JPH0415715A - Robust velocity/position controller - Google Patents

Abstract

PURPOSE:To attain the highly accurate velocity/position control by detecting the velocities of a 1st and a 2nd controlled means to apply the prescribed arithmetic processing to each detected velocity for addition of both velocities and using the velocity feedback information and the derivatives as the acceleration feedback information. CONSTITUTION:A 2nd compensation means 116 performs a desired operation based on the velocity output value Vm of the 1st controlled means 112 and outputs the result of operation to an adder 122. Meanwhile a multiplier 118 performs a desired operation based on the velocity output value Vo of the 2nd controlled means 114 and outputs the result of operation to the adder 122. The adder 122 outputs the sum of both outputs of the means 116 and the multiplier 118 as a synthetic feedback signal Vf. This signal Vf is engatively supplied to an addition/subtraction point 12 at the precedent stage of a 1st compensation means 110. The means 110 includes a velocity compensation element Gv 110a, an addition/subtraction point 110b, and a gain constant Ka. The signal Af is negatively fed back to the point 110b. Thus the highly accurate velocity/position control is attained.

Description

DETAILED DESCRIPTION OF THE INVENTION FIELD OF INDUSTRIAL APPLICATION The present invention relates to an improvement in a robust control device for velocity or position, particularly when the second controlled means is oscillatory.

[Prior Art] Accurate speed control and position control are required in the fields of various processing devices and measuring devices, and such speed or position control devices are It includes a control means and a second vibratory controlled means.

That is, in conventional speed or position control devices, the controlled means is non-oscillatory and the relative order of its transfer characteristic is O, -1.
, -2, and a second controlled means including an oscillatory element whose relative order is -2. Then, the output of the velocity dimension and the output of the position dimension, which are the behavior of the second controlled means, are controlled to desired behavior.

Typical examples of such controlled means include servo motors, which are widely used in industry and are key elements of automated and unmanned production equipment, as well as servo motors that process source energy to generate controlled energy. Examples include a first controlled means composed of a power amplifier that supplies a motor, and a robot arm driven by the motor, or a table for moving a heavy object driven via a joint having spring-like characteristics. The second controlled means is vibratory and has a natural vibration.

FIG. 11 is an explanatory diagram of the physical system of a conventional general speed control device.

The speed control device shown in the figure includes a first compensating means 10 and a first compensating means 10.
It includes a controlled means 12 and a second controlled means 14.

Then, the command speed Vc is processed by the first compensation means 10 in a feedback manner based on the detected speed Vm of the first controlled means 12 or the detected speed VO of the second controlled means 12, and outputs the operation information Ua. It is.

[Problems to be Solved by the Invention] Conventionally, there have been many proposals both theoretically and experimentally to make the kinematic behavior of the non-oscillatory first controlled means desired, such as its velocity and position. There are also examples of this technology being used in products on the market.

However, these
As shown at the end of pages 5 to 832, in both cases, the first controlled means, that is, the relative order is 0. -1. -2 and was limited to non-vibration control devices.

On the other hand, if the rigidity of the second controlled means is sufficient or the accuracy of the speed Vo of the second controlled means 14 does not matter, it is sufficient to feed back only the speed Vm of the first controlled means 12. Speed V.

When placing importance on the accuracy of the speed Vo, it is necessary to feed back the speed Vo.

In such a situation, when the controlled object includes a vibrating second controlled means, few effective methods have been proposed, and even in the examples that have been proposed, (1) disturbance is applied. (2) The physical constants (parameters) that define the characteristics of the model do not change.

(3) In most cases, it is just a theory based on unrealistic assumptions, such as that the relative order of the model characteristics does not change, and the results of applying these to reality are easily affected by external disturbances. As a result of changes in the model, the optimal control state is not maintained, and the stability is significantly degraded, even becoming unstable.
The control methods praised as such failed to have a major impact on industry. Therefore, as a method for compensating the characteristics of the control system, as shown in FIG.
0 in the control system, and in order to converge the difference between the command and the feedback information to zero as quickly as possible, the characteristics of the first compensation means 10 include proportional (p) sufficient integral CI) characteristics and increase for specific frequency components. It has a notch filter characteristic that makes the width small, and the parameters that define the characteristic are determined mainly empirically.It cannot cope with changes in the characteristic of the non-linear model, but it compromises with insufficient suppression of external disturbances. There was nothing else.

The present invention has been made in view of the problems of the prior art described above, and its purpose is to provide robust speed and position control that enables accurate speed and position control even when the second controlled means is vibratory. The goal is to provide equipment.

