JPH0355189A - Force control method of robot - Google Patents

Abstract

PURPOSE:To determine the force receiving from an object (work) accurately by performing calibration for determining the error component of a force sensor itself and the mass and position of center of gravity of force sensor, a tool and work in the condition that they are fixed at the tip of a robot arm. CONSTITUTION:The mass and the position of center of gravity of an object (work) 30 are determined, which might be factors as required to decide the gravitation component to be corrected when a robot 10 is operating, and the error component value included in the output value of a force sensor 12 is determined by a sensor/processor 21 using a sensor environment parameter decided with the service environment, etc., taken into consideration. From the output value of sensor 20, these gravitation compensating component value and sensor error component value are subtracted by a main processor 22 when operation is executed actually. Thereby the actual force receiving from the object 30 can be determined accurately, which enables force control according to program.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is applicable to detecting force from an object, which is included in the output value of force detected by a force detector, when performing force detection with an articulated robot. This invention relates to a force control method for a robot that obtains information on forces other than the force that is present and compensates for the gravitational influence component of the tool and the error component of the force detector itself, which are generated depending on the robot's operating posture.

[Prior Art] FIG. 9 is a schematic diagram of a conventional robot control device.
In the figure, 10 is a robot, 12 is a robot arm 1
1 is a force detector attached to the tip of 1, and 14 is a tool for performing work (collectively refers to objects such as work tools, hand devices, and other objects to be held and objects to be held according to the work, (referred to as a tool) is attached to the tip of the force sensor 12. 20 is a control device that controls the force of the robot 10, which includes a sensor processor 21 that processes the force detection signal of the force sensor 12, corrects the output value of the sensor processor 21, and controls the force of the robot 10 via the servo processor 23. It is equipped with a main processor 22 that performs control. 30 is a workpiece, and 50 is a measuring instrument used for calibrating the sensor 12.

In the conventional robot configured as described above, when the force sensor 12 is attached to the arm 11 to detect force, the detection value detected by the force sensor 12 includes a force other than the force received from the target object including the tool 14 and the workpiece 30. component (that is, the force detection value due to the influence of gravity and the error component of the force sensor itself), so in order to obtain the actual value of the force received from the object, it is necessary to calculate the force value detected by the force sensor 12. It is necessary to correct the gravity component generated by the influence of the mass of the tool 14, etc. included in the output value, and the error component of the sensor itself.
Due to the error component of the sensor 12 itself attached to the robot 1 and the mass, shape, and mounting orientation of the tool 14,
When 0 is made to take various postures, the influence value due to the gravity of the object attached after the sensor, which is included in the output value, affects the measured value of the sensor 12, and the actual distance from the object is This is because the value of the force received will be different.

Therefore, in order to solve this problem, conventionally, as one of the correction methods, the mass and center of gravity of the force sensor 12 itself and the tool 14, which affect the measured value of the force sensor 12 attached to the arm 11, are calculated by adjusting From tool 14
Measure with the measuring device 50 with the robot removed, then attach it to the arm 11, and calculate the influence of gravity when the robot 10 takes an arbitrary posture based on the data obtained from the measuring device 50. However, a method was used in which the gravity influence component value obtained as a result of the calculation was subtracted during actual work to compensate for the influence of gravity on the force sensor 12 (4).

Further, as another method, for example, Japanese Patent Application Laid-Open No. 6274594
There is something like the one shown in the publication, in which a force sensor is installed between the tip of the robot and the hand device, and when the robot takes multiple postures, the output value of the force sensor and the posture information of the robot at that time are collected. This is a method of compensating for the influence of gravity by calculating the influence of gravity on the hand device on the output value output from the force sensor and subtracting it from the output value of the force sensor.

In this case, when performing the gravity compensation described above, errors may occur in the output value measured by the force sensor due to manufacturing errors of the force sensor, installation conditions when installing it on the robot, usage environment, etc., and no correction is made to the value actually detected by the force sensor. If the value is used as a force value without performing this, the accurate value of the force applied by the object cannot be obtained, so it is necessary to correct and use the error component value, which is the output error from the force sensor.

Therefore, we used a method in which the error component of the force sensor that occurs is determined in advance using a measuring device in the state of a single force sensor, and that value is used as the error component of the force sensor to correct the actually detected output value. There is. Alternatively, as shown in the embodiment of the above-mentioned publication, when calculating the influence value due to gravity on the hand device in advance, a method is used in which the value is calculated by a calculator.

