JP3305392B2 - Positioning method and device for operating device in automobile, etc. - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention own using a control loop
A method for determining the position of an operating device of a moving vehicle, wherein a control deviation between a set value and an actual value is determined and supplied to a first controller as an input amount, wherein the first control The control device is configured as a conventional control device having at least one of the control characteristics P, I, and D, and relates to a method and apparatus in which the first control variable is delivered by the first control device.

[0002]

2. Description of the Related Art A method is known from DE 3731 984A for the adaptive control of an electromagnetic drive with friction. In this method, the time course of the magnitude (absolute value) of the frictional force is detected by a model-assisted non-linear observer (observer). Subsequently, the polarity of the frictional force detected by another method is superimposed on the absolute value (magnitude). The following means are connected in parallel to the controller for compensating the frictional force thus detected, that is, the means whose output signal is superimposed on the output signal of the controller are connected in parallel. The signal thus generated is supplied to a control section.

A disadvantage of the above method is that in order to obtain good control characteristics, a model must be found which represents the control system as accurately as possible. The better the control system is represented (represented), the more extensive, large and complex the models used for the representation are usually. This makes it difficult to transfer the control method to another system, and to perform error checking (investigation) and error removal.

[0004] DE 40 12 577 C1 discloses a control system for an operating device with friction. In this control system, a two-point controller is connected downstream of the position controller. The hysteresis width of the two-point controller can be adjusted depending on the operating conditions. This arrangement reduces the effects of friction to improve the dynamic characteristics of the system. A disadvantage of this control system is that the operating device mechanism always oscillates about the set position under the control of the two-point controller. This results in increased wear of the operating mechanism and energy consumption. Furthermore, it is not possible to permanently limit the control deviation to a small value, and a time-varying control deviation between zero and half the hysteresis width remains.

[0005] A similar control device is known from DE 320 78 63 A1. Here, the nonlinearity controller has two points-
The controller is connected downstream. Using the limiting stage limits the switching frequency of the two-point controller.

From US Pat. No. 5,500,133 a vehicle speed control system using fuzzy logic is known.
In that case, the actual value of the speed is detected and the set value is set taking into account the driver requirements. Running state ( eg, gear position)
, An appropriate membership function is selected. Based on the membership function thus found, the setting of the traveling speed, and the actual value, an operation amount for controlling the operating device is obtained by applying a fuzzy control rule.

A disadvantage of the method and apparatus described above is that optimal control cannot be performed under any operating conditions.

[0008]

SUMMARY OF THE INVENTION It is an object of the present invention to provide an accurate and quick positioning of an operating device in a motor vehicle. In this case, the deviation between the set value and the actual value is made as small as possible under as many operating conditions as possible, and the control characteristics are stabilized even in extreme cases. In particular, it seeks to reliably remove even small control deviations despite the presence of friction or other non-linear factors.

[0009]

The above object is achieved by the features of claim 1 and further by the advantageous embodiments and developments described below.

The present invention has the advantage that the disadvantages of the prior art described above are obviated.

Particularly advantageously, the advantages of fast transient (transition) characteristics and good noise characteristics are obtained. Precise control of the set value can be achieved by friction compensation. Another advantage is the good robustness (stability) of the controller in extreme situations. Depending on the adaptation method of the control method of the present invention, adaptation to a specific application area can be performed without any problem without involvement in the structure of the controller and exercise of influence.

Advantageous configurations and developments of the invention are described below. A normalized quantity associated with the present disclosure;
Both denormalized quantities are used. For discrimination, the symbols for the non-normalized amounts are underlined, respectively.

[0013]

BRIEF DESCRIPTION OF THE DRAWINGS FIG.

The construction and operation of the present invention will be described for use in connection with a fuel metering device in a diesel internal combustion engine. In this embodiment, a signal for controlling an electromagnetic operating mechanism of an injection pump for performing fuel metering is transmitted.

FIG. 1 shows a basic connection configuration of a control circuit according to the present invention. The block designated by the reference numeral 10 sends the setpoint W of the position of the electromagnetic operating mechanism of the fuel metering device as a first input signal to the connection point 11. The actual value y of the adjusting device mechanism is supplied to the connection point 11 as another input signal. The actual value y is subtracted from the set value W at the connection point 11.
The result of the above calculation operation is called a control deviation e, and the controller 12
Supplied to This controller 12 thereby controls the manipulated variable U 1
And transmits it to another connection point 13. Bond points 13 as another input signal receives the manipulated variable U 2 obtained by the fuzzy controller 14, as the output signal U, and sends the sum of U 1 and U 2. The fuzzy controller 14 is connected in parallel with the controller 12 and has two inputs. The control deviation e determined by the connection point 11 is added to the first input side,
The time differential value e 'of the control deviation e is applied to the input side, and the time differential value is obtained from the control deviation e by the differentiation stage 15.

