JP3283662B2 - State estimation observer device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a state estimating observer, and more particularly to a state estimating observer used for a control device of various actuators.

[0002]

2. Description of the Related Art Conventionally, control devices for various actuators have been used.
A state estimation observer device is used for the
No. 2-10411). FIG. 4 shows such a conventional state.
Schematic configuration of first example A01 of state estimation observer device
FIG. 5 is a conventional state estimation observer device.
Block showing a schematic configuration in the second example A02 of the device
FIG. As shown in FIG. 4, a conventional first example device A01
Is from the controlled object 1 consisting of the actuator and its load
Output observation value y obtained by observing the output y of_SAnd control object 1
Input observation value u obtained by observing the control input u_SAnd based on
A configuration in which the state of the controlled object 1 is estimated by the server 2
It is used as a disturbance estimation observer as follows.
Move. That is, the observer 2 controls the target 1
The disturbance d acting on the actuator is determined by the speed of the actuator.
Component, acceleration component, and each observed value y of the control input_S, U
_SAnd the disturbance estimation value d_SCompensation based on
The control amount Δu is fed back to the actuator. One
This conventional device A01 is used in an actuator control system.
Estimation of all disturbance forces to be suppressed
And feedback, the actuator
Maintain predetermined control performance even when parameters fluctuate
It was measured. In addition, as shown in FIG.
In the second example device A02, the electric motor
If there is a delay in the current loop 0 of the machine,
Calculate control input u and input control input u to observer 2
Then, the speed of the motor is estimated. And estimate
The error between the set speed and the actual speed is multiplied by the gain and fed.
But the feedback value and the actual torque
The current value is also multiplied by the gain so that
The motor is controlled so that the actual speed matches the estimated speed.
That is, the control target 1 is intended to match the model.
(JP-A-3-93486). Note that FIG.
For convenience, the block diagram of the conventional example is equivalently converted.

[0003]

In the first prior art apparatus A01 as described above, the dead time, torsional vibration of the shaft, delay of the current loop, etc. caused by the sensor system, control calculation, control cycle, sample hold, etc. In the case where the influence of the parasitic element of the above is not considered at all, for example, when the dead time L exists, FIG.
If the convergence gain ω ₀ of the middle observer 2 is ω ₀ > π / 2L, it can be understood that the control system becomes unstable by the known Nyquist stability determination. In addition, the conventional second example apparatus A0
2 merely increases the current feedback gain with respect to the current loop delay as is apparent from FIG. Therefore, if the control target 1 has a parasitic element such as a torsional vibration of the shaft or a dead time, there is a possibility that the same problem as in the above-described first example apparatus A01 may occur. The present invention improves the state estimation observer device in order to solve such problems in the conventional technology, and constantly performs stable state estimation even for control systems including parasitic elements such as higher-order modes and dead time. It is an object to provide a state estimation observer device that can be performed.

[0004]

In order to achieve the above-mentioned object, the present invention provides a method based on an output observation value obtained by observing an output from a controlled object and an input observed value obtained by observing a control input to the controlled object. In a state estimation observer device for estimating a state of a control object, a state estimation observer device which exists in a first route from the control input to the output observation value and is used for state estimation.
From the control input to the input observation value, the product of the first parasitic element , which is a dynamic characteristic element other than the nominal mathematical model, and the dynamic characteristic of the first route, which is a dynamic characteristic element for correcting the output observation value. in dynamic characteristic elements that exist within the second root of
A certain second parasitic element and dynamic characteristics for correcting the input observation value
A dynamic characteristic determining unit that determines a bi-dynamic characteristic so that a product of the dynamic characteristic of the second route as an element is equal to the dynamic characteristic of the first route determined by the dynamic characteristic determining unit; A state estimating unit for estimating the state of the controlled object by using a value obtained by multiplying the characteristic and a value obtained by multiplying the input observation value by the dynamic characteristic of the second route determined by the dynamic characteristic determining unit. This is configured as a state estimation observer device characterized by the above. Further, in a state estimation observer device for estimating the state of the controlled object based on an output observed value obtained by observing an output from the controlled object and an input observed value obtained by observing a control input to the controlled object, y _D = N _Y × y _S and u _D = N _U × u _S provided that the above-mentioned state estimation of the controlled object is performed using the values y _D and u _D given by P _Y × N _Y = P _U × N _U. Is a state estimation observer device characterized by the following. Where y _S is the above output observation, u
_S is the input observation, P _Y is in the first route from the control input to the output observation , and is used for state estimation.
Dynamic characteristic elements other than the nominal mathematical model
A certain first parasitic element, N _Y , corrects the output observation
Dynamic characteristic of the first route is a dynamic characteristic elements, P _U and the second parasitic element is a dynamic characteristic elements that exist <br/> in the second route from the control input to the input observations, N _U is the above input
It is a dynamic characteristic of a second route that is a dynamic characteristic element for correcting an observation value . Further, there is provided a state estimation observer device characterized in that a filter unit for providing a filtering characteristic according to the first or second parasitic element to a result of the state estimation of the control target by the state estimation unit is provided.

