JPH079608B2 - Robot control method and apparatus - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a robot control method and apparatus, and particularly to a control method and apparatus suitable for causing a robot operating in multiple degrees of freedom to perform various tasks. It is about.

[Background of the Invention]

In recent years, robots have been widely introduced in various fields,
One of the major characteristics of this is that the robot is versatile and can be applied to a wide variety of tasks, unlike general dedicated machines.
Because it is one.

When the robot is applied to a certain work, it is rarely used as it is, and it is common to equip the robot with a tool corresponding to an object to be handled.

For example, when introducing a robot into a processing or assembly line at a work site to perform work, a tool according to the shape and size of the object or the content of the work is provided between the hand part at the tip of the robot or the arm part and the hand part. It is installed, and assembly work and painting work are performed. In this case, the tool itself is made to have a certain degree of flexibility so as to absorb and eliminate factors such as variations in accuracy of the target object and work errors such as positioning errors.

However, as the contents of work become complicated and sophisticated, the tools themselves are required to have a high degree of freedom and high functions, and therefore their structure tends to be complicated and expensive. Further, it is not uncommon for one tool itself to deal with it, and even in the same work, it is often necessary to replace the tool depending on the shape of the object and the working conditions.

As a result, not only is it time-consuming to change and adjust the tool, which reduces the operating rate of the robot, but also causes many problems such as maintenance and management of the tool in order to maintain the reliability of the work.

An example of a flexible tool is the Remote Center Compliance (RCC) mechanism disclosed in USP 4,098,001. With this RCC mechanism, if a force or moment is applied to the tip of the grasped object, it will move in parallel or rotate to handle it, and simply move the robot hand in a straight line. It can be combined. However, even this RCC mechanism is not sufficiently flexible with respect to changes in the length and fitting condition of the shaft-shaped part which is the object.

As described above, the robot is different from the conventional dedicated machine in that it can handle only a single part and work, but it handles a wide variety of parts, and different work can be done with the same machine. , Has a great significance. For this reason, if it is necessary to replace the tool every time the type and work of the robot change, it is difficult to show off the characteristics of the robot in terms of versatility or flexibility.

However, the current situation is that there is no choice but to carry out such a tool exchange or something equivalent to this with the current robot technology, and in order to deal with this, it is possible to make the tool versatile, A tool having general versatility often has a complicated mechanism and often includes an unnecessary part, which may be more wasteful than a case of exchanging tools.

In this way, on the contrary, if the problem is solved only by the mechanism, the problem of flexibility has a property that it becomes more difficult.

[Object of the Invention]

It is an object of the present invention to provide a robot control method and its apparatus that can cope with various work conditions without changing mechanically such as tool exchange and the like by simply changing control characteristic parameters.

[Outline of Invention]

The present invention is a robot control method for controlling a robot at each sampling time, in which a force signal is calculated by detecting an actual force applied to a robot hand at a previous sampling time (n-1). The first value, which is the difference between the signal and the force reference value that was the target value at the previous sampling time (n-1), is calculated, and the actual position and posture of the hand at the previous sampling time (n-1) are calculated. A second value is calculated by multiplying the spring constant by the difference between the position signal represented and the position reference value which was the target value of the position and orientation at the pre-sampling time (n-1) to calculate the pre-sampling time (n The value obtained by differentiating the position detection value from the position detector provided in the robot is used as the velocity signal in -1), and the third value is calculated by multiplying this value by the viscosity constant, and the first, second, and With a value of 3 The command speed of the next sampling time (n) of the hand is calculated from the mass coefficient of the hand.

Further, a position detection value from a position detector provided in the robot is used as the position signal.

Further, position information from a non-contact sensor provided in the robot is used as the position detection signal.

Further, the robot control device calculates a force signal based on a force applied to the robot hand, and a position of the hand based on information from a position detector provided in the robot. It is configured by second calculating means for calculating a position signal representing a posture, and fourth calculating means for calculating a joint angular velocity of the robot from a command velocity of the hand based on the force signal and the position signal. is there.

Example of Invention

 Hereinafter, the present invention will be specifically described with reference to the drawings.

FIG. 1 is a configuration diagram showing an outline of the overall configuration of an articulated robot and a control system. Here, the articulated robot 10 shown as an example of the robot to be controlled is provided with a swivel base 11 capable of swiveling on a base around a vertical axis, and provided on the swivel base 11 so as to be rotatable about a horizontal axis. Upper arm 12 and this upper arm 12
There are a total of 6 degrees of freedom including a forearm 13 rotatably provided around another horizontal axis located at the tip of the forearm, and a wrist 14 that operates in 3 degrees of freedom at the tip of the forearm 13 as shown in the figure. It is an articulated robot. The wrist portion 14 has a hand 15 for gripping an object to be handled and a force sensor 16 for detecting a force applied to the robot.

An encoder PE for detecting each rotation angle and a tacho generator TG for detecting each rotation speed are connected to a motor M for driving each of the swivel base 11, the upper arm 12, the forearm 13, and the wrist portion 14. ing.

The control device 17 takes in the position signal of the robot and the force signal v applied to the robot, which are respectively obtained from the encoder PE and the force sensor 16, and determines the robot's position from the signal reference value r and the force reference value c input in advance. Is calculated and output to the servo amplifier 18 to operate the articulated robot 10.

