JP2001001736A - Damping coefficient control system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress pitching and rolling movement of a vehicle without sacrificing controllability in suppressing the vertical oscillation of a vehicle. SOLUTION: In this damping force control system, a first target damping force of each wheel for suppressing a vehicle from oscillating in the heaving direction of the vehicle body is calculated based on a vehicle single-wheel model in reference to the skyhook theory (S102). Then a second target damping force of each wheel for suppressing the vehicle from oscillating in the pitching direction is calculated based on a vehicle front-rear wheel model (S104). Further a third target damping force of each wheel for suppressing the vehicle from oscillating in the rolling direction is calculated based on a vehicle left-right wheel model (S106). Then among the first to third target damping forces, the one having the largest absolute value is selected for each wheel (S108). Finally damping force at each wheel position is set to be the selected target damping force as previously selected (S110-116).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a damping force control device for controlling the damping force of each damper provided between each wheel of a vehicle and a vehicle body.

[0002]

2. Description of the Related Art As a first prior art in this kind of damping force control device, an actual damping coefficient is determined based on the skyhook theory, that is, a vertical speed of a vehicle body and a relative speed of the vehicle body in the vertical direction with respect to wheels. A damping force control device that determines an actual damping coefficient in accordance with the ratio of
JP-A-5-294122). In this device, the vertical speed of the vehicle body at each wheel position is decomposed into a roll motion speed, a pitch motion speed, a heave motion speed, and a warp motion speed of the vehicle body, and a pitch gain that changes according to a differential value of the longitudinal acceleration is calculated. Multiply the decomposed pitch movement speed,
A roll gain that changes according to the differential value of the lateral acceleration is multiplied by the decomposed roll motion speed, and a constant gain is multiplied by the heave motion speed and the warp motion speed, respectively. The motion speed, heave motion speed, and warp motion speed are recombined into the vertical speed of the vehicle body for each wheel position, and the recombined vertical speed of the vehicle body for each wheel position and the wheel speed of the vehicle body for each wheel position are recombined. The actual damping coefficient is determined according to the ratio to the relative speed, so that the vertical vibration of the vehicle body at each wheel position is suppressed, and the pitching and rolling of the vehicle body are also effectively suppressed.

Further, as a second prior art, a relative displacement of a vehicle body in a vertical direction with respect to a wheel at each wheel position is detected, and the relative displacement is calculated based on various equations of motion of the vehicle in consideration of pitching, rolling, and the like of the vehicle body. Using the amount of displacement, calculate the vertical speed of the vehicle body at each wheel position, and calculate the relative speed of the vehicle body in the vertical direction with respect to the wheels at each wheel position by differentiating the relative displacement amount, thereby calculating each of the vertical speeds. By determining the actual damping coefficient according to the ratio of the relative speed and the relative speed, only by detecting the relative displacement amount, a damping force control device that controls the damping force of the damper based on the skyhook theory is also provided. This is known (JP-A-6-344743).

[0004]

However, in the above-mentioned first and second prior arts, the algorithm itself for suppressing the vertical vibration of the vehicle body, that is, the sky itself, in accordance with the pitch movement and the roll movement of the vehicle body. Since the algorithm itself based on the hook theory is directly corrected, the correction has the effect of suppressing vibration in the pitch and roll directions of the vehicle, but it is originally applied to suppress the vibration in the vertical direction of the vehicle. This impairs the basic performance of damping force control based on the designed specifications.

[0005]

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and has as its object to control pitch movement and / or roll of a vehicle body while ensuring control performance for suppressing vertical vibration of the vehicle body. It is an object of the present invention to provide a damping force control device that can also expect a vibration suppression effect on motion.

In order to achieve the above object, the first aspect of the present invention is described.
The damping force control device that controls the damping force of each damper provided between each wheel of the vehicle and the vehicle body suppresses vibration in the heave direction of the vehicle body based on a single wheel model of the vehicle. Target damping force calculating means for calculating a first target damping force for each wheel for each wheel, and a second target damping force for suppressing pitchwise vibration of the vehicle body based on front and rear wheel models of the vehicle. A second target damping force calculating means for calculating each of the wheels; a final target damping force for each of the wheels based on the calculated first and second target damping forces for each of the wheels; Target damping force determination means, and outputs a control signal corresponding to the determined final target damping force to each of the dampers, and sets the damping force of each damper to the determined final target damping force. Output means for setting and controlling Lies in the fact. In this case, for example, the first target damping force calculating means calculates the first target damping force according to the amount of motion of the vehicle body in the vertical direction, and the second target damping force calculating means calculates the first target damping force in the pitch direction of the vehicle body. The second target damping force is calculated in accordance with the motion state amount of the target.

According to the first structural feature, the first and second target damping force calculating means calculate the first and second target damping force, respectively, and the functions of the final target damping force determining means and the output means are provided. Accordingly, the damping force of the damper is set and controlled to the final target damping force determined based on the first and second target damping forces for each wheel. In this case, the first target damping force is calculated based on the single wheel model of the vehicle to suppress the vibration in the heave direction of the vehicle body, and the second target damping force is calculated based on the front and rear wheel models of the vehicle. The first target damping force is calculated independently of the first target damping force in order to suppress the vibration in the pitch direction of the vehicle body. Insufficient damping force for the pitch motion of the vehicle body will be compensated while ensuring the control performance of, and the vertical vibration of the vehicle body will be effectively suppressed, and at the same time, the vibration for the pitch motion of the vehicle body will also be suppressed, Good riding comfort and running stability of the vehicle are maintained.

