improving the reliability of the robot in the sense that the load on the robot during working can be reduced. Further, even in the control of a robot system including a combination of a robot arm and a robot hand, or a robot having a redundant degree of freedom, the above effect does not change. Therefore, an appropriate solution for the case where the system as a whole has a redundant degree. Is also effective in determining the uniqueness of.

[0079]

According to the present invention, according to the contents of the work to be performed by the robot, the displacement of each degree of freedom of the robot, that is, the posture state, is set so that the target can be most easily achieved with respect to the target defined as the specification of the work. Therefore, the work time by the robot is shortened and the work quality is improved.

[Brief description of drawings]

FIG. 1 is a procedure diagram showing an embodiment of the present invention.

FIG. 2 is a perspective view showing a system including a robot and a work target object of the present invention.

FIG. 3 is a perspective view showing a robot system configuration in which each degree of freedom of the robot is clearly shown.

FIG. 4 is a block diagram showing a main configuration of an embodiment of the present invention.

FIG. 5 is a perspective view showing a configuration of a system including a robot arm, a robot hand, and a work target object of the present invention.

FIG. 6 is a diagram illustrating the configuration of a two-degree-of-freedom robot hand and the definition of displacement parameters and the like.

FIG. 7 is a block diagram showing a main configuration of an embodiment of the present invention.

FIG. 8 is a perspective view showing a configuration of a robot having a redundant degree of freedom which is a target of the present invention.

FIG. 9 is a block diagram showing a main configuration of an embodiment of the present invention.

[Explanation of symbols]

1 robot 2 Work object 3 control device 5 robot hand 18 Robot Minions 31 Work specification setting section 32 Control performance evaluation index calculator 33 Optimal posture state determination unit 34 Comparison Department 35 Work Object Position Detection Unit 36 Robot Motion Control Unit 37 Coordinate conversion calculation unit 38 Interpolation calculator 39 Coordinate conversion calculation unit

Claims (8)

[Claims] 1. A robot control method for controlling the position / orientation and / or force / torque of a robot arm or robot hand having at least two degrees of freedom to cause the robot arm or robot hand to perform work. The position / posture control performance evaluation index is set for each posture state of the robot with respect to the predetermined position / posture control direction corresponding to the contents of A force / torque control performance evaluation index is set for each posture state of the robot with respect to the control direction, and each of the robot arm or the robot hand is adjusted so that the weighted total index of the two control performance evaluation indexes approaches the extreme value as much as possible. A robot control method characterized by controlling the degree of freedom. 2. The position / orientation control performance evaluation index is a spatial velocity ratio of a robot tip working point to each degree of freedom velocity of the robot or its reciprocal, or a dimensionless value thereof. 1. The method for controlling the robot according to 1. 3. The force / torque control performance evaluation index is the ratio of the spatial force / torque of the robot tip working point to the force / torque of each degree of freedom of the robot, or its reciprocal, or
The robot control method according to claim 1, wherein the control values are dimensionless values. 4. The robot control method according to claim 1, wherein the weighted comprehensive index is a weighted arithmetic mean of the two control performance evaluations. 5. A robot system having a robot hand having at least two degrees of freedom at one end of a robot arm having at least two degrees of freedom is used for position / positioning.
In a method of controlling a robot for controlling work by controlling at least one of posture and force / torque, one of the robot hand and the robot arm is controlled by the method according to claim 1, and the robot arm is used. It is characterized in that the position / orientation of the robot hand or the other one of the robot arms is controlled so that the position / orientation of the work point of the robot hand viewed from a fixed coordinate system moves along a desired trajectory. Control method of the robot. 6. A position / orientation and / or force of the robot arm and / or robot hand of a robot system including a robot arm and / or a robot hand.
When performing work by controlling torque, the number of independent degrees of freedom that the robot has is greater than the number of independent positions / postures, forces, and torques that should be controlled according to the work. In this case, each degree of freedom of the robot is controlled so that the weighted total index of the control performance evaluation index according to claim 1 takes an extreme value or a maximum value, and thus each degree of freedom of the mechanism having redundant degrees of freedom. A method for controlling a robot characterized by uniquely determining. 7. A robot control method for controlling the position / orientation and / or force / torque of a robot arm or robot hand having at least two degrees of freedom to cause the robot arm or robot hand to perform work. The specifications are described using a coordinate system fixed to the work space, the operation control target in each coordinate axis direction and / or the coordinate axis direction in the coordinate system is clarified, and each coordinate axis direction and / or the coordinate axis direction is specified. The motion control capability of the robot is obtained as a function of the posture state of the robot, and the control performance evaluation index for the work specification in each direction corresponding to each of the posture states of the robot is calculated, and the weighted total evaluation index value is the largest. The robot's posture state, which is a value, is set as the optimum posture state, and the robot is made to perform work in the optimum posture state. A method for controlling a robot characterized by the following. 8. A work space in a robot controller for controlling the position / orientation and / or force / torque of a robot arm or robot hand having at least two degrees of freedom to cause the robot arm or robot hand to perform work. Work specification setting means for clarifying the operation control target in each coordinate axis direction and / or around the coordinate axis in the coordinate system based on the work specification described using the fixed coordinate system, and each coordinate axis direction And / or a control performance evaluation index for obtaining the motion control capability of the robot in the directions around the coordinate axes as a function of the posture state of the robot, and calculating a control performance evaluation index relating to work specifications in each direction corresponding to each of the posture states of the robot. A plurality of calculation means and a plurality of outputs of the control performance evaluation index calculation means for the same posture state of the robot A signal for controlling the degree of freedom of the robot such that the control performance evaluation index is weighted to calculate a weighted total evaluation index value thereof, and the weighted total evaluation index value is in the posture state of the robot having the largest value. A controller for a robot, comprising: an optimum posture state determining means for outputting