FIG. 2 is an explanatory diagram showing the components of the control device and the drive unit of the robot and their relationships. The controller 17 includes a microprocessor 19 for performing various calculations and controls, and an RA
M20, ROM21, operation unit 22 for performing routine function operation,
Interface circuit (I / F) 24 connected to console 23
And a bus line 25 for connecting each of these components. Further, the bus line 25 is output to the motor M in the robot drive unit 26 via the servo amplifier 18, so that the angular velocity digital signal is converted into an analog signal. A D / A conversion circuit 27 to be exchanged, a counter 28 for counting the output from the encoder PE connected to the motor M, and for converting an analog signal related to the force output from the force sensor 16 via the amplifier 29 into a digital signal. Is connected to the A / D conversion circuit 30.

Here, the position reference value r and the force reference value c may be 0, but in this case, the articulated robot 10 operates only by the force v given to the hand 15. That is, it becomes a passive mechanism.

On the other hand, when only the position reference value r and the force reference value c are given, the movement of the articulated robot 10 is determined by these reference values.

With the above configuration, control device 17 performs the following processing operation.

FIG. 3 is a schematic block diagram showing the processing operation of the controller 17.

The controller 17 receives the position signal of the robot by counting the signal from the encoder PE of each joint of the robot in the position / orientation calculation unit 40 by the counter 28, and converts the angle of each joint into the orthogonal coordinate of the hand system. Then, the position of the hand
Calculate the posture. Actually, the operation unit 22 and the like perform the command from the microprocessor 19 in FIG. In addition, the force sensor 16 detects the external force a applied to the robot,
Control device as the voltage v of the strain gauge that composes 16
17 is taken in via an amplifier 29 and an A / D conversion circuit 30.
The force calculation unit 50 takes in v, and first converts the strain gauge voltage v into a force by a strain / force conversion matrix or the like. Actually, as in the case described above, the arithmetic unit 22 and the like perform the command from the microprocessor 19 in FIG.

Next, when the weight of the hand 15 and the object gripped by the hand 15 is large, the value given to the force sensor 16 differs depending on the posture of the hand. Therefore, the posture of the hand 15 is calculated more, the influence due to this is removed, and the force a actually applied to the robot is estimated. The value thus estimated is and is given to the control main part 60. The main control unit 60 calculates the speed at which the hand 15 should operate from the calculated force and the position of the robot and the force and position reference values c and r input in advance. Since this speed is a value in the coordinate system fixed to the hand, it is converted into a speed in the base coordinate system by coordinate conversion and is output.

After this, the joint angular velocity calculation unit 70 calculates from the velocity in the Cartesian coordinate system.
Each axis of the robot should take Is output. However, (Here J ^-1 is called the inverse Jacobian matrix, whose elements are functions of the robot's position signal.) Is converted into an analog signal by the D / A conversion circuit 27 and output to the motor M for each joint of the robot via the servo amplifier 18 to operate the robot.

The operation of the control device 17 has been described above. Of these, the processing method of the control main part 60 mainly constitutes the feature of the present invention, and the principle thereof will be described below.

In general, the movement of a flexible tool is related to the inertial force, the restoring force of a spring, the viscous force of a damper, and the applied force. Here, one degree of freedom will be described to simplify the description.

When the model of the tool having one degree of freedom is expressed by a mathematical expression, the following expression (1) is obtained.

Here, m is the mass of the hand, K is the spring coefficient of the spring, C is the viscosity coefficient of the damper, and xr and Fc are the reference values of the position and force, respectively.

Equation (1) shows that the position and velocity of the tool when the force (F-Fc) is applied to the tool are x and v, respectively. In other words, if the robot hand (F-
When the force Fc) is applied, the robot's hand is x and v
It shows that the robot hand can perform the same movement as when the tool is attached, though it has a position and a velocity as follows, although the robot hand is not actually attached.

In other words, by controlling the robot assuming the tool you want to attach to the robot, you can make the robot move the hand that you would do when you incorporated the tool into the robot. This is one of the major features of the invention.

When actually operating the motor M that drives each joint of the robot, it is useful to calculate the speed v by the calculation means because the speed or a command value corresponding to the speed is given to the servo amplifier. The equation is transformed into the following equation (2).

Furthermore, when the above equation (2) is represented by a discrete value system, the following equations (3) and (4) are obtained.

Here, Δt is the sampling time, and the suffix i or n is the i or n-th value.

As is clear from the above formula, according to the processing method of the present invention, the differential calculation is not included in the calculation of speed, it is resistant to noise, malfunction does not occur, and the robot hand can move stably. .

Further, v is added to the right side of the above equation (4) transformed into a discrete value system.
_{There is n-1} (v before n), and v _n is on the left side.
That is, in order to calculate the speed v of the hand of the n-th robot by the calculation means, only the previous speed v is required, and it is not necessary to detect and feed back the actual speed v from the tachogenerator TJ etc. every time. This is to detect the speed of the robot's hand using a speed detector such as a tachogenerator TG actually attached to the drive system of the robot, to detect the position of the encoder PE etc. also attached to the drive system of the robot. This is a useful feature considering that the reliability of the signal is a little lower than the detection of the position of the robot hand detected by the instrument.