[0008] Further, the final target damping force determining means and the output means, which are the features of the first configuration, are used to determine the larger one of the calculated first and second target damping forces. Selection means for selecting each of the wheels, and a control signal corresponding to each of the selected target damping forces is output to each of the dampers to set and control the damping force of each of the dampers to each of the selected target damping forces. Output means.

According to this, the damping force of the damper is set and controlled for each wheel by the action of the selecting means and the output means to the larger of the first and second target damping forces. The damping force of the damper at each wheel position is based on the front and rear wheel model of the vehicle, provided that the first target damping force is smaller than the second target damping force only when the vibration in the pitch direction of the vehicle body has increased to some extent. Therefore, the second target damping force for suppressing the vibration in the pitch direction of the vehicle body is set. In other cases, the damping force of the damper at each wheel position is set to the first target damping force for suppressing the vibration in the heave direction of the vehicle body based on the single wheel model of the vehicle. Therefore, according to this, the control performance for suppressing the vertical vibration of the vehicle body inherent in the first target damping force calculated by the first target damping force calculating means is more accurately secured, and Insufficient damping force for the pitch motion of the vehicle is compensated, so that the vertical vibration of the vehicle body is more effectively suppressed, and the vibration of the vehicle body for the pitch motion is also suppressed. The properties are kept better.

[0010] A second structural feature of the present invention is that the second target damping force calculating means in the first structural feature is characterized in that the second target damping force calculating means is configured to oscillate the vehicle body in the roll direction based on the left and right wheel models of the vehicle. Is replaced with a second target damping force calculating means for calculating a second target damping force for each wheel. In this case, for example, the second target damping force calculating means calculates the second target damping force according to the amount of motion of the vehicle body in the roll direction.

In the second structural characteristic, the second target damping force for suppressing the vibration in the pitch direction of the vehicle body in the first structural characteristic suppresses the vibration of the vehicle body in the roll direction. For the roll motion of the vehicle body while ensuring the control performance for suppressing the vertical vibration of the vehicle body inherent in the first target damping force. Will be compensated,
The vertical vibration of the vehicle body is effectively suppressed, and at the same time, the vibration due to the roll motion of the vehicle body is also suppressed, so that the riding comfort and the running stability of the vehicle are kept good.

Further, a third structural feature of the present invention is that:
In the first structural feature, the second target damping force calculating means, which is the second structural characteristic, is added as third target damping force calculating means for calculating a third target damping force. The target damping force determining unit is a final target damping force determining unit that determines a final target damping force for each wheel based on the calculated first, second, and third target damping forces for each wheel. It has been replaced. Also in this case,
Selecting means for selecting the maximum target damping force among the calculated first, second, and third target damping forces for each wheel, respectively; Output means for outputting a control signal corresponding to each selected target damping force to each of the dampers and setting and controlling the damping force of each of the dampers to the selected target damping force, respectively.

In this third structural feature, the damping force of the dump is calculated independently instead of the first and second target damping forces in the first and second structural features. The target damping force determined based on the first to third target damping forces, for example, the maximum target damping force among the first, second, and third target damping forces is set and controlled for each wheel. In this case, the first target damping force is calculated based on the single-wheel model of the vehicle to suppress the vibration in the heave direction of the vehicle body,
The second target damping force is calculated based on the front and rear wheel models of the vehicle to suppress vibration in the pitch direction of the vehicle body, and the third target damping force is calculated based on the left and right wheel models of the vehicle in the roll direction of the vehicle body. The first to third target damping forces are calculated independently in order to suppress the vibration. Therefore, according to the third structural feature, the control performance for suppressing the vertical vibration of the vehicle body inherent in the first target damping force calculated by the first target damping force calculating means is secured. In addition, the insufficient damping force for the pitch motion and the roll motion of the vehicle body is compensated, and the vertical vibration of the vehicle body is effectively suppressed, and the vibration of the vehicle body for the pitch motion and the roll motion is also suppressed. In addition, the riding comfort and running stability of the vehicle are favorably maintained.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram showing a vehicle damping force control device according to the embodiment.

This vehicle has a vehicle body BD (shown in FIGS. 3, 5, and 7) and wheels FW1, FW2, RW1, RW at respective wheel positions of left and right front wheels FW1, FW2 and left and right rear wheels RW1, RW2.
Suspension device 10A, 1 provided between W2
0B, 10C, and 10D. These suspension devices 10A, 10B, 10C, 10D
D and each of the wheels FW1, FW2, RW1, and RW2, and is provided between the vehicle body BD and each of the wheels FW1, FW2, RW1, and RW2.
Springs 11a, 11 respectively elastically supporting
b, 11c, 11d and these springs 11a, 11b,
11c, 11d, each wheel FW of the vehicle body BD
1, FW2, RW1, and RW2, respectively, are provided with dampers 12a, 12b, 12c, and 12d. These springs 11a, 11b, 11
c, 11d and dampers 12a, 12b, 12c, 12d
Each of the sets constitutes a shock absorber. The dampers 12a, 12b, 12c, and 12d can change the opening degree of the orifice, and have variable damping coefficients.

The vehicle is provided with dampers 12a and 12a.
An electric control device 20 for variably setting damping coefficients (damping forces) of b, 12c and 12d is provided. The electric control device 20 is composed of a microcomputer and its peripheral circuits, and repeatedly executes the program of FIG. 2 (including the subroutines of FIGS. 3, 5, and 7) every predetermined short time by a built-in timer, and 12a, 12b, 12c, 12
The orifice opening of d is controlled. This electric control device 20
Are connected to sprung acceleration sensors 21a to 21d, relative displacement amount sensors 22a to 22d, pitch angular velocity sensor 23, and roll angular velocity sensor 24.