Further, the filtering characteristic of the filter section is a product of the first parasitic element and the dynamic characteristic of the first route, or the product of the second parasitic element and the dynamic characteristic of the second route. This is a state estimation observer device that has an inverse characteristic that is inversely proportional to the product. Furthermore, the state estimation result x _E of the control object, FαP _T ^-1 or _{Fα (P O × N Y)} -1 _{However, P T = P Y × N} Y = P U × N U = P _This is a state estimation observer device characterized in that a new state estimation result is obtained by multiplying an inverse motion characteristic F given by _O × P _I × N _Y. Here, P _Y is the first parasitic element, N _Y is the dynamic characteristic of the first route, P _U is the second parasitic element, N _U
Is a dynamic characteristic of the second route, and P _O and P _I are partial parasitic elements when the first parasitic element is separated into the control input side and the output side. In addition, the first or second
Is a state estimation observer device in which the gain of the dynamic characteristic of the route is set to 1. Further, there is provided a state estimation observer device wherein the first or second parasitic element is 1. Moreover,
A state estimation observer device in which the first and / or second parasitic element is a higher-order mode. Further, there is provided a state estimation observer device in which the first and / or second parasitic element is a delay of a lower control loop. Further, the first and / or second parasitic element is a state estimation observer device due to a delay of a lower control loop.

[0006]

According to the present invention, when the state of the controlled object is estimated based on the output observed value obtained by observing the output from the controlled object and the input observed value obtained by observing the control input to the controlled object, A product of a first parasitic element present in a first route from a control input to the output observation value and a dynamic characteristic of the first route, and a second route from the control input to the input observation value Both dynamic characteristics are determined by the dynamic characteristic determination unit so that the product of the second parasitic element existing in the second path and the dynamic characteristic of the second route becomes equal. The value obtained by multiplying the output observation value by the dynamic characteristic of the first route determined by the dynamic characteristic determination unit is multiplied by the dynamic characteristic of the second route determined by the dynamic characteristic determination unit by the input observation value. The state estimation of the control object is performed by the state estimation unit using the calculated values. By performing the state estimation by giving the dynamic characteristic corresponding to the parasitic element in this way, it is possible to always perform a stable state estimation without causing the control system to become unstable due to the presence of the parasitic element as in the conventional example. . Furthermore, if a filter unit that provides a filtering characteristic according to the first or second parasitic element is provided in the state estimation result of the control target by the state estimating unit, the influence of each parasitic element can be reduced. ,
More stable state estimation can be performed. In this case, the product of the first parasitic element and the dynamic characteristic of the first route,
Alternatively, the influence of the parasitic element can be completely removed from the estimation result by giving the inverse characteristic that is inversely proportional to the product of the second parasitic element and the dynamic characteristic of the second route to the filtering characteristic of the filter unit. In addition, the first or second
When the gain of the dynamic characteristic of the route is set to 1, the observer estimation formula used for the above state estimation is simplified, and the load on the apparatus at the time of state estimation can be reduced. further,
When the first or second parasitic element is set to 1, the calculation using the equation used for determining the bi-dynamic characteristics is simplified, and the load on the device when determining the dynamic characteristics can be reduced. Further, the first and / or second parasitic element is set to a higher mode. Further, the first and / or second parasitic elements are dead time. Further, the first and / or second parasitic element is a delay of the lower control loop. Therefore, even when the present invention is applied to a control object having a higher-order mode such as a torsional vibration of a shaft, a dead time of a sensor system or a controller, or a delay of a lower control loop such as a current loop, the state can be always stably estimated. .