In the equation (4), x is related to the position x.
_As can be understood from the fact that there is only one _n-1 , the x position needs to be detected every time, but this position x is calculated from the position signal (joint angle) of the robot from the encoder PE.

Here, m, k, and C are values determined when assuming the characteristics of the tool. Assuming a tool, analyze the mechanism of the tool and determine these values off-line before operating the robot, and use these values as data specific to the tool (even if it is a constant May be input from the console 23 and stored in the RAM 20 of the control device 17. In other words, if there are multiple assumed tools, there will be multiple data to be stocked accordingly.

Further, the condition for stabilizing the system represented by the above formula (1) is known in the basic theory regarding ordinary differential equations. It is possible to stabilize or diverge by setting m, c, and k based on it.

Similarly, in order to stabilize Equation (2) above by Equation (4) converted to a discrete value system, in addition to the above conditions, Δt
The value of must also be sufficiently small.

However, since Δt is the sampling time, it takes a positive value, and since m is a value corresponding to the mass, it takes a positive value.
If used for the purpose of stabilization, C has a positive value.

In the above, according to the present invention, the method of making the operation characteristic of the tool to the robot has been described.
It relates to a method based on the values of C, k.

In addition to the above method, the present invention is characterized in that the special characteristic of the tool can be determined based on the force F applied to the robot, which is another variable, the reference value Fc of this force, and the reference value xr of the tool position. is there.

That is, when the three values of F, Fc, and xr are used as command values,
Many combinations are possible and the tools can have different operating characteristics. Here, the following operation example will be described as a specific example.

Fig. 4 shows the first operation example, and the condition is 4mk-C.
² > 0, C> 0, m> 0, xr = 0, Fc = 0, and when F takes the step input Fs27 as shown in FIG. 4, position x is time t
In contrast, the relationship of the following expression (6) is established.

However, it is assumed that x _{t = 0} = 0 and v _{t = 0} = 0.

The operation of the robot in this case is output 28 shown in FIG. The position converges to a value proportional to the size of the Fs step input 27, and the rising time and vibration at that time are determined by m, C, and k.

FIG. 5 shows the second operation example, and the condition is 4 mk-C.
² > 0, C> 0, m> 0, F = 0, Fc = 0, and when xr is given a step input x _rs 29 as shown in FIG. Equation (7) is obtained.

However, it is assumed that xt = ₀ and vt = ₀ .

The operation of the robot in this case is the output 30 shown in FIG.

As can be seen from the output 28 shown in FIG. 4 and the output 30 shown in FIG. 5, and as can be estimated from these outputs and the equation (1), even if a step input is given to F under certain conditions. , Xr gives the same output.

That is, change the value to F to get the same output,
It can be set to xr or selected, and various types of control are possible as required.

FIG. 6 shows a third operation example, the condition of which is k =
0, m> 0, C> 0, xr = 0, Fc = 0, and F as shown in FIG.
When the step input Fs31 is applied to the speed v, the speed v has the relationship of the following expression (8) with respect to the time t.

However, it is assumed that x _{t = 0} = 0 and v _{t = 0} = 0.

In this case, the speed of the robot hand is the output shown in FIG.
As shown in 32, the speed converges unlike the first and second operation examples described above. That is, the force can control the speed of the robot hand.

Although three simple operation examples have been shown as described above, the present invention has a possibility of executing a number of other useful controls.

The principle and operation example of the case of controlling a robot assuming a tool having one degree of freedom have been described above.

However, in reality, the tool is rarely provided with one degree of freedom, and a tool with multiple degrees of freedom is often used.

In the case of multiple degrees of freedom, the following equation can be obtained by generalizing the equation (1) using a vector.

Where: Position vector (translation position and rotation angle of robot in Cartesian coordinate system): Velocity vector (translation speed and rotation angle of robot in Cartesian coordinate system): Force signal detected by force sensor (translation force and rotation) (Including force) r: Vector of force reference value c: Vector of position reference value [M]: Matrix showing virtual mass of tool [C]: Matrix showing virtual viscosity coefficient of tool [K]: Spring of tool This is a matrix showing constant coefficients. Here, [M], [K], and [C] are generally diagonal matrices.

The formula (9) is transformed into the formula (10).

= [M] ^-1 ∫ [-r)-([C] + [K] (
-R))] dt (10) Here, the calculation by the equation (10) is actually carried out by a discrete value calculator in a computer as follows.

However, n = 1,2 In formulas (11) and (12), Δt is a scalar indicating the sampling time [I]: unit matrix Note that the initial value ₀ in formula (12) must be determined separately, and In the case of multiple degrees of freedom, it is necessary to assume a tool with the same or lesser degrees of freedom as the robot.

The case of multiple degrees of freedom has been described above. However, as mentioned in the explanation of one degree of freedom, the method of obtaining the position information and the speed information necessary for calculating the speed from the detector by hardware and the software There is a method of using the value of the calculation result each time, and various combinations are possible. An example is shown below.

FIG. 7 is a block diagram showing a first method of using position information from the encoder PE for position feedback and using a value of a calculation result for speed conversion.

