JPS63257809A - Master/slave controller - Google Patents

Abstract

PURPOSE:To improve operatability by driving a slave motor by a signal for controlling a master/slave positional deviation signal to zero and correctly measuring torque independently of the angles of master and slave joints. CONSTITUTION:When a master 1 is driven, a signal Pm from a master potentiometer Ps is changed. A positional deviation subtractor 28 calculates a positional deviation signal Pfb between the master 1 and the slave 2 by subtracting the slave potentiometer signal Ps from the master potentiometer signal Pm. A slave speed controller 23 outputs a speed command Sps to a slave motor controller 24 so that the deviation signal Pfb is turned to zero and a controller 24 drives the slave 2 based on the command. Consequently, the slave 2 follows the master 1 and executes its operation in accordance with the operator's operation. Even if the joints of the master 1 and the slave 2 are set to any angle, the torque value can be correctly measured.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) The present invention provides a slave-side manipulator (hereinafter referred to as slave) that performs actual work and a master-side manipulator (hereinafter referred to as master) that operates the slave. ) The present invention relates to a master-slave control device for a master-slave manipulator having a device for measuring joint torque using a torque sensor that cannot measure absolute torque, such as a strain gauge.

(Prior Art) When performing complex and detailed work by remote control, it is effective to control using a master-slave system.

Master-slave manipulator system.

There are two broad categories: masks operated by operators.

It consists of a slave remotely operated by a master, and a master-slave controller that controls the master and slaves.

The mask and slave each have several joints.

It is a manipulator with a similar shape.

The control device has the following functions to remotely control the slave with a mask.

1. When the master is pressed by the operator, the mask operates according to the pressing force.

2. If there is a force received by the slave, operate the master according to the reaction force of the force received by the slave.

3 When the master operates, the slave follows the master's operation.

That is, the control device moves the master according to the force with which the operator manipulates the mask, and causes the slave to operate according to the movement of the mask. In this way, the operator can remotely control the slave.

However, since the slave and master are separated, it is difficult for the operator to monitor the situation at the operation site.

For this purpose, by providing a function that transmits the reaction force from the slave to the mask, the force of the slave holding the object can be transmitted to the master, allowing you to perform operations with a sense of realism as if you were actually doing the work. This makes it easy to operate.

Now, in order to move the master as requested by the operator, and in order to transmit the reaction force from the slave to the mask, it is necessary to consider the torque due to the weight of the arm itself.

For example, in the case of a master, since the joint has the weight of the arm, it is necessary to measure the correct operating force with which the operator operated the master, taking into account the torque due to the weight.

This state will be explained using FIG. 4. Regarding the torque Tg+++ applied to the joint due to the arm's own weight, consider the simplest 1-axis case.

Length Tgm=Jushin G・5IN(θ)...
(2) However, Tg+a: Torque applied to the joint due to the weight of the arm G: Weight of the arm Φ Distance from the two joints to the center of gravity θ: Can be determined from the angle of the joint.

As a result, the torque T applied to the entire joint becomes the torque T11 applied to the joint due to the operator's operation.

This is in addition to the torque Tm due to the arm's own weight.

In other words.

T = T −+ T gem ・
...■ However, Tm: Torque applied to the joint due to operator's operation T: Total torque applied to the joint Tm Ugly: Torque applied to the joint due to the weight of the arm.

Thereby, it can be determined by subtracting the torque Tgm due to the arm's own weight from the torque Twit applied to the joint due to the operator's operation and the torque T applied to the entire joint.

For this purpose, it is assumed that the total torque applied to the joints is correctly measured, so the sensor that performs this measurement will be explained.

For example, a strain sensor that measures the torque applied to the joint of the master will be schematically explained using FIG. 3. First, in order to determine the torque applied to the joint (g+lI), the force applied to the shaft of the motor is measured. The shaft of the motor rotates the motor gear 42 and the arm is operated by the joint shaft 44, so a moment is applied to the gear 43 due to the inertia of the arm, so that the gear 42 receives a reaction force in the tangential direction. This reaction force is applied to the support base 4 that supports the bearing.
The strain sensor 4 attached to 1 is distorted.