[0007]

Embodiments of the present invention will be described below with reference to the accompanying drawings to provide an understanding of the present invention. The following embodiments are examples embodying the present invention, and do not limit the technical scope of the present invention. Here, FIG. 1 is a block diagram showing a schematic configuration of a state estimation observer device A according to an embodiment of the present invention, FIG. 2 is a partial block diagram showing a state in which disturbance is applied to a control target, and FIG. It is a schematic diagram which shows an example. As shown in FIG. 1, the state estimation observer device A according to the present embodiment
The output observation value y _s obtained by observing the output y from
For that performs state estimation of the control object 1 based on the input observed value u _s a control input u observed to are the same as the conventional example. However, in this embodiment, the first parasitic elements P _I (S) and P _O (S) existing in the first route from the control input u to the output observation value y _s and the dynamic characteristics of the first route N _Y
And the product of the (S), the dynamic characteristic of the second parasitic element P _U (S) and the second route that is present in the second route from the control input u to the input observed value u _s N _{U (S)} Ryodo characteristic N _Y (S), so that the product is equal to and the dynamic characteristic determining unit 2 for determining the N _U (S), a first determined by the dynamic characteristic determination unit 2 to the output observed value y _s The value y multiplied by the dynamic characteristic N _Y (S) of the route
State estimation performed and _D, and the state estimation of the controlled object 1 by using the value u _D obtained by multiplying the dynamic characteristics N _U (S) of the second route which is determined by the dynamic characteristic determination unit 2 to the input observed value u _s A different point from the conventional example in that the device includes the unit 3. Furthermore, the state estimator 3
The state estimation result x _E of the controlled object 1 according to the first or second
May be provided with a filter section 4 which gives a new state estimation result by giving a filtering characteristic F (S) corresponding to the parasitic element of the above. This point also differs from the conventional example. In that case,
The filtering characteristic F (S) of the filter unit 4 is defined as the product of the first parasitic elements P _I (S), P _O (S) and the dynamic characteristic N _Y (S) of the first route, or the second parasitic element P (S). _Y (S) and the second
May be inversely proportional to the product of the route and the dynamic characteristic N _U (S).

Hereinafter, the basic principle and operation of the device A will be described. [Case of Basic Configuration] First, a case will be described in which the apparatus A includes only a control target 1, a dynamic characteristic determination unit 2, and a state estimation unit 3, which are basic configurations. That is, in the controlled object 1 in which a parasitic element such as a higher-order mode or a dead time exists in the input / output as shown in FIG. 1, the transfer function expression of the nominal mathematical model of the controlled object 1 is represented by P _N (S) or the state equation. The expression is represented by the following equation. dx (t) / dt = Ax (t) + Bu (t) (1) y (t) = Cx (t) (2) Next, an output observation value y _S obtained by observing the output value y and a control input Assuming that only the input observation value u _S that observes u can be observed, the total P _Y (S) of the first parasitic element from the control input u to the output observation value y _S is calculated on each of the control input side and the output side. It can be expressed by the product P _I (S) × P _O (S) of the partial parasitic elements. Here, the dynamic characteristics determining section 2 determines the dynamic characteristics N _Y (S) and N _U (S) of the first and second routes that satisfy the following expression. P _Y (S) N _Y (S) = P _U (S) N _U (S) (= P _T (S)) (3) Also, the dynamic characteristics N _Y (S) of the first and second routes And N
_U (S) and output observation value y _s and input observation value u _{s, respectively.}
The output value when the input preparative When y _D and u _D, the following equation is established. y _D (S) = N _Y (S) y _S (S) (4) u _D (S) = N _U (S) u _S (S) (5) Values y _D and u obtained here as input _D, the state estimation observer is a state estimation unit 3 estimates the estimated value x _E state x is given by the following equation. dz (t) / dt = Dz (t) + Ey D (t) + Hu D (t) ... (6) x E (t) = Pz (t) + Vy D (t) ... (7) where in the formula It is assumed that a matrix T exists such that the matrices D, E, H, P, and V satisfy the following equation. Here, I is a unit matrix, and D is an asymptotically stable matrix. TA-DT = EC, H = TB, PT + VC = 1 (8)