[Means for Solving the Problem] In order to achieve the above object, a speed robust control device according to the present application includes a second compensation means, a multiplication means, an addition means, and a third compensation means. It is characterized by

That is, the second compensation means transmits and outputs the detection information of the first output using the following formula [1].

(Gn-α)/Gm...■Gm: the above G
A polynomial Gn having the same order as fk and a vibration characteristic plate whose natural angular frequency ωm≦ωfk: a polynomial Gn having the same dimension as Gm and whose natural angular frequency ωn
≧ωfk and governing the behavior of the output of the second controlled means.
Constant) Multiply and output.

The addition means adds the outputs of the second compensation means and the multiplication means and outputs composite feedback information.

The third compensation means pseudo-differentiates the synthetic feedback information and negatively feeds back the pseudo acceleration control information to the path of the first compensation means to form an acceleration loop.

The composite feedback information is also used as feedback information in the velocity loop.

Further, the acceleration loop has a synthetic relative order of the first controlled means in the loop and each of the compensation means within the range of 01-1-2 to obtain a desired synthetic amplification degree of the compensation elements in the acceleration loop. The first output is determined so that disturbances applied to the acceleration loop and fluctuations in parameters of the first controlled means do not affect the first output.

The first constant α and the second constant β are characterized in that they are determined so as not to affect the second output.

Further, in the present invention, it is possible for the first compensation means to include a saturation element.

Further, in the present invention, for the constants α and β, 0≦α
, β≦1.

In the present invention, it is preferable that α=β for the constants α and β.

Further, in the present invention, it is preferable that the second controlled means has, in addition to the oscillatory characteristics, a characteristic of relative order -3, which is a first-order lag characteristic having a small time constant.

Further, in the present invention, the speed and position robust control device uses external command information in the position dimension and feedback information in the position dimension obtained by integrating the detection information of the second output to loop the position. The present invention is characterized in that the speed dimension command is generated using the difference between this command information and this position dimension feedback information.

Note that it is preferable that the second controlled means has a characteristic of relative order -3, which is a first-order lag characteristic having a small time constant in addition to the vibrational characteristic.

Further, the present invention is also applicable to the case where the second controlled means has non-oscillatory characteristics, and its relative order has −3 characteristics.

[Operation] As described above, the present invention is a control device that targets a controlled operation that includes a vibratory second controlled means in addition to a non-oscillatory first controlled means, in which a disturbance applied to a controlled object is used. A second compensating means is provided for compensating when there is a change in the parameters of the first and second controlled means.

Then, the detected information of the input of the vibratory second controlled means is supplied to the input of the second compensating means, and the detected information of the output of the second compensating means, the force, and the output of the second controlled means are set to a predetermined constant. (hereinafter referred to as the second constant β) and the multiplied output of the multiplication means to create feedback information, which is firstly fed back to the velocity loop, and secondly, through the third compensation means to generate pseudo acceleration information. A, and this is negatively fed back to the forward path of the first compensation means to form an acceleration loop, and the relative order in the acceleration loop is set to 0. -1. -2 or less, the composite gain Ka of the acceleration loop can be set to a desired large value.

On the other hand, the characteristic of the second compensation means is the characteristic polynomial G of its denominator.
I11 is a vibrational characteristic, one natural angular frequency ωm
, and this ωm is an oscillatory characteristic polynomial Gfk (k=1.2., .) of the oscillatory second controlled means.

The natural angular frequency ωfk (k=1.2., 1) of Gfk is determined in anticipation of variations in the parameters of Gfk, so that it does not become larger than the natural angular frequency ωfk (k=1.2.,,l).

Next, the characteristic polynomial of the numerator of the second compensation means has the form Gn-α, which is obtained by subtracting a predetermined constant (hereinafter referred to as the first constant α) from the characteristic polynomial Gn that governs the behavior of the second controlled means to the desired behavior. and the degrees of the polynomials of Gn and Gm are made equal, the relative degree of the second compensation means is O, and the composite gain Ka is set to a large value.

As a result, (1) the disturbance TL applied within the acceleration loop does not affect the input of the vibratory second controlled means;

(11) Variations in the parameters of the non-oscillatory elements in the acceleration loop do not affect the input of the oscillatory second controlled means.

(III) Oscillatory polynomial Gfk of the oscillatory second controlled means
This effect is canceled by the effect of the oscillatory polynomial Gm of the second compensation means, so that no oscillatory behavior appears in the output of the second controlled means.