In the direction of rotation shown in Figure 3, the motor shaft is in gear 4.
2 is pushed in the tangential direction of the gear 42 by the force A, and the strain sensor 4 is compressed and distorted.Also, even when a force is received from the arm in the direction to move the gear 42, the same happens. distort.

In this way, the force applied to the joint is measured in the form of strain.

However, strain sensors leave residual strain when installed. Also, depending on the daily temperature, etc., the residual strain sensor distortion gradually changes. Therefore, it cannot be said that the distortion value absolutely indicates the current torque value6.Therefore, the amplifier is called zero balance. It has an initial setting function. This initial setting function sets the electrical signal output to the outside to zero at the time when the operator requests the initial setting function.

After making this setting, the relative change in strain amount from the initial setting torque can be measured as the torque applied to the joint.

Now, as you can see in 20, even when there is no operation by the operator, the torque applied to the joints is due to the weight of the arm. In other words, when initializing the sensor, the electrical signal from the amplifier becomes zero, but due to the torque due to its own weight, the actual torque applied to the joint does not become zero, and the value from the sensor and the actual torque It will be different.

Therefore, in order to match the torque value applied to the actual joint with the torque value from the sensor, conventionally the slave and master were manually operated to an angle where no torque was applied due to their own weight, and the initial settings were made at that angle. I was going.

Now, let's create a control device for operating slave 2 by master 1 using the torque value thus measured.

This will be explained with reference to FIG.

The master 1, which is remotely operated by the operator, includes a master motor 3 that drives the joints, a master strain sensor 4 that measures the torque applied to the joints, and a master potentiometer 5 that measures the angle of the joints.

First, in the case of master 1, the speed signal S pm' is inputted to the master motor controller 14 by the master manual operation device 19 until the angle at which no torque is applied due to its own weight, and in the case of slave 2, the speed signal S pm' is inputted to the master motor controller 14 by the slave manual operation device 29. By inputting a speed signal S ps' to the motor controller 24, the respective motors are operated.

Next, the initialization switch 30 initializes the master amplifier 10 and slave amplifier 20. After initial settings are made, the torque applied to the joint is input from the master strain sensor 4 to the master amplifier 10 and converted into an electrical signal Tan.

By subtracting the torque Tgm due to the master's own weight from this torque, the torque Ta+ due to the operator's operation can be calculated, as shown in equation (2).6 Next, if there is no reaction force from slave 2, The mask speed control device 13 operates the master according to the torque Tm generated by the operator's operating force.

A speed command Sp+* is outputted so that the master 1 operates at a speed proportional to the operating torque Tm of the operator.

The master motor controller 14 is the master motor 3
The speed is controlled according to the speed command Spa.

In this way, when there is no reaction force from the slave 2, the master 1 operates according to the force operated by the operator.

Next, a case where the slave 2 hits an obstacle or the like and receives a reaction force will be described.

In this case, a reaction force is applied to the slave 2, the slave strain sensor 7 is distorted, and the amount of distortion Ths becomes an electric signal Tas by the slave amplifier IO.

Next, the torque Ts due to the reaction force actually received by the slave from the obstacle can be determined by subtracting the torque Tgs due to the weight of the slave 2 from the torque Tas applied to the slave joint.

In order to transmit the torque Ts due to this reaction force to the master, the reaction force feedback device 27 converts it into a reaction force feedback signal MPf.

Furthermore, this reaction force feedback signal Mpf is added to the master operating force T11 by the master adder 17,
Master speed control device 13 as master operation command Mt
is input. In this way, the operator operating the master can feel this reaction force.

As described above, the master 1 operates according to the operator's operation command, and at the same time, the operator can feel the force received by the slave 2.

Next, when the master 1 operates, the signal Pm of the joint angle of the master potentiometer 5 changes.
The master-slave position deviation signal Pfb is calculated by subtracting the signal Ps of the potentiometer 8 of the slave 2 from .

