JPH06206559A - Method and device for controlling quantity having effect on vehicular operation dynamic characteristic or operating condition - Google Patents

Abstract

PURPOSE: To possibly optimally control in a closed loop a quantity having an influence on the running dynamic characteristic or the running condition of a vehicle. CONSTITUTION: A manipulated variable u is obtained through a fuzzy controller 20 from a deviation e between a target value w and an actual value y of a signal concerning a quantity having an influence on the running dynamic characteristic or the running condition of a vehicle, and a time differential of the deviation e. The manipulated variable u is fed to an actuator signal generator 16, e.g. a pulsewidth modulator which outputs a corresponding signal uPWM to an actuator 18, and hence this is suited for control of a steering system, a brake system and the chassis. Such constitution enables the deviation between a target value and an actual value to be possibly kept small under a lot of operating conditions and the control characteristics to be stabilized even in extreme cases.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling an amount that influences a running dynamic characteristic or a running state of a vehicle and an apparatus for carrying out the method, and more particularly to a running dynamic characteristic or a running state of a vehicle. The invention relates to a method and a device for controlling influencing quantities according to the deviation between the desired and actual values of the signals associated with these quantities.

[0002]

From DE 37 34 477 A1 a device for controlling the steering angle of the wheels of a vehicle is known. In this device, a reference variable representing the motion of the vehicle is obtained from the turning angle (turning angle) of the steering wheel and the vehicle speed detected using a sensor based on a predetermined mathematical model, and this is used as a target value. Reference variables are also determined for the actual movement of the vehicle. The controller drives the actuator that adjusts the steering angle of the wheels by comparing the two reference variables with each other to reduce the deviation between the two reference variables.

A system for controlling the speed of a vehicle using fuzzy logic is known from US Pat.
In that case, the actual value of the speed is obtained, and the target value is set in consideration of the driver's request. An appropriate membership function is selected according to the running state (eg gear position). The operation amount for controlling the actuator is obtained by applying the fuzzy rule based on the membership function thus obtained and the target value and the actual value of the traveling speed.

[0004]

SUMMARY OF THE INVENTION An object of the present invention is to control the amount of adjustment of the running dynamic characteristics or running condition of a vehicle as optimally as possible (closed loop). The deviation between the desired value and the actual value must then be as small as possible under as many operating conditions as possible, and the control characteristics must be stable even in extreme cases.

[0005]

SUMMARY OF THE INVENTION According to the invention, the object of the invention is to control the quantities which influence the running dynamics or the running state of a vehicle according to the deviation between the desired and actual values of the signals associated with these quantities. In the method and apparatus described above, a manipulated variable is obtained from a deviation between a target value and an actual value and a time derivative of this deviation using a fuzzy controller, and the manipulated variable is further supplied to an actuator signal generator to generate an actuator signal. This is solved by the arrangement in which the signal is supplied to the actuator by the device.

[0006]

According to the present invention, it is possible to optimally control the amount that affects the running dynamic characteristics or running condition of the vehicle. Since the control circuit according to the present invention does not require a mathematical model of the controlled object, it is possible to save the trouble of finding the model. Further, there is an effect that the structure of the fuzzy controller is simplified. It is easy to adapt it to a specific area of use.

Another advantage is that the fuzzy controller is robust even in extreme situations. In particular,
It is advantageous to use the control circuit according to the invention when the actuator is formed as a switching valve device, for example to drive a rear axle actuator. The switching valve device has the same advantages as the fuzzy controller.
That is, the switching valve device has a simple structure, is strong, and can be formed at low cost. This results in very good overall control characteristics.

[0008]

The present invention will be described in detail below with reference to the embodiments shown in the drawings.

The case where the structure and function of the present invention are used in connection with rear wheel steering will be described as an example. The present invention in this embodiment provides a signal for driving an electromagnetic switching valve that regulates hydraulic flow to at least one rear axle actuator that controls the steering angle. In the following, an embodiment with one rear axle actuator will be described. However, it is also possible in principle to drive multiple rear axle actuators. Furthermore, the present invention relates to chassis (suspension) control, for example,
Alternatively it can be used in connection with a braking system.