The convergence of the state estimation observer is determined by the matrix D
And the filtering characteristic given by the poles of the matrix D is given by F _D
(S), the obtained estimated value x _E (state estimation result)
Is represented by the following equation. _{x E (S) = F D} (S) P T (S) P I (S) -1 x (S) (= F D (S) P O (S) N Y (S) x (S)) ... (9) From the asymptotic stability of the matrix D, the filtering characteristic F _D (S)
Is stable and each parasitic element is stable.
Parasitics P _O (S) is stable. Further, from the stability of each parasitic element and the above equation (3), the dynamic characteristic N of the first route is obtained.
_{Since Y} (S) is also stable, the state estimation observer is affected by each parasitic element as shown in the above equation (9), but the state can always be stably estimated. On the other hand, in the conventional example, as described above, instability sometimes occurred. Next, the present method is applied to a control system to which a disturbance d is applied as shown in FIG. For example, the transfer function expression P _N (S) of the nominal mathematical model of the control system is given by the following equation. P _N (S) = J ⁻¹ / S ² (10) Here, assuming that d is an offset disturbance (the time differential value of the disturbance d is 0) and the disturbance estimation observer is configured by this method, the filtering characteristics F _D (S) is ω / (S +
ω), and the disturbance estimation observer is given by the following equation. _{d E (S) = ω /} (S + ω) × {N Y (S) J -1 S 2 y s (S) -N U (S) u S (S)} ... (11) However, S is Laplace And ω is the observer convergence gain. Here, considering that the disturbance estimated value d _E is compensated by negative feedback, the closed-loop transfer function from the disturbance d to the disturbance estimated value d _E is as follows. d _E (S) = P _O (S) N _Y (S) ω / (S + ω) × d (S) (12) Accordingly, the closed loop transfer function from the disturbance d to the output value y is represented by the following equation. y (S) = J ^-1 ｛I-ωP _O (S) N _Y (S) / (S + ω)｝ S ^-2 d (S) (13)

As described above, it is understood that the disturbance is stably estimated, and the disturbance is stably removed even when the feedback is performed. That is, for example, when a dead time L exists as a parasitic element, in the conventional example, the observer convergence gain ω is set to π / 2.
The fact that the control system is destabilized by Nyquist stability determination when the value is L or more has already been described. In contrast, the observer does not become unstable at all. [Case where Filter Unit 4 is Provided] Next, a case where the filter unit 4 is provided will be described. (6), (7) to the state estimation observer shown in equation adds the filtering characteristic F (S), dz (t ) / dt = Dz (t) + Ey D (t) + Ju D (t) ... (14) the use of _{x E (S) = F (} S) {Pz (S) + Vy D (S)} ... state estimation observer that estimates the state by (15), the estimated value x _E (S) the following Given by the formula. _{x E (S) = F (} S) F D (S) P O (S) N Y (S) x (S) ... (16) ( hereinafter simply referred to parasitics) parasitics gave dynamics P
_O (S) N _Y (S) (dynamic characteristic N _Y (S) of the first route
Is the first parasitic element P _I (S), the entire first parasitic element P
_Y (determined by S)) or the parasitic element P
By giving the filtering characteristic F (S) in accordance with _T (S), the influence of each parasitic element can be reduced, so that the state can be more stably estimated. further,
Parasitic element P _O (S) N _Y (S) or parasitic element P
_T (S) of the reversing characteristic filtering characteristic F (S) F _D
Be given the dynamic characteristics in filtering characteristic F (S) to give the (S), the estimated value x _E obtained is expressed by the following equation. x _E (S) = x (S) (17) x _E (S) = P _I (S) ^-1 x (S) (18) The estimated value x _E (S ) And state x
(S) is completely matched, or (1)
8) reversing characteristic P _I (s) ^-1 of the input when feeding back the estimated value x _E by is taking parasitics P _I (S) is canceled as the effect of completely parasitics it is possible to configure a feedback loop estimate x _E in the form to be removed.