Slave speed control bag so that this deviation signal Pfb becomes zero! ! 23 outputs the speed command Sps to the slave motor controller 24. The slave motor controller 24 operates the slave 2 according to the command Sps.

In this way, the slave follows the master 1, and the slave 2 operates according to the operator's operation.

As a result, the slave 2 was remotely controlled by the master 1.

(Problem to be Solved by the Invention) Since the strain sensor has residual strain, it is not possible to measure the absolute value of strain.

For this purpose, it was necessary to initialize using the zero balance function as described above.

However, this initialization process reduces the torque value to zero at this point, so unless the torque applied to the actual joint is also reduced to zero, there will be a difference between the torque measured by the sensor and the torque value actually applied to the joint. . To avoid this, it was necessary to move the master and slave arms to an angle where no torque was applied to the joints, and then perform the initialization work in that state.

That's all, when starting work with master slave.

Unless the joint is moved to a point where the torque due to its own weight is not applied, there will be a difference between the torque applied to the joint and the torque calculated by the sensor, making it impossible to accurately calculate the torque applied to the joint.

In master-slave control, this is the most important sensor for control, and if control is performed with an inaccurate torque value, there is a risk of runaway. Therefore, it is not possible to work as a master slave in this state.

Therefore, when attempting to operate with a conventional manipulator, it is necessary to move the master and slave to an angle where no torque is applied due to their own weight, making it impossible to start freely when starting to use master-slave control, and requiring complicated startup operations. It was necessary.

Further, it is necessary to operate the slave until its own weight torque is zero, but since the slave is far away, it is difficult to understand the surrounding environment and there is a risk of colliding with obstacles. For this reason, it is necessary to temporarily evacuate the equipment to a safe location before operating it, and this operation reduces the operating rate.

Furthermore, if the operator accidentally initializes the amplifier, there is a problem in that the force of its own weight remains and the master may operate on its own, which is dangerous.

Therefore, the present invention provides an initialization device so that the torque value can be accurately measured at any angle of the master and slave joints, and eliminates complicated preparatory operations for starting master-slave control. The present invention aims to provide a master-slave control device with improved operability.

[Structure of the invention]

(Means for solving the problem) As shown in Figure 1, in a master-slave control device that performs master-slave control, the applied joint torque is memorized by the arm's own weight when the sensor is initialized. A device is provided to add the value to the sensor value.

In other words, when the amplifier is initialized, the current torque is calculated separately from the sensor, and that value is stored as an offset so that it always shows the current torque value accurately regardless of the posture.

Calculate the correct torque value in addition to the sensor value.

This torque applied at initialization can be considered to be the torque due to the weight of the arm, so the torque due to one weight can be calculated from the angle of the joint.

In the following manner, even after initialization, the subsequent torque value can be accurately determined.

A device is provided that uses a signal to initialize the amplifier and simultaneously calculates the joint torque due to the arm's own weight from the angle of the joint.

A device is provided that stores the value and outputs the value as an offset. A device is provided that adds the torque value detected by the sensor and the stored torque value.

By using this device, you can always calculate the correct torque value and start master-slave operation.

(for production) Explain about master.

First, if you want to start master-slave control from a position where the arm's own weight is applied, use a switch to request the zero balance function. This request clears the strain sensor's electrical signal to zero.

At this time, the arm's own weight value is also memorized from the angle of the joint.

To calculate the actual torque applied to the joint,
The memorized torque based on the arm's own weight is converted into an electrical signal in accordance with the output level of the master amplifier that converts the signal from the strain sensor.

An electric signal obtained by adding the electric signal based on the self-weight value and the electric signal indicating the torque value from the strain sensor can be used as an electric signal indicating the overall torque value applied to the joint.

For example, if we consider a case where the arm moves to a position where its own weight value becomes zero, the electric signal from the sensor and the electric signal from the own weight value stored at the time of initialization are electrical signals with different signs and the same absolute value. becomes. As a result, the electric signal representing the total torque value of the joint becomes zero, indicating the actual torque.