In other words, the value detected by the encoder is used on the right side of equation (10), and on the right side of equation (10),
Is a method of directly using the value calculated by the calculating means, not the value detected by the tachogenerator TG.

Where from Is converted using a matrix [J] ^-1 ([J] is called a Jacobian matrix), and conversion from to is performed by coordinate conversion from the robot joint coordinate system to the orthogonal coordinate system. .

Next, FIG. 8 is a block diagram showing a second method that uses information from the detector for both position feedback and velocity feedback.

Here, the value n in the discrete value system is obtained by the following equation, for example.

n = Δtn-1 + x _n -1 (13) However, n = 1,2, ... Further, FIG. 9 is a block showing a third method in which the position feedback and the velocity feedback use the calculated values. It is a diagram.

In other words, this method is to feed back both position and velocity feedback using the value of the calculation result.

In this case, the initial values ₀ and ₀ in the equation (12) need to be determined separately.

The three methods have been described above, but each has the following characteristics.

That is, for the purpose of feedback, it is preferable to detect the actual values of both the position and the velocity, and return the feedback, as in the second method (the method of FIG. 8). However, judging from the current technical level, the position detector PE attached to the robot has a fairly good accuracy, but the speed detector TG is inferior to that. For this reason, the first method (the method of FIG. 7) that is an eclectic combination of both is actually the best.

The third method (the method shown in FIG. 9) is characterized in that it requires no detector at all, although the amount of calculation increases.

Since the tool has maximum 6 degrees of freedom, the matrix is a 6 × 6 matrix. However, since the degree of freedom, size, etc. of the tool change depending on the work, the matrix values of [K], [C], and [M] ⁻¹ are found by experiments according to the tool, and the controller 17
It is advisable to store each data in the ROM 21 in correspondence with tools and work contents.

Further, since the tool size is related to the vector amount, it is necessary to store it in each table of the ROM 21 according to the tool and the work content.

As described above, the present invention is characterized in that the tool can have various flexibility by controlling the robot.

Next, the calculation process of the angular velocity in the present invention will be described more specifically. Before that, the structure of the model robot,
Each coordinate system and symbol to be set will be described.

As shown in Fig. 10, the robot structure usually requires 6 elements for positioning in the space of the rigid body, and since the movable range is wider than the installation area, it is assumed that the robot has 6 degrees of freedom. did. This articulated robot has a rotation angle θ _i (i = 1 to 6) of each joint and each arm length d ₁ ,
It is represented by a ₁₂ , a ₂₃ , a ₃₄ , d ₅ (), and each coordinate system described below is set. First, a base coordinate system fixed to the place where the robot is installed (hereinafter referred to as B system), a hand coordinate system fixed to the hand provided at the tip of the robot (hereinafter referred to as H system), and sensor coordinates fixed to the sensor unit. System (hereinafter S
System), an auxiliary coordinate system (hereinafter referred to as system A) that sets the operating conditions and calculates the command speed by using the reference point and the origin of the rotational moment of motion, and is determined by the equilibrium position / posture of the spring. It is determined by the reference coordinate system corresponding to the H system (hereinafter referred to as the RH system) and the equilibrium position / posture of the spring.
There is each coordinate system of RA system).

Furthermore, in order to show the position and posture of the hand, the following four 3
Determine the dimension vector.

: Origin of H system displayed in B system: Unit vector in x axis direction of H system displayed in B system: Unit vector in y axis direction of H system displayed in B system: Z of H system displayed in B system Axial unit vector According to the present invention, the robot structure, coordinate system and symbols above are used. An example of the calculation method of is specifically described.

FIG. 11 shows the block diagram of FIG. 3 in more detail. As described above, the control device 17 is divided into four main blocks of the position / orientation calculation unit 40, the force calculation unit 50, the main control unit 60, and the joint angular velocity calculation unit 70.

The force calculation unit 50 further includes a conversion unit 51, a coordinate conversion unit 52, a gravity component calculation unit 53, and a coordinate conversion unit 54. When an external force a is applied to the hand of the robot, the sensing part of the force sensor 16 provided in the robot, for example, the strain gauge, senses this.
The applied force and moment are output as voltage to the conversion unit 51 via the amplifier 29 and the A / D conversion circuit 30. Converter 5
1 is to multiply the transmitted voltage by the voltage / force conversion matrix [A] ⁺ , and first calculate the force s in the S system. This conversion matrix [A] ⁺ is the force sensor
Calibration is performed by applying standard loads to 16 in each axial direction to obtain a matrix [A] having a relation of = [A] s, and [A] ⁺ = ([A] ^T [A]) ^-1 [A ]
It is obtained from the equation of ^T.

In this way, the conversion matrix [A] ⁺ is calculated and stored in advance, and is treated as a constant during the control calculation.
One force sensor corresponds to one conversion matrix [A] ⁺ , and when changing to another force sensor, the conversion matrix [A] ⁺ corresponding to that force sensor can be obtained and dealt with. Since the calculated force s is a value in the S system, the following coordinate conversion unit is used to coordinate-convert it into the H system.
It is output to 52.

That is, the coordinate conversion unit 52 determines the origin position and orientation of the S system in the H system (unit vector of each axis). Is used as a parameter, the force s of the S system is the force _H of the H system
Convert to. The converted force _H is output to the next correction unit 54.