The fuzzy controller operates on an internally normalized quantity. Therefore, it is assumed that all quantities are normalized when read into the fuzzy controller, and renormalized when output from the fuzzy controller. The notation of unnormalized quantities is underlined so that the normalized quantity can be easily distinguished from the unnormalized quantity. For example, e (unnormalized control deviation) or e (normalized control deviation).

FIG. 1 shows a block diagram of the principle of a known conventional control circuit. The block with the reference number 10 outputs the target value w to the connection point 11. Sensor 12
The actual value y is detected by the
Is used as the second input signal of. The control deviation e at the connection point 11 is determined by forming the difference between the target value w and the actual value y . The control deviation e is supplied to the controller (controller) 14, and the controller based on this, the manipulated variable u.
Is output to the actuator signal generator 16. The actuator signal generator 16 generates a corresponding drive signal u 1 and applies this signal u 1 to the actuator device 18. The position of the actuator is detected by the sensor 12 as the actual value y .

FIG. 2 shows a principle block diagram of the control circuit according to the present invention. The schematic configuration corresponds to the known control circuit shown in FIG. Of importance to the present invention is the internal structure of the individual blocks, in particular block 14. Block 14 is a fuzzy controller 20 and time derivative of control deviation
It comprises means 22 for forming e '. The fuzzy controller uses the control deviation e and the time derivative of the control deviation as input signals.
e ′ is supplied, and the manipulated variable u is output as an output signal. This manipulated variable is output to the actuator signal generator 16. The actuator signal generator 16 produces a signal u PWM, which is pulse-width modulated accordingly and drives the actuator device 18.

The actuator device 18 comprises a rear axle actuator and a switching valve device, and a hydraulic medium is supplied to the rear axle actuator via the switching valve device. The switching valve is opened and closed with the clock of a pulse width modulated signal, which results in a correspondingly modulated hydraulic flow. The thus controlled hydraulic flow moves the piston of the rear axle actuator, thereby adjusting the turning angle of the wheel.

The actual position y of the rear axle actuator is detected by the sensor 12 and is guided to the junction 11 with a negative sign. The target position w of the rear axle actuator is guided to the connecting point 11 with a positive sign. Two input signals at junction 11
y and w are additively superposed. As a result of the superposition, a control deviation e occurs.

To determine the target position w of the rear axle actuator by block 10, the methods known from the prior art are used. In one method, the steering wheel turning angle is detected by a suitable sensor.
This cutting angle is converted via the characteristic curve into the corresponding target position w of the rear axle actuator. Other quantities which characterize the running condition, for example the yawing speed, can also be taken into account when converting.

FIG. 3 diagrammatically illustrates a fuzzy controller 20 which is composed of three functional units 30, 32 and 34 which are internally connected in series with one another. These three functional units are usually fuzzy units (3
0), fuzzy inference unit (32) and defuzzification unit (34).

The control deviation e and its time derivative e'are supplied to the first functional unit 30 as input signals. Each input signal is normalized. That is, the normalized control deviation e
And the normalized time derivative e ′ of the control deviation is obtained.
Two membership functions μe and μe ′ are formed in relation to the quantities e and e ′ as illustrated in FIG.

E to e'in each membership function
The actual value of is converted, followed by the membership function
Is supplied to the functional unit 32.

In the second functional unit 32, the first
Fuzzy rules are applied to the membership function supplied by the functional unit 30 of the above, resulting in a membership function μu for the manipulated variable. A graphical representation of this operation, commonly referred to as fuzzy inference, is shown in FIG. 8 below. Membership function μu is the third
To the functional unit 34 of the third functional unit 34, on the basis of which the manipulated variable u normalized by averaging is formed.
To occur. From this normalized operation amount u, the operation amount u suitable for the system using the fuzzy controller is obtained and supplied to the output of the fuzzy controller 20.

The membership function is shown in FIG. In order to determine the membership function, a plurality of categories are provided for the normalized control deviation e, the normalized time derivative e ′ of the control deviation and the numerical range of the normalized manipulated variable u. In that case, the position of the partial area within the numerical range is qualitatively described by each category. The degree to which each value in the numerical range is related to the category is determined by the membership function. For example, the degree associated with the “negative and small” category with respect to the normalized control deviation e is described by the membership function μe indicated by “NS” in FIG.