The output signal U at the node 13 is supplied to an electromagnetic operating mechanism 16 of the fuel metering device, which meteres the fuel accordingly. The actual position y of the operating device mechanism is determined by a sensor 17 mounted on the operating device mechanism 16 and returned to the connection point 11 via the supply line 18 and responded.

The controller 12 includes at least one of the elements P, I, and D according to a use area. The above-mentioned at least one element is designed as follows, that is, designed to be able to perform the optimal control under the monitoring of the friction of the operating device mechanism 16. The effect of friction is compensated by the fuzzy controller 14. The fuzzy controller 14 is designed in such a way that it has a sufficient effect on the control process, especially with small control deviations (where frictional forces have a particularly detrimental effect), that is, It is designed to give a high contributing component amount to U. This allows for permanent control deviations (which is
Which usually occurs in the situation described above). In the case of very large control deviations, the effect of the fuzzy controller 14 is slight and the control characteristics are mainly determined by the controller 12.

In another embodiment, the fuzzy controller 14 can be additionally connected to the controller 12 or selectively turned on with a small control deviation e . The additional connection or the selective closing connection of the fuzzy controller 14 is performed as follows.
That is, it is performed when the control deviation e falls below a predetermined value.
It will be released again when the value is exceeded. In this case, the switching process is performed only when the switching condition exists at least for a predetermined time interval.

In order to determine the defined (set) position of the electromagnetic operating mechanism 16 by means of the block 10, techniques known from the prior art can be used. Usually, the setting (prescribed) position of the operating device mechanism 16 is determined by using a characteristic curve or a characteristic field in consideration of a driver's request from various operation parameters.

FIG. 2 shows a fuzzy controller 14 internally comprising three series-connected functional units 20, 22, 24.
FIG. The above three functional units are usually a fuzzy conversion circuit 20, fuzzy listening (fuzzy inference).
Circuit 22, Defuzzification
Circuit 24.

The first functional unit 20 receives, as an input signal,
The control deviation e and its derivative e 'are received. By the normalization, a normalized control deviation e and a normalized differential value e ′ of the control deviation are obtained. The quantities e and e 'are associated with two groups of membership or membership functions μe, μe' (shown graphically in FIG. 4). Membership (attribution)
The actual values for e to e 'in the function are stored and subsequently the second functional unit 22
A membership function is transmitted to.

In the second function unit 22, a fuzzy control rule is applied to the membership (assignment) function prepared from the first function unit 20, and as a result, an assignment function μu for the operation amount is obtained. A graphical representation of the above-described computational operation, which is commonly referred to as fuzzy resonating or inference, is shown in FIG. The membership function μu is transmitted to a third functional unit 24, which generates a normalized manipulated variable U by forming an average value. From the normalized operating amount U manipulated variable U 2 which is adapted to the system used in the fuzzy controller 14 is determined and outputted at the output side of the fuzzy controller 14.

FIG. 3 shows another embodiment of the fuzzy controller 14 which additionally has means for adapting the controller. Said measures may be provided only during the phase of development of the controller, or-if constant adaptation is desired-more or partly during the use of the controller in question. . The core of the adaptive fuzzy controller shown here is shown in FIG.
Each of the functional units, that is, the fuzzy circuit 20, the fuzzy resonating or inferring circuits 31 and 32, and the defuzzification circuit (defuzzification circuit) 33. The arrangement and operation of such a functional unit has already been described with reference to FIG. The fuzzy resonating unit in particular differs from that shown in FIG. 2 in that the fuzzy control rules have been removed from the static control base circuit 32 to facilitate adaptation. . The control rules contained in the control base circuit 32 are applied at block 31 to the membership function.

In the adaptive fuzzy controller shown in FIG. 3, two stages of the adaptation are provided. In the first stage, the characteristics of the controller are adapted to the field of use. For this purpose, the static control base circuit 32 is correspondingly controlled by the expert system module 34. Above all, the theory of structurally variable systems is used for fuzzy control. The expert system module 34 contains expertise that applies to all intended fields of use.