On the other hand, the position / orientation of the hand in the B system calculated by the position / orientation calculation unit 40 as described later, (,,
), The gravity component calculation unit 53 uses the weight W of the hand and the object gripped by the hand and the position of the center of gravity as parameters, and the force exerted on the force sensor 16 by the own weight of the hand and the object itself, that is, the hand and the object. The force g corresponding to the gravity is calculated and output to the correction unit 54. The correction unit 54 removes the influence of the gravity of the hand and the object from this g and the above-mentioned _H , and obtains the force actually applied to the hand from the formula = _H - _g . The calculated force after correction is in the H system and needs to be converted into the A system in which the speed is calculated. Therefore, in the coordinate conversion unit 55, the origin position and posture of the A system in the H system (unit of each axis). vector) The corrected force is coordinate-converted to the A system by using as a parameter, and then output to the main control unit 60.

Note that it is not necessary to perform speed calculation in the A system depending on the mode of operation control of the hand.
Coordinate transformation at 55 is unnecessary.

The position / orientation calculation unit 40 includes a calculation unit 41, an A system position / orientation calculation unit 42, an RA system position / orientation calculation unit 43, an A system and RA system position / orientation deviation calculation unit 44, and a coordinate conversion unit 45. Composed of.

First, the calculation unit 41 determines the encoder PE connected to the motor M.
Based on the position signal obtained by the counter 28 from the position information of the robot, each arm length of the robot and the position / posture _E , ( _E , _E , of the hand with respect to the arm tip portion)
Position of hand in B system with _E ) as a parameter
Calculate the posture, (,,), that is, From each expression of ,,,, are obtained respectively.

Here, [A _1h ] represents a transformation matrix from the first coordinate system to the H coordinate system, and is a function of ,, _E , _E , _E , _E. Calculated hand position / orientation, (,,
) Is output to the gravity component calculator 53 in the force calculator 50, the angular velocity calculator 70, which will be described later, and the position / orientation calculator 42 of the A system. In the RA system position / orientation calculation unit 43, the H system (hand)
Equilibrium position / posture r of translation and rotation springs corresponding to
From h, (rh, rh, rh) Is used to correspond to system A. Calculate the equilibrium position / posture ra, (ra, ra, ra) of the translation spring and the rotation spring.

Furthermore, in the position / orientation deviation calculation unit 44 of the A system and the RA system, r
From a, (ra, ra,) and a, (a, a, a), the position deviation Δ and the attitude deviation are calculated by the following equations.

Here, from the posture (ra, ra, ra), (a,
The unit vector in the axial direction for rotating to the posture a, a) means the rotation angle about the axis.

Further, based on these calculation results, the position / orientation deviation Δ in the B system is calculated by the following equation.

As is clear from this equation, the deviation Δx is
It can be changed by changing ra, (ra, ra, ra). The calculated deviation Δx is converted in the coordinate conversion unit 45 from the B system to the A system in which the speed is calculated.

When the speed is calculated in the H system according to the mode of operation control of the hand, the deviation Δx is converted into the H system, but here, the case where the speed is calculated in the A system will be described.

Next, it comprises an operating condition setting unit 61, a command speed calculation unit 62, and a coordinate conversion unit 63.

First, in the operation condition setting unit 61, the external force applied to the hand in consideration of the operation conditions of various hands in the A system *
By calculating, the hand can be made to perform a high degree of exercise.

A specific application example will be described later.

The command speed calculation unit 62 includes a force calculated by the force calculation unit 50 (through the operation condition setting unit) and a position / orientation calculation unit 40.
Using the force reference value c, the mass matrix [M], the viscosity constant matrix [C], and the spring constant matrix [K] as parameters from the position / orientation deviation Δ calculated in step 2, the hand command speed in the A system is calculated by the following equation. To do.

= [M] ^-1 ∫ (-c- [K] Δ [C] v) dt ...
(19) As described in the above-mentioned principle explanation, according to this formula, [M],
Since the behavior of the hand when a force is applied by [K] and [C] is determined, the operating characteristics of the hand should be set by appropriately setting the parameters [M], [K], and [C]. You can

Further, by appropriately setting the force reference value c, the pressing force can be held constant or changed, and various hand movement characteristics can be obtained.

Since the command speed calculated above is the value in the A system, the above-mentioned parameter in the coordinate conversion unit 63 is used. Is converted to the B system by using and is output to the next angular velocity calculation unit 70.

Since the robot is actually driven and controlled by the angular velocity of each joint, the command velocity is calculated in the angular velocity calculation unit 70 by a position signal indicating the rotation angle of each joint of the robot, each arm length of the robot, and the conversion matrix (described above). A _1h ] as a parameter, Is converted to. this Is input to the motor M for driving each arm through the D / A conversion circuit 27 and the servo amplifier 18, and the robot hand operates at a predetermined speed.

Next, a specific application example of the control method according to the present invention will be described below.

First, in the first application example, the RA system does not move and is fixed to the B system, so that the A system matches the H system, so that the RA system matches the RH,
This is an example in which the initial position and orientation of the system match the RA system.