The sprung acceleration sensors 21a to 21d are
In the positions of the wheels FW1, FW2, RW1, and RW2, the vehicle body B
D (spring member), each wheel FW
Of the body BD at the FW1, RW2, RW1, and RW2 positions
Vertical acceleration x against space_pb1'', X_pb2'', X
_pb3'', X_pb4'' Is detected, and the vertical acceleration x
_pb1'', X_pb2'', X_pb3'', X_pb4''
Power. The vertical acceleration x_pb1'', X_pb2'' 、
x_pb3'', X_pb4'' Indicates that the upward direction is positive and the downward direction
Is negative. The relative displacement sensors 22a to 22d are
Body BD at FW1, FW2, RW1, RW2 positions
(Spring member) and wheels FW1, FW2, RW1, RW2
(Spring unsprung member)
Body BD at FW1, FW2, RW1, RW2 positions
Up and down for each wheel FW1, FW2, RW1, RW2
Relative displacement x_pw1-X_pb1, X_pw2-X_pb2, X_pw3−
x_pb3, X_pw4-X_pb4Are detected, and the same displacement x
_pw1-X_pb1, X_pw2-X_pb2, X_pw3-X _pb3, X_pw4-X
_pb4Is output. Note that these relative changes
Order x_pw1-X_pb1, X_pw2-X_pb2, X_pw3-X_pb3, X
_pw4-X_pb4Is the displacement relative to the predetermined reference displacement.
And the direction of the decrease (shrinkage direction of the damper) is positive
The direction in which both increase (the direction in which the damper extends) is negative.
You.

The pitch angular velocity sensor 23 is constituted by rate sensor assembled near the center of gravity of the vehicle body BD, the detection signal 'to detect the angular velocity P _a' angular velocity P _a pitch direction of the vehicle body BD represents a Is output. The roll angular velocity sensor 24 is constituted by a rate sensor assembled near the center of gravity of the vehicle body BD, detects the angular velocity _Ra ′ in the roll direction of the vehicle body BD, and outputs a detection signal representing the angular velocity _Ra ′. .

Next, the operation of the vehicle damping force control device configured as described above will be described. After turning on an ignition switch (not shown), the electric control device 20 operates as shown in FIG.
Is repeatedly executed every predetermined short time. The execution of this program is started in step 100, and the first to third damping force calculation routines are executed in steps 102, 104 and 106, respectively.

The first damping force calculation routine is shown in detail in FIG. 3, and each damper 12a, 12b for suppressing the vibration in the heave direction (vertical direction) of the vehicle body BD based on a single wheel model of the vehicle. , 12c, and 12d are used to calculate the first target damping force.

Before describing the first damping force calculation routine, a method of calculating the first target damping force will be described. FIG. 4 shows a single-wheel single-degree-of-freedom model of a vehicle. In the figure, _Mb is the mass of the vehicle body BD. x _pb
Is the amount of vertical displacement of the vehicle body BD (spring member) in the absolute space with respect to the reference position. x _pw is the amount of vertical displacement of the wheel W (unsprung member) with respect to the reference position in the absolute space. These displacement amounts x _pb , x _pw are both positive in the upward direction. Ks is the value of the springs 11a, 11
It is a spring constant of the spring 11 representatively representing b, 11c, and 11d. C _s is the value of each damper 12a, 12b, 12c, 1
It is a damping coefficient of the damper 12 representatively representing 2d. C _sk
Is a damping coefficient of the virtual damper 12A that has been skyhook based on the skyhook theory, and is a predetermined constant.

First, a virtual model using a virtual damper 12A instead of the damper 12 will be considered. _Assuming that the vertical acceleration of the vehicle body BD is x _pb ″ and the vertical speed of the vehicle body BD is x _pb ′, the equation of motion of the vehicle body BD in the heave direction of the virtual model is represented by the following equation 1.

[0023]

M _b x _pb ″ = K _s (x _pw −x _pb ) −C _sk x _pb ′ Further, considering a real model using a real damper 12, the equation of motion of the real model in the vertical direction Is expressed as in the following Expression 2.

[0024]

M _b x _pb ″ = K _s (x _pw −x _pb ) + C _s (x _pw ′ −x _pb ′) The following equation 3 is derived from the above equations 1 and 2. Then, according to the relationship defined by Equation 3, the vehicle body BD is driven by the damper 12 (or the damper 12A) based on the skyhook theory.
Is given by the following equation (4).

[0025]

## _EQU3 ## C _s (x _pw '−x _pb ') + C _sk x _pb '= 0

[0026]

Equation 4] _{Fd = C s (x pw '} -x pb') = - C sk x pb ' Next, a description will be given of the first damping force calculating routine, execution of the routine begins at step 150 in FIG. 3 And
In step 152, sprung acceleration sensors 21a to 21d
Each wheel from FW1, FW2, RW1, RW2 and below the vehicle body BD at positions acceleration _{x pb1 '', x pb2 '} ', x pb3 '', x
_{Enter pb4} '' respectively. Next, at step 154, the input vertical accelerations x _pb1 ″, x _pb2 ″, x _pb3 ″,
By integrating x _pb4 ″, each wheel FW1, FW2,
The vertical speeds x _pb1 ′, x _pb2 ′, x _pb3 ′, and x _pb4 ′ of the vehicle body BD at the RW1 and RW2 positions are calculated, respectively.

Next, at step 156, the calculated respective vertical velocity x _pb1 was _{', x pb2', x pb3} ', x pb4' following the number to be multiplied beforehand appropriately set and the skyhook damping coefficient C _sk respectively 5, the respective dampers 12a, 12a
b, 12c, 12d, the first target damping force Fd1, Fd2,
Fd3 and Fd4 are calculated respectively.