This eliminates the need to operate the master and slave for initialization.

Therefore, the master and slave can freely initialize the sensor no matter what angle they are at.
You can freely start remote control using master-slave operation.

(Example) FIG. 1 shows the configuration of the present invention.

Master 1 and slave 2 are the same as before.

The master amplifier 10 is a device that amplifies the signal Thm from the master strain sensor 4 and has a zero balance function.

The master initial torque setting device 11 is the master amplifier 10
Torque due to self-weight when zero-balanced TgI11
is stored and subsequently converted into an electrical signal Tea indicating the initial indirect torque offset.

The master dead weight calculation device 12 is a device that calculates the dead weight value Tgm from the angle of the master.

The master torque adder 15 uses the stored torque value T6.
11 and the torque value Tan from the sensor amplifier 10 to obtain the torque value Trm applied to the joint.

The master torque subtractor 16 is a device that calculates the force Tm operated by the operator by subtracting the electric signal Tg, which indicates the torque due to the body's own weight, from the electric signal Tr, which indicates the torque value applied to the joint.

On the other hand, when explaining the slave 2, the slave amplifier 20 is as follows.
This is a device that amplifies the signal Ths of the slave strain sensor 7 and has a zero balance function.

The slave initial torque setter 21 is connected to the slave amplifier 20.
The torque Tgs due to its own weight when zero-balancing is stored, and an electric signal Tgs indicating the initial joint torque offset is stored.
Convert to as and output.

The slave dead weight calculation device 22 is a device that calculates the dead weight value Tgs from the slave angle Ps.

The master torque adder 25 uses the stored torque value To.
s and the torque value from the sensor is added to Tas to obtain the torque value Trs applied to the joint.

The slave subtracter 26 is a device that calculates the reaction force Ts received by the slave 2 by subtracting the electric signal Tgs indicating the torque due to its own weight from the electric signal Trs indicating the torque causing the joint to stiffen.

The reaction force feedback device 27 is a device that calculates the amount of reaction force Ts of the slave 2 to be transmitted to the master 1.

The adder 17 calculates the force Tm operated by the operator and the slave 2.
A value Mt for operating the master 1 is calculated by adding the reaction feedback force Mpf.

The master speed control device 13 gives a speed command Sp- according to a value Mt for operating the master 1. The master motor controller 14- controls the speed of the master motor 3 according to the speed command Spw+.

The position error subtractor 28 receives the potentiometer signal P1 of the master 1.
The angular deviation Pfb is obtained by subtracting the potentiometer signal Ps of slave 2 from 1. The slave speed control device @23 sets the speed fd Sps so as to make the angular deviation Pfb zero.

The slave motor controller 24 has a speed signal Sps
The speed of the slave motor 6 is controlled as follows.

The operation of the master-slave control device described above will be explained below.

Before operating the master slave 2, the operator presses the switch 30 to turn on the switch signal Stt in order to initialize the torque of the master 1.

When this switch No. 48 SV is turned ON, the master amplifier 10 performs zero balance.

It is activated when the switch signal Sw is turned ON, and the initial torque setting device 11 sets the value T from the master self-weight calculation device 12.
gm is stored (101).

Convert this memorized value to the necessary electric signal ROM (1
02), this converted electric signal Tea+ is converted to a value matching the output level of the sensor signal TaIm outputted by the master amplifier 10.

This converted value is output as an electrical signal (10
3). This output continues even after activation and completion of the operation.

This stored signal Tag and the electric signal Ta from the sensor
tm is added by the master adder 15 to become an electric signal Trm indicating the total torque value applied to the joint.

The operator operation signal Tug is obtained by subtracting the master 1's own weight value signal Tg+s obtained from the joint angle Pa from the joint torque signal Trm by the master subtractor 16.

The signal Mt for operating the master 1 is obtained by adding the operator operation signal Tm and the reaction force signal MPf from the slave 2 to the adder 1.
It can be found by adding 7. In other words,
With this calculation, the reaction force Mpf from slave 2 can be transmitted to master 1.