(rV) polynomial G of desired characteristics in the molecule of the second compensation means
By controlling the behavior of the output of the second controlled means by n, the influence on the disturbance TL and the fluctuation of the parameters of the first and second controlled means is extremely small, that is, the degree of robustness is high (hereinafter referred to as robust). The present invention provides a control device that controls the speed command Vc and the speed-dimensional output Vo of the second controlled means by a characteristic polynomial Gn having desired characteristics, and is widely used in industry. The aim is to meet the expectations of

[Example code] Hereinafter, preferred embodiments of the present invention will be described based on the drawings.

FIG. 1 shows a physical system of a speed robust control device 100 according to an embodiment of the present invention, and parts corresponding to those in the prior art are indicated by the reference numeral 100 and their explanation will be omitted.

A characteristic feature of the present invention is that in addition to the first compensation means 110, a second compensation means 116, a multiplication means 118, and a third compensation means 120 are provided.

The second compensating means 116 is controlled by the first controlled means 11.
A desired calculation is performed based on the speed output Vm value of No. 2, and the result is output to the adder 122.

Further, the multiplier 118 performs a desired calculation based on the speed output vo value of the second controlled means 114 and outputs the result to the adder 122.

Then, an adder 122 converts the sum of the respective outputs of the second compensation means 116 and the multiplier 118 into a synthesized feedback signal V
f and is negatively supplied to the addition/subtraction point 124 in the preceding stage of the first compensating means 110.

On the other hand, the composite feedback signal Vf is pseudo-differentiated by the third compensating means 120, and outputs an acceleration-dimensional feedback signal Af.

Here, the first compensation means 110 includes a speed compensation element (G
v) 110a, addition/subtraction point 110b, gain constant Ka
The feedback signal Af includes an addition/subtraction point 110
negative feedback to b.

The speed robust control device according to this embodiment is roughly constructed as described above, and its operation will be explained next.

First, the first controlled means has a transfer characteristic of 1/
When G a is used, and the operator is represented by S using the Laplace transform well known in the automatic control field, motor: kt
/ (Js+D), power amplifier (characteristics improved by current feedback): 1/(Tas+1)
It is expressed as If the motor is an AC servo motor, the coefficient of friction is negligibly small, so it is t.

J: Intensity of the motor and rotating system kt: Energy conversion constant Ta: Time constant (current applied to the motor - torque generated from the motor).

The input of this element is operation information Da in the torque dimension or acceleration dimension, and the output is the motor speed Vm. The dimension of Jo is rad/5ec2・A (A is ampere, the unit of electric current), the dimension of Ta is sec, and the dimension of s is 5ec-
'is.

Note that a hydraulic cylinder may be used instead of the motor, and a servo valve may be used instead of the power amplifier. In these cases, the characteristics can also be approximately expressed as 1/Ga.

The characteristic 1/Gf of the second controlled means 114 is composed of several vibrational elements, and the second characteristic 1/Gfk is expressed as follows using the tamping coefficient ξ and the natural angular frequency ωfk. .

When using a motor, a power amplifier, or electric energy, the energy type used is electric power, and when a motor, power amplifier, or electric energy is used, the characteristic of only 1/Ga is sufficient in the motor or power amplifier, and the characteristic of 1/Gfk is not noticeable.

However, even in such cases, when industrial robots, machine tools, measuring instruments, etc. are driven by motors, one or more vibrational transmission characteristics are usually applied to these mechanical elements due to their mass and rigidity. This phenomenon occurs during high-precision control and cannot be ignored in the case of high-precision control. It is well known that in such a case, the vibrational characteristics are expressed as...■.

In the case of high-precision control where the form of energy used is air or fluid whose compressibility cannot be ignored, vibrational behavior such as (2) appears in the characteristics of cylinders and servo valves, and this also cannot be ignored.

On the other hand, the transfer characteristic 1/Gf of the second controlled means 114 is dimensionless, and the input of the second controlled means 114 is the output Vm of the first controlled means 112 in the speed dimension, and the second controlled means 114
The output is the velocity-dimensional output VO.

In engineering, what is important about the characteristics of elements used in control systems is that if we use the operator S used in the Laplace transform, these characteristics can be expressed as a polynomial in S for the numerator and a polynomial in S for the denominator. The degree (=the degree of S in the numerator−the degree of S in the denominator)≦0. That is, it is important that these elements be obtained by combining only integral characteristics.

The apparatus of this embodiment has input and output information Vm, V of the vibrational element of the second controlled means 114, 1/Gf.