The operation of the hand in this case will be described with reference to FIG. First
As shown in FIG. 12A, when a force in the y-axis direction is applied to the hand 15 in the H system, the external force in the H system becomes Is shown. Here, in the equation (19), c = 0 Then, the hand 15 receiving the external force stands still in the position / posture shown in FIG. 11 (b). Here, Δy = _fy / h ₂₂ (24) Δθx = mx / h ₄₄ (25), and according to the control method of the present invention, the parameter [K]
It means that the same operation as when a spring is added to the wrist of the robot is obtained by appropriate setting of.

That is, when the force is removed instantaneously origin O _H hand is a simple harmonic motion comprising amplitude Δy around the initial position of the origin O _H origin i.e. the hand reference value r of the position and orientation of gripper 15 Figure 12 As shown in (b), a simple oscillation of rotation indicated by an angle having an amplitude Δθx with the initial posture as the center is performed.

That is, the vibration oscillations in the attitude of the vibration and the hand of the origin O _H hand were synthesized hand 15 is performed, in particular the position of the hand 15 since the attenuation represented by the formula (6) Is as follows.

In other words, if any translational force / rotational force is applied to the hand 15,
The tip of the hand 15 is at the initial position, and the posture of the hand 15 is a restoring force and a damping force for returning to the initial posture.

Therefore, although there is actually no spring, the hand 15 moves as if it were suspended in the air by a spring having 6 degrees of freedom that satisfies the above conditions.

Next, by setting [K] = [◯], it is possible to cause the hand to perform qualitatively different actions. In this case, since there is no spring force, that is, the restoring force, the hand does not try to return to the original position or posture, and the hand translates in the direction of the translational force applied to the H system and rotates around the axis that is the center of the rotational force. Rotate to. That is, in the midair, it is possible to realize a motion in which the translational force and the rotational force are applied in the direction and magnitude as they are, and when the application of force is stopped, the motion is stationary.

For example, if a force as shown in Fig. 12 (a) is applied, the speed at this time is calculated from equation (8). Becomes

In the above description, the spring constant matrix [K] of 6
We considered all the cases where the diagonal elements are not all 〇, but if you set some of these six diagonal elements to 〇 and not to some 〇, for example, if you apply force, the hand As for the posture, the restoring force works, and it always tries to return to the original posture, but the position moves as it is pushed.
The hand can have a motion characteristic different from one motion characteristic.

Next, a second application example will be described.

The second application example is a case where the RA system moves, and the A system is made to match the H system, so that the RA system and the RH system match and the RA system moves.

The operation of the hand in this case will be described with reference to FIG.

In the first application example described above, the RA system is fixed to the B system. For this reason, the hand exerts a restoring force to the position / posture indicated by the RA system, and since the restoring force was proportional to the deviation from the RA system, the hand actually works only near the RA system. You cannot do it.

However, this example is characterized in that the position / orientation of this RA system is moved on the B system. To move the RA system means to move the equilibrium position / posture of the restoring force, so the hand moves in a manner that follows the movement.

For example, as shown in FIG. 13, when the RA system position / orientation is moved, the hand position / orientation also moves,
At the same time, it also has the operating characteristics shown in the first application example.

That is, if a force is applied while the hand is moving, the RA system moves out of the moving path according to the force, but at that moment, a restoring force to the RA system works at the same time, so the hand returns to the moving path.

That is, in a robot having a mechanism in which a spring is added to the wrist, a motion as if the hand is moving can be realized simply by giving a control parameter. The operation at this time is shown by the equation (7), but it is basically the same when comparing this with the equation (6) showing the operation when a force is applied when the RA system is at rest. Change. This suggests that if the force applied to the robot (when the RA system is stationary) and the amount of movement of the RA system are appropriately determined, the hand can perform the same operation even if both are replaced.

As a third application example, an example in which the operating condition of the hand is set will be shown.

In the first example, the operating conditions are only rotation speed, A system does not match H system, therefore RA system does not match RH system and RA system is B
It is fixed to the system and matches the initial position and orientation of the A system.

The operation of the hand in this case will be described with reference to FIG.

The tip of the shaft 80 of the hand 15 is the origin O _{A of the} A system. The basic operation is the same as that shown in FIG. 12, except that the A system is at a different position from the H system.

Further, the translational force and the rotational force about the Z axis were set to O as the constraint condition of the operation.

That is, the translational force in the A system is O, and therefore the origin O _{A of the} A system is
Only rotational force around is allowed. However, the rotational force around the Z axis is O
And Therefore, within the range that can be detected by the force sensor, no matter which point of the hand and shaft is applied in any direction (however, the line of action does not pass the origin),
The hand and shaft perform only a combined rotational movement about the x-axis and the y-axis. And if the force is removed, it will return to the original posture.

This operation realizes the rotation of the shaft (workpiece) around the tip point, which is the essence of the RCC mechanism described above, and is useful for automatic assembly, particularly for precision fitting work.

In the following example, the motion condition of the hand is only translational force, and the A system is H
Therefore, the RA system matches the RH system, and the RA system is fixed to the B system to match the initial position and attitude of the A system.

The operation of the hand in this case will be described with reference to FIG.

Although the basic motion is the same as that shown in FIG. 12, the rotational force is set to O as a constraint condition of the motion.