[0028]

Fdi = _−Csk × _pbi ′ where “i” in the above equation 5 is a positive number of 1 to 4.
In this example, each wheel FW1, FW2, RW1, RW
Although the two common skyhook damping coefficients _Csk are used, different values may be used for the front wheels FW1, FW2 and the rear wheels RW1, RW2. And step 15
At 8, the execution of the first damping force calculation routine ends.

The second damping force calculation routine is shown in detail in FIG. 5, and is based on the front and rear wheel model of the vehicle.
Each damper 12a for suppressing the vibration in the pitch direction of
The second target damping force is calculated by using 12b, 12c, and 12d.

Before describing the second damping force calculation routine, a method of calculating the second target damping force will be described. FIG. 6 shows a two-degree-of-freedom model of the front and rear wheels of the vehicle. In the figure, x _bf is the amount of vertical displacement of the vehicle body BD (spring member) at the front wheel FW1, FW2 position with respect to the reference position. _xbr is the amount of vertical displacement of the vehicle body BD (spring member) at the positions of the rear wheels RW1 and RW2 with respect to the reference position. x _wf is the front wheel FW1, FW2
This is the amount of vertical displacement of the (unsprung member) with respect to the reference position. x _wr is the amount of vertical displacement of the rear wheels RW1, RW2 (unsprung members) with respect to the reference position. These displacement amounts x _bf , x _wf , x _br , x _wr are all positive in the upward direction. K _sf is the spring 11 at the front wheel FW1, FW2 position.
It is a spring constant of the spring 11f representatively showing a and 11b. K _sr is the spring 11 at the position of the rear wheels RW1, RW2.
It is a spring constant of the spring 11r representatively representing c and 11d. C _sf is the damper 1 at the position of the front wheels FW1 and FW2.
It is an attenuation coefficient of the damper 12f representatively representing 2a and 12b. C _sr is a damping coefficient of the damper 12r representing the dampers 12c and 12d at the positions of the rear wheels RW1 and RW2. C _p is a damping coefficient of the virtual damper 12p for suppressing vibration in the pitch direction of the vehicle body BD, and is a constant appropriately determined in advance.

First, consider a virtual model using a virtual damper 12p instead of the actual dampers 12f and 12r. The vertical acceleration of the body BD is x _pb '',
Assuming that the angular acceleration and the angular velocity of the body BD in the pitch direction are P _a ″ and P _a ′, respectively, the equations of motion of the body BD in the heave direction and the pitch direction of the virtual model are expressed by the following equations (6) and (7), respectively. You. Note that the angular acceleration P _a ″ and the angular velocity P _a ′ in the pitch direction are both equal to the front wheel F of the vehicle body BD.
The rotation direction in which the W1 and FW2 sides rise and the rear wheels RW1 and RW2 sides fall is defined as positive.

[0032]

[6] _{M b x pb '' = K} sf (x wf -x bf) + K sr (x wr -x br)

[0033]

[Equation 7] _{I p P a '' = -L} f K sf (x wf -x bf) + L r K
_sr however _{(x wr -x br) -C p} P a ', I p is the pitch direction of the inertia moment of the vehicle body BD, L _f is the distance between the front axle and the center of gravity, L _r is the rear wheel axle This is the distance between the centers of gravity.

Considering a real model using the actual dampers 12f and 12r, the equations of motion of the vehicle body BD in the heave direction and the pitch direction of the real model are as follows:
9, respectively.

[0035]

(Equation 8)

[0036]

(Equation 9)

The following equations (10) and (11) are derived from the above equations (6) to (9).

[0038]

C _sf (x _wf '-x _bf ') + C _sr (x _wr '-x _br ') = 0

[0039]

[Number 11] _{L f C sf (x wf '} -x bf') -L r C sr (x wr '-x
_br ′) + C _p P _a ′ = 0 Further, according to the relationship defined by these equations (10) and (11), the front-wheel damper 12f and the rear-wheel damper 12r (or the damper 12r) suppress the vibration of the vehicle body BD in the pitch direction. 1
2p), the ideal damping force P applied to the vehicle body BD
F _f and PF _r are represented by the following equations (12) and (13).

[0040]

[Number 12] _{PF f = C sf (x wf} '-x bf') = C p P a '/ (L f + L r)

[0041]

[Expression 13] PF_r= C_sr(x_wr'-X_br') =-C_pP_a'/ (L_f+ L_rNext, the second damping force calculation routine will be described.
Execution of the routine is started in step 160 of FIG.
In step 162, the vehicle body B is detected from the pitch angular velocity sensor 23.
Pitch angular velocity P of D_a'. Next, step 16
4, for suppressing vibration in the pitch direction of the vehicle body BD
Ideal and appropriately determined damping coefficient C _pEnter the above
Pitch angular velocity P_a', And the predetermined front wheel and
Each distance L between the rear wheel axle and the center of gravity_f, L_rEquation 1 using
By executing the operations of 4, 15 in the pitch direction of the vehicle body BD
In order to suppress vibration, a front wheel damper 12f and a rear wheel damper 12f are provided.
Target damping force PF of damper 12r_f, PF_rCalculate each
I do.

[0042]

[Number 14] _{PF f = C p P a '} / (L f + L r)

[0043]

Equation 15] _{PF r = -C p P a '} / (L f + L r) Next, at step 166, the target damping force PF _f of the front wheel damper 12f described above calculations the front left and right wheels FW1, for FW2 and sets each as 2 target damping force PFD1, PFD 2, respectively set the target damping force PF _r of wheel damper 12r after the calculated as the second target damping force PFd3, PFd4 of the left and right rear wheels RW1, for RW2. Then, in step 168, the execution of the second damping force calculation routine ends.