A second compensation means 116 having Gm and Gn consisting of a constant α and a polynomial of S having the same degree as Gf.
and a multiplier 118 having a constant β, the sum Vf of the outputs of these elements is used as speed feedback information in the first
It is used in place of Vm to Vo in Figure 1.

Here, the relative order of the second compensation means 116 is 2-2=0. The relative order of the multiplication means 118 is 0-0=0.

Furthermore, in the apparatus of this embodiment, the speed feedback information Vf allows the third
Feedback information Af in the acceleration dimension is generated via the compensating means 120, and this is negatively fed back to the forward route of the first compensating means 110.

The first compensation means 110 has a commonly used configuration,
It includes a gain constant Ka, an addition/subtraction point, and a speed compensation element Gv.

For the element Gv, a proportional+integral, first-order lag, advance element, or notch filter is used, and its relative order is usually O or -1.

In the system of FIG. 1, the first and second controlled means 112 and 114
is a real system with physical substance, the characteristics 1/Ga, 1/
The parameters J°, Ta, ωfk, ξ, etc. that define Gf vary slightly.

On the other hand, each element of each means 110, 116, 118, and 120 is an element that inputs, outputs, and processes information, and is usually composed of digital circuit elements such as a microcomputer and memory. Since it is configured to take in and output information at a sufficiently fast sampling period, it nominally has the planned function, but the physical entity is a digital circuit system. As a result, the parameters defining the characteristics of the nominal elements of each of the means 110, 116, 118, and 120 are not affected by the outside world at all.

One of the objects of the present invention is that even if the characteristics of the actual system vary somewhat due to the novel control structure and characteristics of such nominal elements,
The behavior of the system output ■0 is governed by the nominal elements,
The objective is to provide a device that is not affected by fluctuations in the actual system.

Another object of the present invention is that even if the disturbance TL is applied,
Due to the novel control structure and characteristics of such nominal elements,
The objective is to provide a device in which the behavior of the system output Vo is not affected by the disturbance TL.

Next, ensuring robustness by the apparatus according to this embodiment will be explained.

FIG. 2 is a block diagram obtained by equivalently transforming FIG. 1.

As is clear from the figure, an acceleration loop is constructed with Vc' as input and Vo as output. What should be noted about the acceleration loop is the round-trip transfer characteristic G of the acceleration loop.
It is the relative order of A.

Since GaXGf, Gm, and Gn are all polynomials of degree 2 in S, the relative degree of GA=−2 (=degree 5 of numerator − degree near denominator) .

Therefore, the gain ka can be made sufficiently large according to the theory of the known automatic control system. As a result, the disturbance TL is also 1
/ ka, which can be ignored in practical terms.

As a result, the transfer characteristic Vo/Vc' of the acceleration loop is Vo (TcS + 1)Gm-...
・■ Vc' GnGf + (βGm −αGf).

Here, Gm is an approximation of Gf. In particular, when Gf is composed of only one vibrational element (for example, Gfl),
If Gm=Gf, ■ becomes. Even if Gf fluctuates at least a little, Gm remains Gf
When approximated, ■ holds true. (2) indicates that the real system Gm and Gf have no relation to the characteristics of the acceleration loop, that is, they are not subject to real system fluctuations.

If Gm is an approximation of Gf, then the feedback element Gm of the velocity loop of FIG. 2 is approximated.

Therefore, FIG. 2 becomes like FIG. 3.

In Fig. 3, the transfer characteristic GB for one round of this speed loop is
The relative order of is, since Gi is, the relative order = relative order of Gv = 0 to 1
... [phase] So,
In the speed loop shown in FIG. 3, the roof gain (=proportionality constant of Gv) can be set sufficiently large. At this time, the characteristics of the system shown in FIG. 3 are expressed by the reciprocal of the feedback element Gn+ (ρ-α).

In this way, the behavior of Vo is governed by Gn, α, and ρ, and Ga
, Gf and the disturbance TL.

In (①), in order for Vo = Vc (the value of Laplace transform operator S is s - 0), ωnlI
ωn.

In this case, the numerator of ■ should be multiplied by (1+β-α), or α=β.

In any case, it is sufficient to use the nominal elements α, p, or 1+ρ−α, and these are not affected by fluctuations in TL, Gn, and Gf at all.

When there is no modeling error (that is, when Gm = Gf)
As explained above, if the real system Ga and Gf fluctuate, modeling errors will occur.

Therefore, this effect will be considered next.