In the second adaptation stage, adaptation to a special control section is performed by selecting appropriate parameters for adding the input and output signals of the adaptation fuzzy controller. Control deviation
e ′ is multiplied by the parameter K ₁ in a block 35, and the thus generated normalized control deviation e ′ is transmitted to a block 30. Time differential e of the control deviation 'is multiplied by the parameter K ₂ at block 36, the time derivative e of such a control deviation is generated' is transmitted to the block 30. Normalized operating amount U which is output from the block 33 is multiplied by the parameter K ₃ at block 37, the operation amount U 2 produced in this way forms an output signal of the adaptive fuzzy controller.

The parameters K ₁ , K ₂ , K ₃ are determined using the adaptation module 38. The data required for the adaptation can be determined manually using the expert system module 34 or automatically by the weighting module 3.
9 is obtained by weighting the prediction with respect to the control deviation. The prediction of the control deviation e (K + i) at the time K + 1 is based on the knowledge of the set value W (K + i) and the actual value Y (K).
It is implemented in.

The value of the case following control deviation e to determine the parameters K ₁ is calculated, i.e. the value of the control deviation e which is to follow the set value W for the controller is no longer possible is obtained from the value . The control deviation e _o obtained in this way should form just half of the maximum normalized control deviation e _max after multiplication with the parameter K ₁ .

[0028] To determine the _{K 1 = e max / 2 e} o K 3 static friction F _S ⁺ and F _S ^- it requires. Parameter K ₃ is chosen designed as follows, i.e., the force for friction compensation is at least F _S ⁺ when e _max / 2, F _S ^- set is designed to be equal to the maximum of.

[0029] ^{_{K 3 ≧ max (2F S +}} / e max '2F S - / e max) parameter K ₂ are chosen as follows, i.e., selected so that the dynamic characteristics of the fuzzy controller is optimized Is done. This can be done manually based on the expert system module 34 or automatically by the above-mentioned weighting of the predicted control deviation e (K + i).

FIG. 4 shows the membership function. In order to define the membership function, a plurality of categories are respectively associated with the normalized control deviation e, the normalized time derivative e ′ of the control deviation, and the value range of the normalized manipulated variable U. Each category quantitatively represents the position state of the sub-region in the value region. The strength (degree of membership) of the degree of association of each value of one value region with one category is determined and set by the membership function. For example, the category "negative small" (NS)
Membership function μe represented by the strength of the degree of association or degree of association of the normalized control deviation to
(FIG. 4-a).

FIG. 4 shows the category “negative bi”.
g ”(NB),“ negativesmall ”(N
S), “positive small” (PS) and “positive big” (PB) membership functions. The abscissa shows the normalized amount e (FIG.
a), e '(FIG. 4-b) and U (FIG. 4-c) are respectively converted from the minimum values e _min , e' _min to U _min to the maximum values e _max , e '.
_max to U _max and the membership functions μe (FIG. 4-a), μe ′ (FIG.
b), μu (FIG. 4-c) are shown for various categories. The membership functions take on values between 0 and 1, respectively. A value of zero indicates that there is no membership for the considered category, and a value of one indicates perfect membership. In the embodiment described here, the membership functions of the normalized control deviation e (FIG. 4-a) and the normalized time derivative e ′ (FIG. 4-b) of the control deviation are the same for the same category. Basically, different membership functions can be chosen for each of these quantities. Depending on the application, it is also possible to select a function curve different from that shown in FIG.

FIG. 5 shows a normalized control deviation e and a phase plane for the normalized time derivative e 'of the control deviation.
Depending on this phase plane, it is possible to derive a fuzzy control rule for obtaining the normalized manipulated variable U from the normalized control deviation e and the normalized time derivative e 'of the control deviation particularly effectively and clearly. The abscissa shows the normalized control deviation e,
The ordinate shows the normalized time derivative e 'of the control deviation. The phase plane is subdivided into a plurality of regions indicated by characters A to H. A straight line “e + e ′ = 0” extends from the upper left to the lower right through the diagonal line of the phase plane.
The fuzzy control rule is generated as follows. That is, the normalized operation amount U is made to correspond to each one region of the phase plane or a combination of a plurality of regions.