The third damping force calculation routine is shown in detail in FIG. 7, and is based on the left and right wheel models of the vehicle.
Each damper 12a for suppressing the vibration in the roll direction of
This is for calculating the third target damping force by 12b, 12c, 12d.

Before describing the third damping force calculation routine, a method of calculating the third target damping force will be described. FIG. 8 shows a two-degree-of-freedom model of the left and right wheels of the vehicle. In the drawing, x _bm is the amount of vertical displacement of the vehicle body BD (spring member) at the right wheel FW2, RW2 position with respect to the reference position. x _bh is the amount of vertical displacement of the vehicle body BD (sprung member) at the left wheel FW1, RW1 position with respect to the reference position. x _wm is the right wheel FW2, RW2
This is the amount of vertical displacement of the (unsprung member) with respect to the reference position. x _wh is the amount of vertical displacement of the left wheels FW1, RW1 (unsprung members) with respect to the reference position. These displacement amounts x _bm , x _wm , x _bh , x _wh are all positive in the upward direction. K _s is the spring 11 at the right wheel FW2, RW2 position.
b, 11d and the spring constant at the position of the left wheel FW1, RW1.
a and 11c are spring constants of the spring 11h representatively shown. C _s is the damper 1 at the right wheel FW2, RW2 position.
This is the damping coefficient of the damper 12m representatively representing the dampers 12b and 12d, and the damping coefficient of the damper 12h representatively representing the dampers 12a and 12c at the positions of the left wheels FW1 and RW1. C _r is the attenuation coefficient of the virtual damper 12r for suppressing vibration in the roll direction of the vehicle body BD, is a predetermined constant properly.

First, consider a virtual model using a virtual damper 12r instead of the dampers 12m and 12h.
The vertical acceleration of the vehicle body BD is _defined as x _pb ''
Assuming that the angular acceleration and the angular velocity in the roll direction of D are R _a ″ and R _a ′, respectively, the equations of motion of the vehicle body BD in the heave direction and the roll direction of the virtual model are represented by the following equations (16) and (17).
Are represented as follows. Note that these angular accelerations R
_a ″ and the angular velocity R _a ′ are both the left wheels FW1, R of the vehicle body BD.
The rotation direction in which the W1 side rises and the right wheel FW2, RW2 side descends is defined as positive.

[0047]

M _b x _pb ″ = K _s (x _wm −x _bm ) + K _s (x _wh −x _bh )

[0048]

[Number 17] _{I r R a '' = (} K s + K) (x wh -x bh -x wm + x
_{bm) T / 2-C r} R a ' , however, I _r is the rolling direction of the inertia moment of the vehicle body BD, K is the spring constant of the stabilizer, T is a tread of the vehicle.

Considering the real model using the actual dampers 12m and 12h, the equations of motion of the vehicle body BD in the heave direction and the roll direction of the real model are as follows:
8, 19, respectively.

[0050]

(Equation 18)

[0051]

[Equation 19] From the above equations 16 to 19, the following equations 20 and 21 are derived.

[0052]

[Number 20] _{C s (x wh '-x bh} ' + x wm '-x bm') = 0

[0053]

C _s (x _wh '-x _bh ' -x _wm '+ x _bm ') T / 2 +
C _r R _a ′ = 0 Further, according to the relationship defined by the equations (20) and (21), the right-wheel damper 12m and the left-wheel damper 12h (or the damper 1h) are used to suppress vibration in the roll direction of the vehicle body BD.
The ideal damping force applied to the vehicle body BD by 2r) is represented by the following Expressions 22 and 23.

[0054]

RF _m = C _s (x _wm '−x _bm ') = C _r R _a '/ T

[0055]

Equation 23] _{RF h = C s (x pwh} '-x pbh') = - C r R a '/
T Next, the third damping force calculation routine will be described. Execution of the routine is started in step 170 of FIG.
At step 172, the vehicle body B is detected from the roll angular velocity sensor 24.
The roll angular velocity R _a 'of D is input. Next, step 17
At 4, ideal and attenuation coefficient Do was predetermined appropriate for suppressing vibration in the roll direction of the vehicle body BD C _r, the input roll angular velocity R _a ', and be predetermined tread T of the vehicle by executing the operation of the following equation 24 using, for calculating the target damping force RF _m of the right wheel damper 12m and left wheel damper 12h to suppress vibration in the roll direction of the vehicle body BD, the RF _h respectively.

[0056]

## EQU24 ## RF _m = C _r R _a '/ T

[0057]

Equation 25] Then _{RF h = -C r R a '} / T, at step 176, the third target damping force for the right wheel target damping force RF _m of the calculated right wheel damper 12m FW2, RW2 RFd2, and sets respectively as RFd4, third target damping force for the left wheels FW1, RW1 the target damping force RF _h of the left wheel damper 12h was the calculated RFd
1 and RFd3. Then, in step 178, the execution of the third damping force calculation routine ends.

Returning to the description of the program of FIG. 2, the first target damping forces Fd1, Fd2, Fd3, Fd4 and the second target damping forces PFd1, Pd by the processing of steps 102 to 106 will be described.
Fd2, PFd3, PFd4 and the third target damping force RFd1, RFd2,
After the calculation of RFd3 and RFd4, the first to
Of the third target damping force, the one having the largest absolute value is determined for each wheel FW.
1, FW2, RW1, and RW2. That is,
First to third target damping forces Fd1, PFd1, R for the left front wheel FW1
The absolute values | Fd1 |, | PFd1 |, and | RFd1 | of Fd1 are compared, and these absolute values | Fd1 |
|, | RFd1 |
One of the third target damping forces Fd1, PFd1, and RFd1 is set as the target damping force F1 of the damper 12a. The same processing as the processing for the left front wheel FW1 is sequentially performed for the right front wheel FW2, the left rear wheel RW1, and the right rear wheel RW2.
2b, 12c, and 12d, each target damping force Fi (where i
Is an integer of 2 to 4).