1) Fluctuations in Ga can be ignored by increasing the gain Ka of the acceleration loop (Ga does not appear in equation 0).

2') Fluctuation of Gf: In this case, the denominator of equation 0 or a stable polynomial of S is required. That is, GnGf +
(βGm-a Gf) (S'+Al53+A252+A3S+A4)/ω2
In nωj2-@, if Hurwitz's condition is applied, A, -2(ξfωf+ξn(JJn) > 0
・-@A2: ωf2+ωn2+4ξfωfξn
(JJn+(αωf2−βωm')ωn2/ωm">
O-QA3=2 [ξnωnωf2+ξfωfωn”
+(ξmωmaωf2−ξfωfβωm”)ωn2/
0m” > 0...[Phase] A4-ωn2ωf2〉0... Natural establishment and Hl・A,>0...
Same as 0 H2=AIA2-Al > O... [phase] Hi=AtAJs-As'-At 'A4''A3 (
AIA2-All) AI "A4 > 0...0 0 ~ From the @ formula, as a safe condition, ■≧α≧0. 1≧ρ≧0. ωn≧ωf≧ωm ・・
・Obtain [phase]. It is sufficient if ωm, ξm, ωn, ξn, α, and β are determined from [phase] for ωf and ξf of the real system, and these satisfy 0. Note that if α and β are made thicker, the influence of disturbance torque on the characteristic fluctuations of the actual system will be suppressed to a smaller extent.

In addition, in the device of this embodiment, even if Vc includes a saturation element as shown in FIG. 5, even if Vc suddenly fluctuates due to the application of TL, the proportional constant of Gv is large, so that Vc is quickly reduced.
Since the deviation between and Af becomes small and these state quantities fall within the proportional region of D, there is a characteristic that there is almost no need to be influenced by such saturation factors.

Furthermore, as shown in FIG. 6, the speed control system 100 in FIG. It is also suitable to use it in a control system that outputs, in addition to Vo, the control position Po relative to the command position Pc.

A feature of the present invention is that even in such a position control system, the behavior of the output Po is not affected by fluctuations in Ga and Gf and by TL.

Simulation results when the above-described operation of the present invention is applied to the position control shown in FIG. 6 are shown in FIGS. 7 to 10.

However, the parameters of the actual system are as follows.

Ta = 0.0005 sec, kt =
5.0 kgcm/AJ = 0.1 kgcm se
c” D = 0ωf = 100~rad/se
e, ξf = 0.005~0.01~0.02T
L = 20 kgcm Nominal system parameters are ωm = 100 rad/sec, ξm = 0
．． 01ωn = 200 rad/sec ξn
= 1.0α:β= 1.0 Tc = 0.001 sec, (however, 0.01 sec during the time
sec rectangular wave disturbance) Ka =
1000Asec”7m.

The command Vc is an isometric step input using Pc which is proportional to time, and its amplitude is 0.50 m/sec.

At this time, the behavior of Ua, Vo, and TL is shown in FIG.

In other words, Fig. 7 shows the case where there is no modeling error, that is, G
m= G f ((Ll f= 100 rad/s
ec, ξf=o, o1). The influence of the disturbance TL appears in Ua, but not in ■0.

Figure 8 shows the case where there is a modeling error, ωf=150
rad/sec ξf゛=0.01.

In addition, ξf = 0.02 in Fig. 9, and ξf = 0.02 in Fig. 10.
o, oos.

In both cases, the influence of disturbance TL does not appear on VO, and this V
o is extremely insensitive (has no effect) to variations in ξ.

Comparing Fig. 7 and Figs. 8 to 10, it is found that at the time of the rise of VO, ωm = 100 rad/see≠ωf =
The influence of 15Q rad/sec can be seen as only a very slight difference in the waveform.

There is a great advantage that Vo is not affected by modeling errors when ωm≧ωf. In other words, if Gf includes multiple vibrational characteristics and is expressed as
All of these (K=1.2.,.

! This is because it becomes possible to cancel the vibration element using only Gm.

As mentioned above, even if Gm defined for Gf includes a modeling error, the second output Vo is robustly defined as Gn, so in addition to the vibrational characteristics of Gf, which is the number of characters 2, Even when including a first-order lag characteristic (T s s + 1 ) having a small time constant T3, the robust characteristic defined by Gn of the present invention is maintained.

In addition, Gf is not an oscillatory characteristic but a non-oscillatory characteristic (ξfk
It goes without saying that even in the case of ≧1), the robust characteristics dominated by Gn of the present invention hold true.