FIG. 6 is a table showing the correspondence between each area of the phase plane (left column) and the normalized manipulated variable U (center column). From the above correspondence, a fuzzy control rule (right column) is obtained as follows: the normalized control deviation e and the logical combination of each category with respect to the normalized time derivative e ′ of the control deviation are used to calculate the phase plane. The region is represented.

FIG. 7 shows a list of the fuzzy control rules (R1 to R8). A set of control rules for two regions of the phase plane is defined, ie, “e ′ ≧
-E "(R1 to R4) and"e'≤-e"(R5 to R
8) shows a set of control rules for the two areas. Each fuzzy control rule is introduced by the word "IF", which is followed by a premise (antecedent). The premise part (antecedent part) is composed of one category display information or a plurality of category display information combined by a logical operator (logic operator) with respect to the normalized control deviation e. The amount of each category display information is 2
Represents the category between two abbreviations, for example, N
The meaning of EB is as follows: the normalized control deviation e is "negative big". The conclusion part is the category display information for the normalized manipulated variable, for example, “N
UB ".

FIG. 8 shows an example value e _o of the normalized control deviation e, and a normalized differential value e ′ of the control deviation.
Four fuzzy control rules (R1 to R4) defined for the phase plane region “e ′ ≧ −e” are applied.
The upper part of FIG. 8 shows the membership functions μe and μ of FIG.
e ′ and μu are shown. Below this, an example of application of the fuzzy control rules R1 to R4 is shown. At that time, the application of each individual control rule can be taken from left to right, in other words, the results of the control deviations are respectively shown on the right side of FIG. The superposition of the results of the application of the control rules is performed from top to bottom, and the final result shown in the lower right of FIG. 8 is obtained.

The application of a certain fuzzy control rule proceeds as follows: First, the category indication information (for example, R3: PE'S AND NES) of the premise is evaluated as follows: At the positions (local) of the vertical _lines indicated by the example values e _o to e ′ _o , the values of the membership functions μe to μe ′ are changed to the quantities e to e ′ represented by the category display information for each category display information. (Spot). When the premise consists of a plurality of category display information, the maximum function value of the category display information combined by "OR" or the minimum function value of the category display information combined by "AND" is sequentially selected. The function value thus obtained is diverted to the category display information (for example, R3: PUB) of the conclusion part (the corresponding membership function μu of the normalized manipulated variable U is cut at the height of the function value) As a result, the area indicated by hatching in FIG. 8 is obtained. Control rule R
In the case of 1, R2, the cut is made at the zero-height line, and as a result, there is no remaining area.

The above processing method is implemented for each control rule, and subsequently (integration) of all the membership functions μu (areas indicated by hatching) thus obtained of the normalized manipulated variable U. An aggregate quantity is formed. The numerical value U _o for the normalized operation amount is a membership function (area indicated by hatching) obtained by the (integrated) set operation operation of the normalized operation amount U (shown at the lower right).
By averaging. U2 manipulated variable U2 by adaptation of the control section by using the parameter K ₃ is obtained from the operation amount U2 is superimposed on the operation amount of the conventional controller
(See FIG. 1).

In addition to the fields of use described here, the invention can also be used to advantage in motor vehicles for controlling the throttle control of a travel mechanism or the electromagnetic or hydraulic control of a front or rear wheel control. Application examples other than automobiles, for example, application in machine tools or robots are also possible.

[0039]

According to the present invention, it is possible to accurately and quickly determine the position of the operating device in a vehicle, and to minimize the deviation between the set value and the actual value under as many operating conditions as possible. Further, there is an effect that the control characteristics are stabilized even in extreme cases. In particular, high-speed transient characteristics and good noise characteristics can be obtained.

[Brief description of the drawings]

FIG. 1 is a basic configuration diagram of a control circuit of the present invention having a fuzzy controller.

FIG. 2 is a schematic structural diagram of a simple fuzzy controller.

FIG. 3 is a schematic diagram of a configuration of an adaptive fuzzy controller.

FIG. 4 is a characteristic diagram of a membership function with respect to a normalized control deviation e, a normalized time differential value e ′ of the control deviation, and a normalized manipulated variable U;

FIG. 5 is a phase plan view with respect to a normalized deviation e and its normalized time derivative e ′.

FIG. 6 is a diagram illustrating a table for associating each area of a phase plane with a category of a normalized manipulated variable U;

FIG. 7 is an explanatory diagram of listing up a fuzzy control rule.

FIG. 8 is a graphic display diagram of an application example of a fuzzy control rule.