Next, at step 110, the relative displacement amount is set.
Relative displacement x from the sensors 22a to 22d_pw1-X_pb1, X
_pw2-X_pb2, X_pw3-X_pb3, X_pw4-X_pb4Each
In step 112, the relative displacement amounts x input in the same manner
_pw1-X_pb1, X_pw2-X_pb2, X _pw3-X_pb3, X_pw4-X
_pb4Of the vehicle body BD by differentiating
Relative speed x for wheels FW1, FW2, RW1, RW2
_pw1'-X_pb1', X_pw2'-X _pb2', X_pw3'-X_pb3', X
_pw4'-X_pb4'Respectively.

After the processing in step 112, step 1
At 14, by referring to the relative speed-damping force table, each target damping force F1, F2, F3, F4 determined by the processing of step 108 and the relative speed x _pw1 ′ calculated by the processing of step 112. −x _pb1 ′, x _pw2 ′ −
_{x pb2 ', x pw3' -x} pb3 ', x pw4' -x pb4 'x pw' -x p b '
And the respective orifice opening degrees OP1, OP2, OP3, OP4 of the dampers 12a, 12b, 12c, 12d are determined. The relative speed-damping force table is built in the microcomputer in advance, and the dampers 12a, _12p for the relative speed x _pw '-x _pb ' as shown in FIG.
The data storing the change characteristics of the damping force F of b, 12c, and 12d is stored for each orifice opening OP. Then, the respective orifice opening degrees OP1, OP2, OP3, OP4
When determining the damping force F on the graph of FIG.
The curves closest to the point determined by i and the relative speed x _pwi '−x _pbi ' (where i is an integer of 1 to 4) are respectively searched, and the opening OP corresponding to the searched curves is determined by the damping force and the opening degree OP. A selection is made for each set of relative speeds.

After the processing of step 114, step 1
At 16, each control signal representing the determined OP 1, OP 2, OP 3, OP 4 is transmitted to the dampers 12 a, 12 b, 12 c,
12d, and output to each damper 12a, 12b, 12c,
The orifice opening of 12d is changed to the orifice opening OP1, O
P2, OP3, and OP4 are respectively set and controlled. as a result,
Each of the dampers 12a, 12b, 12c, 12d generates the determined target damping force F1, F2, F3, F4.

As described above, according to the above-described embodiment, the first target damping force Fd1, which suppresses the vibration in the heave direction of the vehicle body by the processing of the first damping force calculation routine (FIG. 3).
Fd2, Fd3, and Fd4 are calculated for each wheel based on a single wheel model of the vehicle using the skyhook theory. Second
By the processing of the damping force calculation routine (FIG. 5), the second target damping force PFd1, for suppressing the vibration in the pitch direction of the vehicle body.
PFd2, PFd3, PFd4 are calculated for each wheel based on the front and rear wheel model of the vehicle. The third target damping force RFd1, RFd2, RFd3, R for suppressing the roll vibration of the vehicle body by the processing of the third damping force calculation routine (FIG. 7).
Fd4 is calculated for each wheel based on the left and right wheel models of the vehicle. Then, by the processing of step 108 in FIG. 2, the one having the largest absolute value among the first to third target damping forces is the target damping force F1, F2,... For each wheel FW1, FW2, RW1, RW2 position. Selected as F3, F4,
By the processing of steps 110 to 116, each wheel FW1, F
Dampers 12a, 12b at W2, RW1, RW2 positions
12c and 12d are the target damping forces F1, F2, F
3 and F4 are respectively set and controlled. That is, the dampers 12a, 12
The respective damping coefficients b, 12c and 12d are respectively set and controlled to target damping coefficients corresponding to the target damping forces F1, F2, F3 and F4.

As a result, the first target damping force Fd1, Fd2, Fd3, Fd4 is reduced to the second target damping force PFd1, Pd only when the pitch or roll vibration of the vehicle body BD is increased to some extent.
Fd2, PFd3, PFd4 or the third target damping force RFd1, RFd2,
Each wheel FW, provided that it is smaller than RFd3, RFd4
Dampers 12a, 12 at positions 1, 1, FW2, RW1, RW2
The second target damping forces PFd1, PFd2, PFd3, b, 12c and 12d are used to suppress the vibration in the pitch or roll direction of the vehicle body based on the front and rear wheel or left and right wheel models of the vehicle.
PFd4 or third target damping force RFd1, RFd2, RFd3, RFd
4; otherwise, the dampers 12a, 1
The damping forces 2b, 12c, and 12d are the first damping force for suppressing the vibration of the vehicle body in the heave direction based on the single wheel model of the vehicle.
The target damping force is set to Fd1, Fd2, Fd3, Fd4. Therefore, according to this embodiment, the first target damping forces Fd1, Fd
2, the insufficient damping force for the pitch motion and the roll motion of the vehicle body BD is compensated while securing the control performance for suppressing the vertical vibration of the vehicle body BD inherent in Fd3, Fd4. The vertical vibration of the vehicle body BD is effectively suppressed, and at the same time, the vibration of the vehicle body BD with respect to the pitch movement and the roll movement is also suppressed, so that the riding comfort and running stability of the vehicle can be kept good.