FIG. 4 shows "negative big" (NB), "negative small (negative small)" (NS), "positive small (positive small)" (PS) and "positive". 7 illustrates the membership function for the category of "positive big" (PB). The horizontal axis is the normalized amount e,
e ′ and u are shown, and the vertical axis shows membership functions μe, μe ′ and μu for various categories. The membership functions take values between 0 and 1, respectively. The value 0 means not belonging to the category of interest, and the value 1 means completely belonging. In the embodiment described here, the membership functions of the normalized control deviation e, the normalized time derivative e ′ of the control deviation, and the normalized manipulated variable u are the same for the same category. is there. However, it is also possible in principle to choose different membership functions for these three quantities. It is also possible to select different function curves than those shown in FIG. 4, depending on the respective application. The embodiment shown here has proved to be particularly preferred when driving the rear axle actuator via a switching valve arrangement.

FIG. 5 shows the phase plane for the normalized control deviation e and the normalized time derivative e'of the control deviation, whereby the normalized control deviation e and the control deviation are normalized. It is possible to efficiently and easily derive a fuzzy rule for obtaining the normalized manipulated variable u from the calculated time differential e ′. The horizontal axis shows the normalized control deviation e, and the vertical axis shows the normalized time derivative e ′ of the control deviation. The phase plane is divided into a large number of areas indicated by reference signs A to F. A straight line “e + e ′ = 0” extends from the upper left to the lower right along the diagonal of the phase plane. The fuzzy rule is formed by using the phase plane by associating the category of the normalized manipulated variable u with one region or a combination of a plurality of regions of the phase plane.

FIG. 6 is a table showing the correspondence between the phase plane area (left column) and the normalized manipulated variable u (center column). Based on this correspondence, by describing the region of the phase front by logically combining the categories of the normalized control deviation e and the normalized time derivative e ′ of the control deviation, the fuzzy rule (right Column) can be obtained.

The fuzzy rules (R1 to R4) are listed in FIG. Two areas of phase plane "e '>-e"
A set of rules is defined for each of (R1 and R2) and “e ′ ≦ −e” (R3 and R4). Each fuzzy rule is introduced by the word "IF", followed by the premise (condition). The premise is followed by the word "THEN" and the conclusion. The premise is one of the normalized control deviation e and / or the normalized time derivative e ′ of the control deviation.
It consists of one category display or multiple category displays connected by logical operations. In that case, the quantity for each category indication lies between the two abbreviations for the category. For example, in NEB, the normalized control deviation e (E) is “negative (N: negative)” and large (B:
Big) ”. As a result of the inference, a category display such as “NUB” is obtained for the normalized control amount u (U).

FIG. 8 shows an example in which the fuzzy rule is applied to the exemplary value e0 of the normalized control deviation e and the exemplary value e'0 of the normalized time derivative e'of the control deviation. First, which region of the phase plane (“e ′> − e” or “e ′ ≦
-E ") is required to correspond to the example values e0 and e'0, and rules R1 and R2 or R3 to be applied accordingly
And R4 are selected. For the example values e0 and e'0 shown in FIG. 8, "e'0>-e0" holds. As a result, the fuzzy rules defined for the region "e '>-e" of the phase front, namely R1 and R2, apply.

In the upper part of FIG. 8, the membership functions μe, μe 'and μu of FIG. 4 are shown. Below that the application of fuzzy rules R1 and R2 is illustrated.
In that case, the application of each individual rule shall be read from left to right. In other words, the results of applying the rules are listed on the right side of FIG. 8, respectively. The superposition of the results obtained by applying the rules is done from top to bottom, yielding the results illustrated in the lower right part of FIG.

The application of fuzzy rules is performed as follows.