This method is applied to a control system as shown in FIG.
If so, the disturbance estimation observer at this time is given by the following equation.
Can be d_E(S) = F (S) ｛N_Y(S) S^Twoy_S(S) -N_U(S) u_S(S)｝ (19) However, ω / (S +
ω). Now, the disturbance estimation value d_E
Consider compensating for
In other words, the disturbance estimate d_EClosed loop transfer function to
The numbers are as follows: d_E(S) = F (S) P_O(S) N_Y(S) d (S) (20) The closed loop transfer function from the disturbance d to the output value y is
It is represented by y (S) = J^-1｛IF (S) P_T(S)｝ S^-2d (S) (21) From this, it can be seen that the disturbance d is stably estimated.
You. Here, the inverse characteristic of the parasitic element is given to F (S).
The disturbance estimate d_EAnd the disturbance d
Can be The effect of the parasitic element on the output value y
Since it can be completely eliminated, the closed loop from the disturbance d to the output value y
The transfer function becomes 0 and can be completely zeroed.
Noh. [Simplified configuration] Next, when the above configuration is simplified
Is described. That is, in the above configuration, the control input u
From output observations y_SOf the first parasitic element to
_YAs compared with (S), the input observation value u from the control input u_SFirst to
2 parasitic elements P _U(S) is small or the second
Parasitic element P_uParasitic element P compared to (S)_Y(S) is fine
If small, the dynamic characteristic N of the first route_Y(S) or
Dynamic characteristic N of the second route_UBy setting (S) to 1
Even if a part of the dynamic characteristics is omitted, the conditional expression (3) is
Satisfied, the state estimation observer is expressed by the following equation.
Become. u_D(S) = N_U(S) u_S(S) (5) dz (t) / dt = Dz (t) + Ey (t) + Ju_D(T) ... (22) x_E(S) = F (S) {Pz (S) + Vy (S)} (23) or y_D(S) = N_Y(S) y_S(S) (4) dz (t) / dt = Dz (t) + Ey_D(T) + Ju (t) (24) x_E(S) = F (S) ｛Pz (S) + Vy_D(S)｝ (25) Since the estimation formula can be simplified,
The load on the device can be reduced. Further, the first parasitic element P
₁(S), P₀(S) or the second parasitic element P_U(S)
If 1, that is, if some of the parasitic elements can be omitted
Simplifies the calculation by the conditional expression (3).
Therefore, the load on the device when determining the dynamic characteristics can be reduced.

[Specific Application of Parasitic Element] Next, a specific application example of the parasitic element will be described. As shown in FIG. 3, in a motor drive system in which the control target 1 has a torsional vibration of the shaft as a parasitic element, the motor rotational speed θ _M is output y, the load-side inertia is J _L , and the motor inertia is J _M
And a model that does not consider torsional vibration as in the following equation is a nominal constraint model. P _N (S) = (J _M + J _L ) ⁻¹ / S ² (26) At this time, the first parasitic element P _O (S) is represented by P _O (S) = (S ² J _L + K) / ^{_{(S 2 J L J M /}} (J M + J L) + K) ... using the state estimation observer obtained as (27), it is possible to perform state estimation in consideration of the torsional vibration. Alternatively, when the rotation angle θ _L on the load side is selected as the output y, the first parasitic element P _O (S) is represented by P _O (S) = K / (S ² J _L + J _M / J _M + J _L ) + K) (28) The state estimation in consideration of the torsional vibration can be performed by using the state estimation observer obtained as follows. Where K
Is the spring constant of the shaft. Next, when the dead time of the sensor system or the controller is a parasitic element of the control target 1,
The dead time L _I input, L _O a dead time of the output, if the dead time of the input observations u _s and L _U from the control input u,
The first parasitic element P _I (S), P _O (S), and the second parasitic element P _U (S) can be represented by the following equations, respectively. P _I (S) = e ^−Lis , P _O (S) = e ^−Los , P _U (S) = e ^−Lus (29) The state in which the dead time is taken into account using the state estimation observer thus obtained. An estimate can be made. further,
When the control target 1 is a motor drive system having a current loop delay as a parasitic element, assuming that the time constant of the current loop is T, the first parasitic element P _I (S) and the second parasitic element P
_U (S) can be expressed as the following equation. P _I (S) = 1 / (TS + 1), P _U (S) = 1 / (TS + 1) (30) State estimation taking into account the current loop delay can be performed using the estimation observer obtained thereby. .