In the above embodiment, among the first to third target damping forces, the one having the largest absolute value is determined for each wheel FW.
The target damping force F1, F2, F3, F4 is selected for each of the positions 1, 1, FW2, RW1, RW2. However, when the roll motion of the vehicle body BD does not matter so much, the calculation of the third target damping force is omitted, and the first target damping force Fd1, Fd2, Fd3, Fd4 and the second target damping force PFd1, PFd.
The damping force having the larger absolute value among d2, PFd3, and PFd4 may be selected as the target damping force F1, F2, F3, F4 for each wheel FW1, FW2, RW1, RW2 position. This also allows the first target damping force Fd1, Fd2,
Since the insufficient damping force for the pitch motion of the vehicle body BD is compensated for while maintaining the control performance for suppressing the vertical vibration of the vehicle body BD inherent in the vehicle body BD, the vertical vibration of the vehicle body BD Is effectively suppressed, and at the same time, the vibration with respect to the pitch motion of the vehicle body BD is also suppressed, so that the riding comfort and running stability of the vehicle can be kept good.

If the pitch movement of the vehicle body BD does not matter much, the calculation of the second target damping force is omitted, and the first target damping forces Fd1, Fd2, Fd3, Fd4 and the third target damping force RFd1 are obtained. , RFd2, RFd3, and RFd4, the damping force having the larger absolute value is applied to each wheel FW1, FW2, RW1, RW.
The target damping force F1, F2, F3, F4 may be selected for each of the two positions. This also ensures that the first target damping forces Fd1, Fd2, Fd3, and Fd4 have sufficient control performance to suppress the vertical vibration of the vehicle body BD and that the vehicle body BD has insufficient damping force against roll motion. Since the compensation is performed, the vertical vibration of the vehicle body BD is effectively suppressed, and at the same time, the vibration due to the roll motion of the vehicle body BD is also suppressed, so that the riding comfort and the running stability of the vehicle are maintained well.

Also, in the above embodiment, the first to third
Among the target damping forces, the one having the largest absolute value was determined to be the final target damping force. However, the magnitude relationship between the absolute values of the first to third target damping forces was determined, and the absolute value was increased. As the weight increases, the first to third target damping forces are added and synthesized, or the two target damping forces are extracted from the larger one of the first to third target damping forces and added and synthesized, The final target damping force may be determined based on the first to third target damping forces by adding and combining the two extracted target damping forces with the larger weight being increased. . Further, in the modified example, the larger absolute value of the first and second target damping forces, or the larger absolute value of the first and third target damping forces is used as the final target damping force. The first and second target damping forces, or the first and third target damping forces or the first and third target damping forces are determined by increasing the weight of the larger absolute value of the two target damping forces and adding and combining them. The final target damping force may be determined based on the target damping force. Thus, the final target damping force is determined based on the first to third target damping forces, the first and second target damping forces, or the first and third target damping forces calculated independently. Therefore, the insufficient damping force for the pitch motion and the roll motion of the vehicle body is compensated while securing the control performance for suppressing the original vertical vibration of the vehicle body which the first target damping force has, The vertical vibration of the vehicle body is effectively suppressed, and at the same time, the vibration of the vehicle body due to the pitch motion and the roll motion is also suppressed, so that the riding comfort and running stability of the vehicle can be kept good.

In the above embodiment, each wheel FW
1, FW2, RW1, vertical acceleration x _pb1 of the vehicle body BD as vertical motion variable with respect to the absolute space of the vehicle body BD at RW2 position _{'', x pb2 '',} x pb3 '', x pb4 ' respectively detect' FW1, FW2, RW
The vertical velocity x _pb1 ′, x _pb2 ′, x _pb3 ′, x _pb4 ′ of the body BD at the positions 1, RW2 with respect to the same absolute space may be detected, and the wheels FW1, FW2, RW1, RW may be detected.
Vertical displacements x _pb1 , x _pb2 , x _pb3 , x _pb4 in the vertical direction with respect to the same absolute space of the vehicle body BD at the two positions are detected to calculate vertical velocities x _pb1 ′, x _{pb 2} ′, x _pb3 ′, x _pb4 ′. You may make it. Also, each wheel FW1, FW2, RW1, RW2
Wheels FW1, F of Body BD (Spring Member) at Position
Vertical displacement relative to W2, RW1, RW2 x
_{pw1 -x pb1, x pw2 -x pb2} , x pw 3 -x pb3, x pw4 -x
_{As for pb4} , relative speeds x _pw1 '-x _pb1 ', x _pw2 '-x
_pb2 ', _{xpw3'- xpb3} ', _{xpw4'- xpb4} '
Relative acceleration x _pw1 ″ −x _pb1 ″, x _pw2 ″ −x _pb2 ″, x
_pw3 '' -x _pb3 '', to detect the _{x pw4 '' -x pb4 ''} , the relative speed _{x pw1 '-x pb1', x} pw2 '-x pb2', x pw3 '-x pb3',
x _pw4 '-x _pb4 ' may be calculated.

In the above embodiment, the vehicle body BD
A pitch angular velocity sensor 23 composed of a rate sensor is provided to detect the angular velocity P _a ′ in the pitch direction of the vehicle body. However, the angular velocity P _{a is} determined by the difference in the amount of displacement in the vertical direction between the front part and the rear part of the vehicle body BD. 'May be detected. Further, a roll angular velocity sensor 2 constituted by a rate sensor for detecting the angular velocity _Ra 'of the vehicle body BD in the roll direction.
Instead of 4, the angular velocity R _a ′ may be detected based on the difference between the amount of vertical displacement between the left part and the right part of the vehicle body BD.

Further, instead of some of the sprung acceleration sensors 21a to 21d, the relative displacement sensors 22a to 22d, the pitch angular velocity sensor 23 and the roll angular velocity sensor 24 for directly detecting various motion state quantities, an observer is provided. The motion state quantity of the same part may be detected by estimating a part of the various motion state quantities.