[Problems to be Solved by the Invention] However, in the conventional method as described above, the influence of gravity caused by objects such as tools attached after the sensor on the force sensor, and the error component of the force sensor itself are ignored. It is difficult to compensate and accurately determine the external force received from the object, and even with some of the methods shown above, it is difficult to actually accurately determine and compensate for gravity and error components.

For example, in the example of the above publication, most of the force sensors are constructed using a strain gauge, and the output value of the force sensor includes an error component of the sensor. Moreover, the error component varies depending on the operating environment temperature and the heat generated by the sensor, etc. For this reason, when detecting force, it is necessary to compensate for the output error of the force sensor itself, taking into account the influence of the operating environment temperature and the like. In order to compensate for such force detection errors, a force detection error value is obtained in advance when compensating for the gravity exerted on a force sensor such as a hand device, and this value is then used to compensate for the gravity. The determined error component and gravity compensation value are valid only under the first conditions, and do not take into account changes in the sensor, operating environment temperature, etc. Therefore, a problem arises in that the usage conditions must always be kept constant, but this is practically impossible.Especially with a sensor, the output value of the sensor does not change for several minutes to several tens of minutes after the power is turned on. Since large fluctuations occur in the robot, the robot must wait for a long period of time, which poses a problem in terms of robot usage efficiency.

In addition, another Japanese Patent Application Laid-Open No. 62-114892 describes a method of re-correcting the error of the sensor several times during the work, but with this method (7), it is difficult to constantly correct the error. At the time of updating, it is necessary to stop at a specific position and perform calculations, and considering the dynamic characteristics of the robot, it is necessary to spend some time for this stopping, causing problems such as takt time.

Furthermore, conventional devices do not have a function that allows the user to perform these calibrations in a programmable manner, so calibration cannot be performed in places where there is not enough work area.

The present invention has been made to solve the above-mentioned problems, and is a calibration method for determining the mass, center of gravity position, and error component of objects such as force sensors and tools attached to the tip of the robot. The purpose of the present invention is to provide a force control method for a robot that can compensate for the gravity component applied to the force sensor and the force sensor error component, taking into account changes in the force sensor's output due to changes in the use environment and the temperature of the force sensor itself. do.

[Means for Solving the Problems] The robot force control method according to the present invention includes attaching a tool (8) to the tip of the arm of the robot via a force sensor, and with the tool holding a workpiece, The posture of the robot arm is changed according to the command value from the robot control device according to the specified procedure, the posture data of the robot at that time and the output value from the sensor are measured, and based on the posture data of each posture measured at that time. By performing calculations by the sensor processor, the masses and the positions of the centers of gravity of objects such as the sensor attached to the tip of the arm, tools, and workpieces are determined, which are the determinants of the gravitational component exerted on the sensor.

At the same time, the error component of the sensor itself included in the output value of the sensor is virtually determined by the sensor processor.

Next, the virtual sensor error obtained above is selected from the sensor environment parameter table stored in the internal storage section in advance, based on the environment variables input from the teaching pendant and the elapsed time since the power was turned on. The sensor processor calculates the value and the selected sensor environment parameter to obtain a sensor error component value.

Then, during actual operation of the robot, the actual external force received from the object is determined by subtracting the sensor error component obtained above and the gravity compensation component from the output value detected by the force sensor. It is.

[Function] In the present invention, the mass and center of gravity position of the object are determined, which are necessary elements when determining the gravitational component to be corrected during robot operation, and the conditions such as the usage environment are taken into consideration. By calculating the error component value included in the sensor output value using the sensor environment parameters determined by The actual force exerted by the object can be determined accurately, thus making it possible to control the force through a program.

[Example code] Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a schematic diagram of a control device used in the robot force control method of the present invention. In FIG. 1, the same reference numerals as in FIG. 9 indicate the same or equivalent parts, and 24 stores in advance a parameter matrix (see FIG. 3) used to correct sensor error components due to differences in environment. An internal storage section 25 is a teaching pendant for inputting the external environment parameters.