When there are a plurality of parasitic elements, each of the parasitic elements can be represented by a product of the above parasitic elements. For example, when the dead time L _U and the current loop delay T exist simultaneously, the second parasitic element P
_U (S) is represented by the following equation. P _U (S) = e− ^Lus / (TS + 1) (31) Using the estimated observer obtained in this way, it is possible to perform state estimation in consideration of both the dead time and the delay of the current loop. Other combinations are possible as well. The following operations can be performed by using the state estimation observer device A including the dynamic characteristic determination unit 2 and the state estimation unit 3 that embody each state estimation observer as described above. When a state estimation observer device is configured in this way for a controlled object in which a parasitic element such as a higher-order mode or a dead time exists in the input / output, in the conventional example, if the observer convergence gain is increased, the control system is delayed by the delay of the parasitic element. While the device may become unstable, the present device can improve the stability. In particular, even when disturbance is estimated and fed back as in the conventional example, the control system does not become unstable in this device. However, the estimated value is affected by parasitic elements. By providing a filtering characteristic according to the parasitic element, the state can be more stably estimated. In addition, by giving the inverse characteristic of the parasitic element to the filtering characteristic, the influence of the parasitic element can be completely removed from the estimated value. By partially omitting the dynamic characteristics, the observer estimation formula can be simplified, and the load on the device during state estimation can be reduced. If some of the parasitic elements can be omitted, the calculation using the equation for determining the dynamic characteristics is simplified, and the load on the device when determining the dynamic characteristics can be reduced. Even if the above device is applied to a controlled object that has a higher mode such as torsional vibration, or a dead time of a sensor system or a controller, or a delay of a lower control loop such as a current loop,
State estimation can be performed stably. As a result, it is possible to obtain a state estimation observer device that can always perform stable state estimation even for a control system including a parasitic element such as a higher-order mode or a dead time.

[0014]

Since the state estimation observer device according to the present invention is configured as described above, the state estimation observer device can be used for a controlled object having a parasitic element such as a higher-order mode or a dead time at the input / output. In the conventional example, if the observer convergence gain is increased in the conventional example, the control system may become unstable due to the delay of the parasitic element. On the other hand, the present device can improve the stability. In particular, even when disturbance is estimated and fed back as in the conventional example, the control system does not become unstable in this device. However, the estimated value is affected by parasitic elements. By providing a filtering characteristic according to the parasitic element, the state can be more stably estimated. In addition, by giving the inverse characteristic of the parasitic element to the filtering characteristic, the influence of the parasitic element can be completely removed from the estimated value. By partially omitting the dynamic characteristics, the observer estimation formula can be simplified, and the load on the device during state estimation can be reduced. If some of the parasitic elements can be omitted, the calculation using the equation for determining the dynamic characteristics is simplified, and the load on the device when determining the dynamic characteristics can be reduced. Even if the above device is applied to a controlled object that has a higher mode such as torsional vibration, or a dead time of a sensor system or a controller, or a delay of a lower control loop such as a current loop,
State estimation can be performed stably. As a result, it is possible to obtain a state estimation observer device that can always perform stable state estimation even for a control system including a parasitic element such as a higher-order mode or a dead time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic device of a state estimation observer device A according to one embodiment of the present invention.

FIG. 2 is a partial block diagram showing a state where disturbance is applied to a control target.

FIG. 3 is a schematic diagram illustrating an example of a control target.

FIG. 4 shows a first example A0 of a conventional state estimation observer device.
FIG. 2 is a block diagram showing a schematic configuration in FIG.