That is, the rotational force in the A system was set to O. Therefore, only translational movements are allowed. Therefore, no matter what force is applied to any point, the hand makes a translational motion with the posture kept constant. Then, if the force is removed, it returns to the original position.

Further, in the equation (23), h ₁₁ and h ₂₂ are set to be large and h ₃₃ is set to be small.

In other words, it works as a soft spring in the x-axis and y-axis directions of H system,
It works as a hard spring in the Z direction. Therefore, as shown in Fig. 15, if the same force is applied, it will swing greatly in the Z-axis direction, and xy
The shake is small on the plane.

Further, here, since the A system is fixed to the H system, as shown in FIG. 15, this property (great swing in the Z axis direction) is
Even if the posture of the hand changes, the hand always swings largely in the central axis direction (that is, the Z axis direction of the H system).

In the last example, the motion condition of the hand is the translation condition only, and the origin of the A system is made to coincide with the H system, but its axial direction is B
Match the system.

The operation of the hand in this case will be described with reference to FIG.

In this example, since the direction of the A system is different from that of the H system, the direction in which the hand swings greatly changes when viewed from the hand system. However, it is unchanged from the B system. That is, the direction of large swing is always the Z axis direction of the B system.

Therefore, it can be seen from the above two examples that the operating condition of the hand can be changed in either the H system or the B system by one method of defining the A system.

Next, an application example in which information from a non-contact sensor is incorporated in the control method according to the present invention will be described.

Fig. 17 shows that if a robot enters the danger area while it is being position-controlled in a certain work area space, it will avoid entry into the danger area by a non-contact sensor. Is. As described above, the position control of the robot is performed by operating the position reference value ra, but if this is done, if ra is in the work area space, the robot will be in the work area space no matter how it is operated. Is expected to stay in.

However, in reality, the set parameter [M],
Due to [C], [K] and delay of robot drive system,
Alternatively, the robot may enter the work area, that is, the dangerous area due to malfunction or disturbance. In the example shown in FIG. 17, the walls 116,
There is 117, and a region within a distance of Δlc from the walls 116 and 117 is set as a dangerous region. Therefore, as a robot, at least two non-contact sensors 11 are required to measure the distance between the walls 117 and 116 in the positive and negative directions of the x-axis.
8 is required, and it is known by these non-contact sensors 118 that the robot hand has entered any of the dangerous areas.

For example, assuming that the robot hand is approaching the wall 117, it is known that the distance Δlx + from the wall 117 has reached Δlx + <Δlc, and that the robot hand has entered the dangerous area. If the robot's hand 15 enters the dangerous area, Δlx + should be Δlc so that the robot's hand 15
It is necessary to push back.

As shown in FIG. 18, when it is detected that the dangerous area is invaded at a certain time point tc, it immediately takes the action as shown in FIG. 5 and smoothly exits the dangerous area. Of course, as has been done conventionally, it is possible to make an emergency stop at that time tc, or to move out of the dangerous area with a large acceleration immediately instead of smoothly.

By the way, the above-mentioned control by the non-contact sensor can be easily achieved by adding the feedback signal from the non-contact sensor to the control method according to the present invention described above.

In addition to the control described above, control shown in FIG. 19 is also considered. 19 (a) and 19 (b), when the robot approaches a human and the distance between them becomes smaller than Δlc, x +, the robot is controlled to return from the human to the position of Δlc, x +, and On the contrary, it is shown that when a human approaches a robot, the same control is performed. The control shown in FIG. 19 corresponds to a case where the wall can be moved, so to speak, and non-contact direct teaching is possible by such control.

In each of the above two application examples, the distance between the robot and the object is detected, and the robot is controlled based on the relationship with a certain reference value. As an extension of such control, the one shown in FIG. 20 is considered. In the case shown in Fig. 20, a non-contact sensor
In addition to 118, a force sensor 16 as a contact sensor is attached to the robot, and hybrid control is possible.

For example, assuming the case where the shaft-shaped work 119 is assembled to the work 120 having a hole, the robot needs to convey the work 119 to the vicinity of the work 120 after gripping the work 119.
It is desirable that the conveyance to the work 120 is performed at high speed, but in the case of high-speed conveyance, a large impact is applied at the time of contact with the work 120, so it is necessary to reduce the conveyance speed immediately before the contact. Generally, when only the contact sensor is attached, the contact is detected for the first time with a large impact, or the work is conveyed at low speed to avoid a large impact, but in any case, a large impact or This will cause problems such as low speed.

However, when the non-contact sensor is also used, a large impact can be avoided and the work can be conveyed at high speed.

That is, the contact speed is reduced immediately after the contact of the work 119 with the work 120 is detected by the non-contact sensor 118,
After the work 119 comes into contact with the work 120 in the low-speed conveyance state, the work 119 is to be attached to the work 120 by the contact sensor 118.

In the application example described above, if the non-contact sensor detects that the distance between the robot hand and an object such as an obstacle is within a certain range, the control mode for the robot is changed at the time of detection. It is the one.

Therefore, the robot is appropriately controlled in relation to the target object, and when a contact sensor is also used, it is possible to perform coordinated control with it, and various control operation characteristics can be obtained by simply changing the control parameters. There is an effect that it can be easily realized.