[Effects of the Invention] As explained above, according to the robust speed and position control device of the present invention, the speeds of the first and second controlled means are detected, and predetermined arithmetic processing is performed on each of them. Since the velocity feedback information and the differentiated value are used as the acceleration feedback information, highly accurate velocity or position control is possible even when the second controlled means is vibratory. .

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of a physical system of a speed robust control device according to an embodiment of the present invention, FIG. 2 is a block diagram obtained by equivalently transforming the physical system of FIG. 1, and FIG. Figure 2 shows a simplified configuration of the speed loop when the gain ka is made sufficiently large, and Figure 4 shows the simplified characteristic configuration when the value of the gain 1Gvl is made sufficiently large in Figure 3. FIG. 5 is a block diagram when the first compensation means includes a saturation element; FIG. 6 is an explanatory diagram of a speed and position robust control device including the device shown in FIG. 1; Figure 7 is an explanatory diagram showing the simulation results when there is no modeling error of the device shown in Figure 6, and Figure 8 is an explanatory diagram showing the modeling error (ξf
Figure 9 is an explanatory diagram showing the simulation results for the case where = 0.01), and Figure 9 shows the modeling error (ξf
Figure 10 is an explanatory diagram showing the simulation results for the case where = 0.02), and the modeling error (ξ
FIG. 11 is an explanatory diagram showing the simulation results in the case of f-0,05), which is an explanatory diagram of the physical system of the conventional speed control device. 10.110...First compensation means 12.112...First controlled means 14.114...Second controlled means 116...Second compensation means 118...Multiplication means 120...Third means of compensation

Claims (7)

[Claims] (1) Has a physical reality corresponding to the energy form used, and has a transfer characteristic that defines its characteristics. a first controlled means which has a characteristic and generates a first output in a velocity dimension by inputting operation information in an acceleration dimension; and one or more transfer characteristics 1/Gfk (k=1, 2, .
), each denominator polynomial Gfk
a second controlled means which has an oscillatory characteristic of order 2 and has a natural angular frequency ωfk, and which inputs the first output and generates a second output in the velocity dimension; Negatively feeds back a speed dimension command and detection information of the first or second output to form a speed loop together with the first and second controlled means, and calculates the difference between the command and the feedback information. a first compensating means for generating operation information in the acceleration dimension based on the acceleration dimension; a second compensating means for transmitting and outputting the detection information of the first output according to the following formula [1]; )/Gm...[1] Gm: Polynomial having the same order as Gfk and a vibration characteristic root whose natural angular frequency ωm≦ωfk Gn: Having the same dimension as Gm and its natural angular frequency ωn is determined by ωn≧ωfk, and a polynomial α: a first constant governs the behavior of the output of the second controlled means; a multiplication means for multiplying the detection information of the second output by β (second constant) and outputting the result; an addition means for adding the outputs of the two compensating means and the multiplication means and outputting synthetic feedback information; and pseudo-differentiating the synthetic feedback information and feeding back the pseudo acceleration control information negatively to the path of the first compensating means to create an acceleration loop. and third compensation means for forming a compensating element in the acceleration loop, so that the composite relative order of the acceleration loop can be set to one of 0, -1, and -2, and the composite amplification degree of the compensation elements in the acceleration loop can be set to a desired value. the degree of amplification is determined so that disturbances applied to the acceleration loop and fluctuations in parameters of the first controlled means do not affect the first output, and the composite feedback information is feedback information in the velocity loop. Also used as the first constant α
and a second constant β are determined so as not to affect the second output. (2) A robust speed control device according to claim 1, wherein the first compensation means includes a saturation element. (3) A robust speed control device according to claim 1 or 2, characterized in that the constants α and β are set to 0≦α and β≦1. (4) A robust speed control device according to claim 1 or 2, characterized in that for the constants α and β, α=β. (5) In the device according to any one of claims 1 to 4,
A position loop is constructed using position dimension command information from the outside and position dimension feedback information obtained by integrating the detection information of the second output, and this command information and this position dimension feedback information are used. A robust speed and position control device, characterized in that a command for the speed dimension is generated using the difference between the speed and the position. (6) The device according to any one of claims 1 to 4,
A speed robust control device characterized in that the second controlled means has, in addition to the oscillatory characteristics, a relative order -3 characteristic in which a first-order lag characteristic having a small time constant is added. (7) A robust speed control device according to claim 6, characterized in that the second controlled means has a non-oscillatory characteristic and a relative order of -3.