FIG. 5 shows a second example A0 of the conventional state estimation observer device.
2 is an equivalent block diagram showing a schematic configuration in FIG.

[Explanation of symbols]

A ... state estimation observer apparatus 1 ... controlled target 2 ... dynamic characteristic determining unit 3 ... state estimation unit 4 ... filter section _{P I (S), P O} (S) ... first parasitic element P _Y (S) ... second N _Y (S): Dynamic characteristic of the first route N _U (S): Dynamic characteristic of the second route F (S): Filtering characteristic u: Control input y: Output y _S : Output observation value u _S ... input observed value y _D, (corresponding to the values used in the state estimation) u _D ... value x _E ... state estimation result

Claims (10)

(57) [Claims] 1. A state estimation observer device for estimating a state of a control target based on an output observation value obtained by observing an output from the control target and an input observation value obtained by observing a control input to the control target, Nominal mathematics of the controlled object that exists in the first route from the input to the output observation and is used for state estimation
Out the first parasitic element is a dynamic characteristic elements other than model and the
And the product of the dynamic characteristics of the first route is a dynamic characteristic element for correcting the force observations, from the control input second is a dynamic characteristic elements that exist in the second route to the input observations The second is a parasitic element and a dynamic characteristic element for correcting the input observation value.
A dynamic characteristic determining unit for determining both dynamic characteristics so that a product of the first and second routes is equal to a dynamic characteristic of the first route; and a value obtained by multiplying the output observation value by the dynamic characteristic of the first route determined by the dynamic characteristic determining unit. And a state estimator for estimating the state of the controlled object using a value obtained by multiplying the input observation value by the dynamic characteristic of the second route determined by the dynamic characteristic determiner. State estimation observer device. 2. A state estimation observer device for estimating a state of a control target based on an output observation value obtained by observing an output from the control target and an input observation value obtained by observing a control input to the control target, wherein y _D = N _Y × y _S and u _D = N _U × u _S where the state estimation of the control object is performed using the values y _D and u _D given by P _Y × N _Y = P _U × N _U. What is done is a state estimation observer device. Here, y _S is the output observation value, u _S is the input observation value, and P _Y is a control that exists in the first route from the control input to the output observation value and is used for state estimation.
The first parasitic element, N _Y , which is a dynamic characteristic element other than the nominal mathematical model of interest, is a dynamic characteristic that corrects the output observation value.
Dynamic characteristic of the first route is an element, the dynamic P _U is that exists in the second route from the control input to the input observations
The second parasitic element, N _U , which is a characteristic element, is the above input observation value.
Is the dynamic characteristic of the second route, which is the dynamic characteristic element for correcting . 3. A filter unit for providing a filtering characteristic according to the first or second parasitic element to a state estimation result of the control target by the state estimation unit. State estimation observer device as described. 4. The filtering characteristic of the filter unit is a product of the first parasitic element and the dynamic characteristic of the first route, or a product of the second parasitic element and the dynamic characteristic of the second route. 4. The state estimation observer device according to claim 3, wherein the state estimation observer device has an inverse characteristic inversely proportional to the following equation. 5. A state estimation result of the controlled object x _E, FαPT ^-1 or Fα (PO × NY) ^-1 _{However, P T = P Y × N} Y = P U × N U = P O 3. The state estimation observer device according to claim 1, wherein a new state estimation result is obtained by multiplying the inverse motion characteristic F given by * P _I * N _Y. Here, P _Y is the first parasitic element,
N _Y is the dynamic characteristic of the first route, P _U is the second parasitic element, N _U is the dynamic characteristic of the second route, P _O , P _I
Is a partial parasitic element when the first parasitic element is separated into the control input side and the output side. 6. The state estimation observer device according to claim 1, wherein the gain of the dynamic characteristic of the first or second route is set to 1. 7. The state estimation observer device according to claim 1, wherein the first or second parasitic element is 1. 8. The state estimation observer device according to claim 1, wherein said first and / or second parasitic element is a higher-order mode. 9. The state estimation observer according to claim 1, wherein the first and / or second parasitic element is a dead time. 10. The state estimation observer device according to claim 1, wherein said first and / or second parasitic element is a delay of a lower control loop.