〔The invention's effect〕

As described above, according to the present invention, flexibility can be provided by controlling the robot, and the following operational effects can be obtained.

That is, it is possible to assume that a tool that performs a required operation is assumed, the robot can be caused to perform an operation equivalent to that tool, and the assumed tool can be converted.

Conventionally, the tool was replaced successively, but according to the present invention, this replacement can be performed by simply replacing the parameter for control.

Further, although the same tool is used, even when the robot is caused to perform an operation with slightly different characteristics, conventionally, it was necessary to adjust a part of the sequential tool, but according to the present invention, online With, it is possible to add a convenient function that can convert the characteristics of the tool only in terms of control.

As described above, the present invention has been consistently described in terms of the tool substitute function. However, according to the present invention, a simulation function can be provided in addition to this.

For example, when designing a wrist mechanism to be attached to a robot, and when it is desired to see the behavior of the mechanism before actually making an object, the method of the present invention can be applied to cause the robot to actually perform the operation.

The higher the accuracy of the robot, the higher the effects of the above-mentioned tool substitution and simulation functions. However, in reality, in a robot, unnecessary vibration and delay may occur due to mechanical factors such as vibration due to elasticity of the arm portion and insufficient rigidity of the reduction gear, and due to other reasons such as electric circuits.

In such a case, if the vibration and delay of the robot can be analyzed and predicted, a compensating operation for reducing the vibration can be reversely added according to the present invention.

As described above, the present invention has various useful effects.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic configuration diagram showing an overall configuration of a robot control method of the present invention, FIG. 2 is a configuration diagram showing components of a robot control device of the present invention, and FIG. FIG. 4 is a schematic block diagram showing the robot controller of FIG. 4, FIG. 4 is an explanatory view showing a first operation example of the robot by the control method of the present invention, and FIG. 5 is a second operation example of the robot by the control method of the present invention. 6 is an explanatory view showing a third operation example of the robot according to the control method of the present invention, and FIG. 7 is a block diagram showing a first embodiment of a calculation flow in the control method of the present invention. , 8th
FIG. 9 is a block diagram showing a second embodiment of the operation flow in the control method of the present invention, and FIG. 9 is a block diagram showing a third embodiment of the operation flow in the control method of the present invention.
FIG. 10 is an explanatory view showing the configuration and coordinate system of the robot, FIG. 11 is a detailed block diagram showing the robot control device of the present invention, and FIGS. 12 (a) and 12 (b) are specific examples of the control method of the present invention. FIG. 13 is an explanatory view showing a first specific application example, FIG. 13 is an explanatory view showing a second specific application example of the control method of the present invention, FIG. 14, FIG.
FIG. 16 is an explanatory view showing a third specific application example of the control method of the present invention, and FIGS. 17, 18, 19, 19 and 20 are the control method and non-contact sensor according to the present invention. It is explanatory drawing which shows the Example which combined. 10 …… (joint type) robot, 15 …… hand, 16 …… force sensor, 17 …… control device, 18 …… servo amplifier, 40 …… position / orientation calculation unit, 50 …… force calculation unit, 60… … Control main part,
70 ... Joint angular velocity calculator.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Shiyuki Sakagami 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Inside the Hitachi, Ltd. Institute of Industrial Science (56) References JP-A-60-254209 (JP, A) JP-A-60-205716 (JP, A)

Claims (4)

[Claims] 1. A robot control method for controlling a robot at every sampling time, wherein a force signal is calculated by detecting an actual force applied to a robot hand at a previous sampling time (n-1),
The first value, which is the difference between this signal and the force reference value that was the target value at the previous sampling time (n-1), is calculated, and the actual position and posture of the hand at the previous sampling time (n-1) are calculated. A second value is calculated by multiplying the difference between the position signal that represents the position signal and the position reference value that was the target value of the position and orientation at the pre-sampling time (n-1) by the spring constant, and the pre-sampling time of the hand ( As a velocity signal in n-1), a value obtained by differentiating a position detection value from a position detector provided in the robot is used, and a third value is calculated by multiplying this by a viscosity constant, and the first, second, A method for controlling a robot, comprising calculating a command speed of a next sampling time (n) of the hand from a third value and a mass coefficient of the hand. 2. The robot control method according to claim 1, wherein a position detection value from a position detector provided in the robot is used as the position signal. 3. A robot control method according to claim 1, wherein position information from a non-contact sensor provided in the robot is used as the position detection signal. 4. A robot controller for controlling a robot at every sampling time, wherein a force signal is calculated by detecting an actual force applied to a robot hand at a previous sampling time (n-1),
Calculating means for calculating a first value which is the difference between this signal and the force reference value which was the target value at the pre-sampling time (n-1), and the actual value of the pre-sampling time (n-1) of the hand. Calculating means for calculating a second value by multiplying a spring constant by a difference between a position signal representing the position and the attitude and a position reference value which was the target value of the position and the attitude at the previous sampling time (n-1); A calculating means for calculating a third value by using a value obtained by differentiating a position detection value from a position detector provided in the robot as a velocity signal at the hand pre-sampling time (n-1) and multiplying this by a viscosity constant. And a calculation means for calculating a command speed for the next sampling time (n) of the hand from the first, second and third values and the mass coefficient of the hand.