The robot 10 shown in FIG. 1 is equipped with a force sensor 12 at the tip of an arm 11, and a workpiece 3 is attached to the tip of the arm 11, as in FIG.
It has a hand device 14 for holding 0. The sensor processor 21 of the control device 20 has a function of converting the detection signal from the force sensor 12 into a force/torque value, performing necessary calculations, and calculating a gravity compensation component and a sensor error component, which will be described later. In addition, the main processor 22 subtracts the gravity compensation component and sensor error component obtained by the sensor processor 21 from the detected value of the force sensor 12, outputs the actual force received by the object to the sub processor 23, and outputs the actual force received by the object to the robot. The entire 10 is controlled by force.

It will be explained in further detail. The robot 10 is placed in a state where the hand device 14 and the workpiece 30 are held in front of the force sensor 1 (l1) 2, and in this state, the robot 10 observes the sensor output value applied to the force sensor 12 by the force sensor 12.
The processor 21 processes the force detection signal, and the hand device 14,
Various information necessary to calculate the gravitational component generated by the weight of the workpiece 30 and the force sensor 12 themselves, that is, the mass and center of gravity position of these objects, and the virtual sensor error factor, are obtained and stored in the internal storage device 24. The sensor processor 21 calculates the gravity compensation component and the sensor error component by calculating the sensor environment parameters.

The above operation will be explained with reference to FIGS. 2 to 4.

Figure 2 is a block diagram for determining the weight and center of gravity position necessary for gravity compensation, Figure 3 is a sensor environment parameter table for compensating for error components included in the output value due to the sensor environment, and Figure 4 The figure is a block diagram (12) showing the process of determining the force received from an object during actual operation.

In FIG. 2, the sensor signal 31 detected by the force sensor 12 when the robot 10 takes an arbitrary posture is manually input, and then the sensor signal is processed 32 and converted into force/torque values. Store 33 the data. Next, the robot 10 is made to take several postures, at least two postures (Fig. 5), and the force for each of these postures is calculated.
Torque data is similarly stored 33. On the other hand, regarding the posture of the robot, posture data 34 obtained when sensor signals are manually generated are stored 35 respectively. After the data for calculating the mass and center of gravity position is collected, the sensor processor 2
In step 1, calculations 36 are performed to calculate the mass and center of gravity position, and the mass, center of gravity position, and virtual sensor error fraction value 38 are obtained, and the results are stored 37.

As shown in FIG. 4, the force 39 received from the object during the actual force control operation is determined by the relationship between the elapsed time since the power was turned on and the external environment parameters input from the teaching pendant 25, as shown in FIG. Using the selected sensor environment parameter 40, the sensor processor 21 calculates the sensor error component value 38 with the virtual sensor error component value 38 obtained when calculating the mass and center of gravity position to obtain and update the sensor error component 41. Then, during actual force control operation, the stored data is used to calculate the gravity component 42, and the main processor 22 performs gravity compensation and sensor data compensation to perform force control.

Next, a method for calculating the above mass and center of gravity position will be described.

As shown in FIG. 5, an inertial coordinate system I and a parallel coordinate system G parallel thereto are considered, with the center of gravity of the load L including the force sensor 12, hand device 14, and workpiece 30 as the origin.

and GF　1(0,0,mg,0,0,O)　tg It can be expressed as

Let S be a coordinate system fixed to the sensor 2, and let G' be a coordinate system whose origin coincides with the center of gravity and is parallel to the sensor coordinate system S.

At that time, Image of sensor coordinate system expressed in inertial coordinate system force becomes.

next, Coordinate system G′ expressed by sensor coordinate system S rank Place P shall be.

Coordinate system S and coordinate system G' are the attitudes of becomes.

becomes.

Therefore, the mass and the position of the center of gravity can be determined by making the sensor expressed as described above take a certain attitude, detecting the respective sensor data at that time, and solving the simultaneous equations based on the values.

In addition, the force/torque values actually detected by the force sensor include
Since it includes a sensor error value, if this virtual sensor error value e is, then the detected force SFOR is as follows.

From this, the unknown quantity efx' e ry"fz" +n
By determining x'emy'eII12, a virtual sensor error value is determined.

Below, an example will be shown in which the five-point posture is changed and the data at that time is used to obtain the mass, center of gravity position, and virtual sensor error value.

The inertial coordinate system and the sensor coordinate system are shown as shown in FIG. 6(a).