[Explanation of symbols]

 11, 13 connection point 12 controller 14 fuzzy controller 15 differentiating stage 16 electromagnetic operating device mechanism 17 sensor 18 supply line

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G05D 3/00-3/20 G05B 11/00-13/04

Claims (12)

(57) [Claims] An operating device (1) for an automobile using a control loop.
A method of performing position defined in 6), set value (W) and the actual value (Y) and <br/> determined Me a control deviation (e) between, the control deviation (e), the first controller supplied as an input variable (12), said first controller (12), the control characteristics (P, I, D) is configured as a conventional controller having at least one of, and the first operation sends the amount (u1), the first controller second controller (2) (14) is connected in parallel
Are, the second controller (14) is constructed as a fuzzy controller, the second controller (14), the time differential value of the additionally the control deviation to the control deviation (e) (e ') is supplied as an input quantity, the second controller (14), second operation amount (u2) sent summarizes the first and second operation amount on the total operation amount, the fuzzy control (14) to construct the fuzzy
Controller (14), friction force has no disturbing effect
When the control deviation is small, the manipulated variable with a large absolute value (u2)
At the time of an extremely large control deviation,
The operation amount (u2) of the fuzzy controller (14) is
A method for determining the position of an operating device of an automobile using a control loop , characterized in that it is made smaller than at the time of a small control deviation . 2. The parallel connection of the two controllers (12, 14).
At time , the conventional controller (12) and the fuzzy controller (1
Operation amount obtained by 4) (U 1, U 2 ) are configured to overlap additively method of claim 1, wherein. 3. The fuzzy controller (14) is additionally connected or selectively switched on when the control deviation ( e ) falls below a selectable value for a selectable time interval. 3. The method according to 1 or 2. 4. The method as claimed in claim 1, wherein the fuzzy controller is adapted to the area of use by varying the fuzzy control rules. 5. An input and output signal of the fuzzy controller (14) is multiplied by a parameter, and the fuzzy controller (14) is adapted to a control section by manually or automatically varying the parameter. Item 5. The method according to any one of Items 1 to 4. 6. The value ranges of the normalized control deviation (e), the normalized time derivative (e ') of the control deviation, and the normalized manipulated variable (U) are respectively associated with a plurality of categories. -Each category qualitatively represents the position of a sub-region within the value domain;-one membership function is created for each category, and the membership function assigns the value domain to the category. The method according to any one of claims 1 to 5, wherein the degree of belonging or the degree of belonging of each value is set and fixed. 7. The membership function of the control deviation (e) of the normalization and the normalized time derivative (e ') has the following characteristic course for the following various categories: -For the category "negative big", the normalized control deviation (e) or the time derivative of the control deviation (e ')
_Exhibits a characteristic course that drops from 1 to 0 when taking values from the minimum value (e _min to e ′ _min ) to 1, and for the category “negative small”, the normalized control deviation (e ) Or when the normalized time derivative (e ′) of the control deviation exhibits a characteristic course in which the operation progresses from the minimum value (e _min to e ′ _min ) to 0, the rise is made from zero to 1; In the case of “positive small”, the normalized control deviation (e) or the normalized time differential value (e ′) of the control deviation is from 0 to the maximum value (e _max to e ′).
_max )), so as to exhibit a characteristic course that elapses up to 1), from 1 to zero, and for a category “positive big”, a normalized control deviation (e) or a normalized time of the control deviation The differential value (e ') changes from zero to the maximum value (e _max to e'
_max ), from 0 to 1
7. The method according to claim 6, wherein the pressure is increased. 8. The membership function (μu) of the normalized manipulated variable (U) has the following characteristic course for the following various categories: the case of the category “negative big” On the other hand, when the normalized manipulated variable (U) exhibits a characteristic course in which the operation progresses from the minimum value (U _min ) to 0, first, it rises from 0 to 1,
Then, it falls back to 0 again, when the normalized operation amount (U) exhibits a characteristic course in which the operation progresses from half of the minimum value (U _min ) to 0 for the category “negative small”. From zero to one
The normalized manipulated variable (U) is from 0 to _{の of the} maximum value (U _max ) for the category “positive small”
1 to 0 when exhibiting a characteristic course in which the operation progresses to