Next, the signal Mt for operating this master 1 is input to the master speed control device 13, where the speed signal Spa
is calculated, and the master motor controller 14 controls the master motor 3 according to this command Spa.

Next, how to obtain the reaction force MPf from the slave 2 will be explained.

As with the master 1, the operator presses the switch sv to turn it on in order to initialize the torque of the slave 2 before operating the master slave.

When this switch signal Sv changes to ON, the slave amplifier 20 performs zero balance.

Furthermore, when this switch signal Sv changes to ON, the slave initial torque setter 21 sets the dead weight value Tg at that moment.
Remember s. Furthermore, this stored self-weight value Tgs is converted into an electric signal Tas matching the output of the slave amplifier 20.

This signal Tea and the electric signal Tag from the sensor are added by the slave adder 25 to form an electric signal Trs indicating the total torque value applied to the joint.

The reaction torque signal Ts of the slave 2 is the slave subtracter 2
6, the self-weight value signal Tgs of the slave 2 obtained from the joint angle Ps is subtracted from the joint torque signal Trs.

The reaction force torque signal Ts of the slave 2 is converted into an appropriate value by the feedback control device 27, and becomes the slave reaction force signal Mpf.

Furthermore, when the master 1 operates, the 0 position deviation subtracter 28, which changes the signal Pm of the master potentiometer, subtracts the signal of the slave potentiometer Ps from the signal of the master potentiometer Pa. The position deviation signal Pfb of 2 is calculated. The slave speed control device outputs the speed command Sps to the slave motor controller 24- so that the deviation signal Pfb becomes zero. Slave motor controller 24-
operates slave 2 according to the command.

In this way, the slave 2 follows the master 1, and the slave 2 operates according to the operation of the operator.

Therefore, according to the embodiments described above, the sensor can be initialized and the correct torque value can be determined regardless of the posture of the arm.

Conventionally, in order to obtain the correct torque value, it was necessary to move the arm to a position where no torque was applied.
With the above, master-slave operation can be started no matter what posture the arm is in.

This makes it easier to start the operation, improving operability. Furthermore, operations can be started safely even near obstacles, reducing the burden on the operator.

In the above explanation, the sensor initialization operation was started by pressing the switch 30, but if the Sv signal is turned on when the control power is turned on, the device can be made with fewer operational errors. be able to.

〔Effect of the invention〕

There is a sensor to determine the torque applied to the joint.

Even if it is an absolute sensor, the correct torque can be determined without the need for special operations.

This allows the operator to operate the arm without requiring complicated pre-operation to move the arm to the required position, improving work efficiency.

Furthermore, even if the operator mistakenly requests sensor initialization, an incorrect torque value will not be displayed, allowing safe operation.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of the master-slave control device of the present invention, Fig. 2 is a block diagram of a conventional master-slave control device, Fig. 3 is a diagram of the relationship between the motor and strain sensor, and Fig. 4 is a diagram of the torque applied to the arm. It is an explanatory diagram. 1...Master 2...Slave 3.6...Motor 4,7...
Strain sensor 5.8... Potentiometer
10.20... Amplifier 11.21... Initial torque setting device 12.22... Own weight calculation device 13.23
... Motor controller 14.24 ... Speed control device 15.25 ... Master torque adder 1
6.26...Master subtractor 17...Master adder 27...Cafe feedback device [28...Position deviation subtractor 29...Slave manual operation device 30...
Switch agent Patent attorney Nori Chika Ken Yudo Daishimaru Ken

Claims (1)

[Claims] In a master-slave control device for a manipulator system that includes a slave-side manipulator and a master-side manipulator that remotely controls the slave-side manipulator, when initializing joint torque from a torque sensor, A master-slave control device characterized in that a signal from an initial torque setting device in which a self-weight torque value of a manipulator is stored is added to a signal from a manipulator strain sensor corresponding to a torque applied to a joint to obtain a joint torque.