When this is rotated by θ around the X-axis in the sensor coordinate system, it becomes (Fig. 6 (b)). Furthermore, when it is rotated by 1/2π around the Z-axis of the tool system, it becomes (19) (Fig. 6 ( c) ) When the tool system is rotated by 1/2π around the Y axis, it becomes (20) (Figure 6 (d)) When it is rotated by 1/2π around the Z axis (Figure 6 (e)) First, find mg, e, , e and fx
When fy fz, (23) is obtained, which can be obtained by solving (24).

do.

Next, i , t , i , e , e , e
x y z mx
When my mz is determined, (27) is obtained by setting (27) and solving (28) using the pseudo inverse matrix as described above.

As described above, it is possible to calculate the mass and center of gravity of the tool and the sensor itself, which are elements of the gravitational force acting on the sensor. The flowchart is shown in FIG.

FIG. 8 shows an example of a program for a robot language, which is an example of a method for implementing programmable calibration.

Next, a method for determining the force received from an object in actual operation will be described.

Based on the elapsed time since the sensor power was turned on and the value entered manually from the teaching pendant, an environmental parameter is selected from the sensor environment parameters stored in the internal storage (Fig. 3), and the virtual sensor error value obtained above is selected. The sensor error component is determined from the relationship.

At this time, the virtual sensor error value is E = (e , e , e , e , e , e
)tfx fy fz mx
my mz sensor error value Er = (e', e' f, e' fz
'e'IIIX'fxe' , e' )t my mz When the environmental parameter is α, the sensor error value is expressed as E' = EXα.

From the above, when actually performing work, the actually detected force/torque values are expressed as F= (rx, fy, fz, mx, m
y, mz) The force ●torque value that the t sensor receives due to the gravitational component of the hand device, etc. is F' - (fx', ry', fz', m
x', +ny'IIlz')t At this time, F' can be calculated from the mass, center of gravity position, and sensor attitude information obtained above, and the force actually received from the object and! - Luk F' can be found by F' = F - F' - E'.

[Effects of the Invention] As described above, according to the present invention, objects such as a force detector, a tool, and a workpiece are attached to the tip of the robot arm to measure the force and torque values actually received from the object. Calibration can be performed to determine the mass, center of gravity position, and error components of the force detector itself, and the gravitational component applied to the force detector and the Since the detector error component can be compensated for, the force exerted by the object can be accurately determined. Another advantage is that a special measuring device for determining the mass, center of gravity position, and detector error component is not required.

Furthermore, since the calibration itself can be set programmably, it is possible to obtain variability that corresponds to the working environment and the work to be installed. Further, effects such as calibration being possible in places where there is not enough work area can be obtained.

It should be noted that the present invention is applicable not only to robot control but also to machine tools and special purpose machines related to force control.

[Brief explanation of drawings]

Figure 1 is a schematic configuration diagram of a control device that implements the robot force control method of the present invention, Figure 2 is a block diagram for determining the mass and center of gravity position necessary for gravity compensation, and Figure 3 is a sensor Figure 4 is a block diagram showing the process of determining the force received from an object in actual operation, and Figure 5 shows the calibration. Figure 6 is an explanatory diagram of the method of implementing the calibration method, Figure 7 is a flowchart for performing calibration, and Figure 8 is a programmable implementation of calibration. 9 is a schematic diagram of a conventional robot control device. 10... Robot 11... Robot arm 12... Kasensor 14... Tool 20... Robot controller 21... Sensor processor 22... Main processor 23... Servo processor 24...Internal storage unit 30...Work 40...Sensor environment parameters In the drawings, the same reference numerals indicate the same or corresponding parts.

Claims (1)

[Scope of Claims] A method for controlling the force of the robot by attaching a tool to the tip of the arm of the robot via a force detector, and detecting the force received from an object including the tool and the work with the force detector, With the object attached, the robot is given a movement instruction and operated so that the force detector takes at least two predetermined postures, and the posture of the force detector is changed. From the relationship between the output value of the force detector and the posture information of the force detector at that time, use a calculator to determine the mass and center of gravity position of the object, and the force detector included in the output value. The own error component is virtually determined, and then the virtual detector error component value obtained above is corrected using the sensor environment parameter selected from the sensor environment parameter table stored in advance to obtain the detector error component. The sensor error component value and the gravity compensation component value obtained from the mass and center of gravity position of the object are subtracted from the output value of the force detector during actual operation of the robot, and A force control method for a robot characterized by determining the actual external force being received.