-For the category "positive big", the normalized manipulated variable (U) rises from 0 to 1 when exhibiting a characteristic course in which the operation progresses from 0 to the maximum value ( _Umax ). 7. The method of claim 6, wherein the method comprises: 9. A normalization operation amount (U) is obtained by a fuzzy controller (14). For this purpose, a fuzzy control rule is applied to the normalization control deviation (e) and the time derivative of the control deviation. 9. The method as claimed in claim 1, wherein the results of the application of the control rules are subsequently superimposed by averaging. 10. If the normalized differential value (e ') of the control deviation is greater than the negative value of the normalized control deviation (e), the first set of four fuzzy control rules is applied; 1 According to the content of the fuzzy control rule (R1), the normalized control deviation (e) is "positive small".
And when another condition is satisfied, the normalized manipulated variable (U) is set to “positive big”, and the other condition is a normalized time differential value (e ′) of the control deviation. ) Is "negativesmall" or "p
The first fuzzy control rule (R2) defines that the control deviation (e) is equal to "positive small" or "poise big".
When the condition is “sit big” and another condition is satisfied, the normalized manipulated variable (U) is “positive big”.
e small ”, and the other condition is that the normalized time derivative (e ′) of the control deviation is“ positive ”.
"ve big" or "normalized control deviation (e) is" po
the third time fuzzy control rule (R3) indicates that the normalized time differential value (e ′) of the control deviation is “positive sma”.
11 ”and the control deviation (e) is“ negat ”
ive small ”, the normalized manipulated variable (U) is set to“ positive big ”. According to the fourth fuzzy control rule (R4), the normalized time differential value (e ′) of the control deviation is “Positive b
ig ”and when another condition is satisfied, the normalized manipulated variable (U) is“ positive small ”.
And the other condition is that the normalized control deviation (e) is “negative big” or “negative big”.
ive small ", and a second set including four fuzzy control rules is applied, and the application of the second set of control rules is based on the normalized time differential value (e ') Is less than or equal to the negative value of the normalized control deviation (e), according to the fifth fuzzy control rule, when the normalized control deviation (e) is "negativesmall", When the condition is also satisfied, the normalized manipulated variable (U)
Is "negative big", and the other condition is that the normalized time derivative of the control deviation is "negaive big".
"gative small" or "positive"
small ", and according to the sixth fuzzy control rule, when the first and second conditions are satisfied, the normalized manipulated variable (U) is" negative ".
e small ”, and the normalization control rule (e) satisfies“ neg
"active small" or "negative b"
ig ", and the second
The condition is satisfied when the normalized time differential value (e ′) of the control deviation is “negative big” or the normalized control deviation (e) is “negative big”. According to the above, the normalized control deviation (e) is “positivesmall” and the normalized time differential value (e ′) of the control deviation is “negative”.
small ", the normalized manipulated variable (U) is" neg "
, and according to the eighth fuzzy control rule, when the normalized time derivative (e ') of the control deviation is "negative big" and another condition is satisfied, The manipulated variable (U) is set to “negative small”, and if the above other condition is satisfied, the normalized control deviation (e) is set to “positive small” or “positive small”.
10. A method as claimed in any one of claims 1 to 9, wherein the method is performed when "ve big". 11. Using the control loop performs the position defined for car operating device (16) device, the actual value (Y) and the set value of the signal associated with the position of the operating device (16) and (W) Means (11) for determining a control deviation (e) between the control characteristics P, I and D having at least one of the control characteristics P, I and D are provided.
A first controller (12) configured as a next control is provided.
And, said first controller (12) sends out the receive control deviation (e) and the first manipulated variable as an input variable (U1), said position determined apparatus is implemented as a fuzzy controller It has second controller (14), said second controller is connected in parallel with the first controller (12)
Cage, said second controller, said sending the receive and the second manipulated variable (U2) the time derivative of additionally control deviation to the control deviation (e) a (e ') as an input variable, first the And the second manipulated variable are summarized as the total manipulated variable.
And constitutes the fuzzy controller (14),
Controller (14), friction force has no disturbing effect
When the control deviation is small, the manipulated variable with a large absolute value (u2)
At the time of an extremely large control deviation,
The fuzzy controller (14) has a relatively small absolute value.
The operation amount, characterized in that to supply, to position the vehicle operating device by using a control loop unit. 12. A parallel connection of said two controllers (12, 14).
When connection, as a conventional controller (12) and by an operation amount determined by the fuzzy controller (14) (U 1, U 2) is superimposed additively, it is transmitted subsequently to the operation device (16) to it It is configured, according to claim 11, wherein.