First, for each category display (vertical line, dotted line) at the exemplary values e0 to e'0, the value of the membership function μe to μe 'of the quantity e to e'indicated by the category display is obtained. , Prerequisite category display (eg R1: ((PES AND (P
E'S OR NE'S)) OR (NES ORN
EB)) is evaluated. The values thus obtained for rule R1 are indicated by dots in FIG. 8 and the individual category representations of the premise of rule R1 are: PES, PE'S, NE'S, N.
I, II, III, IV and V corresponding to ES and NEB
Display with. Therefore, the premise of rule R1 is (I AND (II
OR III)) OR (IV OR V) can also be displayed. When the premise consists of a plurality of category displays (in the case of rule R1, there are five category displays), the largest function value among the category displays combined by “OR” or by “AND” is combined. The minimum function value is sequentially selected from the category display. When expressed in parentheses, it begins with the inner parentheses.

Therefore, in the rule R1, first (II O
R III) is evaluated. Therefore, of the points indicated by II and III, the point representing the larger value of the membership function μe ′ is selected. This is point II. This first
In the above step, the premise of rule R1 is as follows:
(I AND II) OR (IV OR V). The expression (I AND II) is then evaluated. That is, points I and II are compared with each other, and the smaller one of them, ie point I
Are selected for further evaluation. The premise of rule R1 is simplified as follows: I OR (IV OR
V). Therefore, the equation IV OR V is then evaluated, in which case the result IV is obtained. So the premise of rule R1 is: I OR IV. The result of this equation is IV.

The function value (for example, point IV when the assumption regarding rule R1 is evaluated) obtained by the above-mentioned method is
The corresponding membership function of the normalized manipulated variable u is moved horizontally to the category display of the conclusion by horizontally cutting at the level of the function value, so that the area hatched in FIG. 8 is obtained.

In the case of rule R2, the processing is performed in the same manner. Since the function value thus obtained is equal to 0 in this case, no area remains.

The result of the rule application is superimposed by forming the total amount of area determined by the rule application. The area thus obtained is shown in the lower part of FIG. This is the same as the area obtained by applying the rule R1. This is because the example values e0 and e'0 have no component of rule R2.

The individual values for the normalized manipulated variable u are obtained by averaging the areas obtained by the total calculation. This averaging is performed, for example, by defining a geometric center point of the area and projecting it on the horizontal axis. From the operation amount u thus normalized,
For example, the manipulated variable u suitable for the controlled object is obtained by multiplying by the parameter obtained experimentally.

[0034]

As is apparent from the above description, according to the present invention, the amount that affects the traveling dynamic characteristics or the traveling state of the vehicle can be controlled in a closed loop as optimally as possible, and in that case, the target The deviation between the actual value and the actual value is kept as small as possible under a large number of operating conditions, and the control characteristics can be stabilized even in extreme cases.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a known control circuit.

FIG. 2 is a block diagram showing the configuration of a control circuit according to the present invention having a fuzzy controller.

FIG. 3 is a block diagram showing a schematic structure of a fuzzy controller.

FIG. 4 is an explanatory diagram showing a membership function.

FIG. 5 is an explanatory diagram showing a phase plane related to a normalized control deviation e and a normalized time derivative e ′ of the control deviation e.

FIG. 6 is a table showing the correspondence of the normalized operation amount u to the category of the area of the phase plane.

FIG. 7 is a table showing a list of fuzzy rules.

FIG. 8 is an explanatory diagram showing application of fuzzy rules in a graph.

[Explanation of symbols]

 10 Target Value Forming Device 14 Controller 16 Actuator Signal Generator 18 Actuator 20 Fuzzy Controller

 ─────────────────────────────────────────────────── ———————————————————————————————————————————————————————————————————— Inventor Thomas Kürten Germany Federal Republic of Germany 47239 Duisburg Kapelenerstraße 41

Claims (9)

[Claims] 1. A method for controlling an amount which influences a running dynamic characteristic or a running state of a vehicle according to a deviation ( e ) between a target value ( w ) and an actual value ( y ) of a signal related to these amounts, Using the fuzzy controller (20), the manipulated variable ( u ) is obtained from the deviation ( e ) between the target value and the actual value and the time derivative ( e ') of this deviation, and this manipulated variable ( u ) is further calculated as an actuator signal. A quantity which influences the running dynamics or the running state of the vehicle, characterized in that it is supplied to a generator (16) and a signal ( u PWM) is supplied to an actuator (18) by an actuator signal generator. Method. 2. The normalized control deviation (e), the normalized time derivative (e ′) of the control deviation, and the normalized manipulated variable (u) are respectively classified into a plurality of categories (“negatively”). "Large", "Negative small", "Positive small" and "Positive large") are defined as qualitative quantities, and the normalized control deviation (e) or the normalized time derivative of the control deviation is defined. The membership function (μe, μe ′ and μu) that determines the degree of belonging to a category is defined as a function of (e ′) or the value of the normalized manipulated variable (u). the method of. 3. A membership function (μe) of the normalized control deviation (e), a membership function (μe ′) of the normalized time derivative (e ′) of the control deviation, and a normalized manipulated variable. Method according to claim 2, characterized in that the membership functions (u) of (u) are respectively identical for the same category. 4. The value assigned to the membership function is -1.
From 0 to 0, the membership function for the category of "negative and large" decreases from 1 to 0, and when the value assigned to the membership function passes from -1 to 0, "negative" When the membership function for the category of "small in" rises from 0 to 1, and the value assigned to the membership function passes through the range of 0 to 1, the membership function for the category of "positive and small" goes from 1 to The membership function for a category of "positive and large" rises from 0 to 1 when the value assigned to the membership function passes through the region of 0 to 1 which decreases to 0. The method according to 3. 5. A fuzzy controller (20) applies a fuzzy rule to the normalized control deviation (e) and the normalized time derivative (e ') of the control deviation, followed by the application of this rule. The method according to any one of claims 1 to 4, characterized in that the operation amount (u) is obtained by superimposing the results and forming an average value to obtain a normalized operation amount (u). 6. The first set of two fuzzy rules is applied when the normalized time derivative (e ′) of the control deviation is greater than the negative of the normalized control deviation (e), In this case, the first fuzzy rule (R1) is that the normalized control deviation (e) is “positive and small” and the normalized time derivative (e ′) of the control deviation is “positive and small”. Or “negative and small”, the first condition is satisfied, and when the normalized control deviation (e) is “negative and small” or “negative and large”, the second condition is satisfied. When the first or second condition is satisfied, the normalized manipulated variable (u) is “negative and small”, and the second fuzzy rule (R2) is: When the control deviation (e) is “positive and large”, the first condition is satisfied, and The second condition is satisfied when the controlled deviation (e) is “positive and small” and the normalized time derivative (e ′) of the control deviation is “positive and large”. When the first or second condition is satisfied, the normalized manipulated variable (u) is set to “negative and large”, and the normalized time derivative (e ′) of the control deviation is normalized. If the deviation (e) is less than or equal to the negative of the two, then a second set of two fuzzy rules is applied, in which case the third fuzzy rule (R3) is: The first condition is satisfied when e) is “negative and small” and the normalized time derivative (e ′) of the control deviation is “positive and small” or “negative and small”, and When the normalized control deviation (e) is “positive and small” or “positive and large”, Assuming that the second condition is satisfied, the normalized manipulated variable (u) is “positive and small” when the first or second condition is satisfied, and the fourth control rule (R4) is The first condition is satisfied when the normalized control deviation (e) is “negative and large”, and the normalized control deviation (e) is “negative and small” and When the normalized time derivative (e ′) is “negative and large”, it is assumed that the second condition is satisfied, and when the first or second condition is satisfied, the normalized control amount ( The method according to claim 5, wherein u) is “positive and large”. 7. The actuator signal generator (16) comprises:
A pulse width modulated signal ( u PW corresponding to the manipulated variable ( u ).
Method according to claim 1, characterized in that it is a pulse width modulator providing M). 8. The method according to claim 1, wherein at least one of the steering, braking and chassis systems is controlled. 9. A device for controlling quantities affecting the running dynamics or running conditions of a vehicle according to the deviation ( e ) between the desired value ( w ) and the actual value ( y ) of the signals associated with these quantities. Means for obtaining the manipulated variable ( u ) from the deviation ( e ) between the target value ( w ) and the actual value ( y ) and the time derivative ( e ') of this deviation are provided in the form of a fuzzy controller (20), and Further, a means for generating a signal ( u PWM) according to the manipulated variable ( u ) and supplying the signal to the actuator (18) is provided to control the running dynamic characteristics of the vehicle or the quantity affecting the running characteristics. Device to do.