Next, various modifications of the first damping force calculation routine will be described. The various modifications are described in the first
The target damping forces Fd1, Fd2, Fd3, and Fd4 are calculated based on a control theory that can handle a nonlinear plant and provide design specifications in the frequency domain. In the following description, it is assumed that a single wheel model of the vehicle is used.
An embodiment will be described in which only one wheel of W1, FW2, RW1, and RW2 is focused on and only the first target damping force Fd representing the first target damping forces Fd1, Fd2, Fd3, and Fd4 is calculated. It should be noted that the first target damping force can be similarly calculated for the other wheels. First, before describing various modifications, a single-wheel model of a vehicle according to the modification will be described.

A. Model First, a model of the suspension device is considered, and the state space of the device is represented. FIG. 10 is a functional diagram of a suspension device per wheel of a vehicle. _Mb is the mass of the vehicle body (sprung member) BD, _Mw is the mass of the wheel WH (strictly, the unsprung member including the lower arm, etc.), and K
_t is a spring constant of the tire TR. K _s is a spring constant of the spring 11, C _s0 is a linear part (hereinafter referred to as a linear damping coefficient) of a damping coefficient C _s of a damper provided in the suspension device, and C _v is a damping coefficient C _{v This} is a non-linear part of s (hereinafter referred to as a non-linear damping coefficient). The sum of these linear damping coefficients C _s0 and nonlinear damping coefficients C _v is
This is the total damping coefficient of the damper 12 (C _s = C _s0 +
_Cv ). RD is a road surface, and if the displacement amounts of the vehicle body BD, the wheels WH, and the road surface RD are x _pb , x _pw , and x _pr , the following equations of motion 26 and 27 are established.

[0072]

(Equation 26)

[0073]

[Equation 27]

Note that the symbol "'" in the equations (26) and (27) and each mathematical expression to be described later represents one-time differentiation, and the symbol """represents two-time differentiation.

The control input u in this suspension device is a nonlinear damping coefficient _Cv . Accordingly, when the road surface disturbance w ₁ is set as the road surface speed x _pr ′ and the nonlinear damping coefficient C _v is used as the control input u, the suspension device is expressed in a state space, and the following expression 28 is obtained.

[0076]

X _p ′ = A _p x _p + B _p1 w ₁ + B _p2 (x _p ) u where x _p , A _p , B _p1 , and B _p2 (x _p ) are the following Expressions 29 to 32.

[0077]

(Equation 29)

[0078]

[Equation 30]

[0079]

(Equation 31)

[0080]

(Equation 32)

The objective of improving the characteristics of the suspension device in this modified example is a vertical speed x _pb ′ (hereinafter, referred to as a sprung speed x _pb ′) of the vehicle body BD which greatly affects the vibration of the vehicle body BD (spring member). Vertical acceleration x _pb ″ (hereinafter referred to as sprung acceleration x _pb ″) of the vehicle body BD that greatly affects the ride comfort of the vehicle and the vertical relative velocity x _pw ′ − of the wheel WH with respect to the vehicle body BD that greatly affects the vibration of the wheel WH. x _pb '
(Hereinafter referred to as relative speed x _pw ′ −x _{p b} ′). Therefore, as the evaluation output z _p , the sprung speed x _pb ′, the sprung acceleration x _pb ″, and the relative speed x _pw ′ −
x _pb ′ is used. In the suspension device, the sprung acceleration x _pb ″ and the wheel WH with respect to the vehicle body BD
Relative displacement x _pw −x _pb (hereinafter simply referred to as relative displacement x _pw −
because it is easy to detect the) and that x _pb, basically observed output y _p a a sprung acceleration x _pb '' and the relative displacement amount x _pw -x _pb. Moreover, the the observed output y _p contains observation noise w _2, which upon state-space representation, the number below 33,3
It looks like 4.

[0082]

(33) z _p = C _p1 x _p + D _p12 (x _p ) u

[0083]

Y _p = C _p2 x _p + D _p21 w ₂ + D _p22 (x _p ) u where z _p , y _p , C _p1 , D in the above equations 33 and 34
_{p12 (x p), C p2} , D p21, D p22 (x p) are as respectively following equations 35-41.

[0084]

(Equation 35)

[0085]

[Equation 36]

[0086]

(37)

[0087]

(38)

[0088]

[Equation 39]

[0089]

(Equation 40)

[0090]

[Equation 41]

However, the state space representation of the suspension device is a bilinear system since the state quantity x _p is included in the coefficient B _p2 (x _p ) as shown in the equation (28).
In the bilinear system, even if the control input u is changed at the origin x = o, B _p2 (o) = o, so that the control is impossible near the origin. Therefore, the control system of the suspension device cannot be designed by the linear control theory, and the design of the control system is designed to obtain the desired control performance by using the nonlinear H 非 線形 control theory, that is, the sprung speed x _pb ′ , _Sprung acceleration x _pb ''
And a control system that suppresses the relative speed x _pw '-x _pb '. Hereinafter, various design examples of the non-linear H 系 control system according to the present invention and various specific calculation examples of the first target damping force Fd according to various modifications for calculating the first target damping force Fd will be described.

B. First modified example b1. Design examples of the non-linear H∞ state feedback control system First, in order to try Nonlinear H∞ state feedback control system design, the evaluation output z _p and the control input state feedback control system as shown in FIG. 11 plus the frequency weight for u Assume a generalized plant of In this case, the frequency weight is a weight whose magnitude changes according to the frequency, and is a dynamic weight given by a transfer function.
By using this frequency weight, it is possible to increase the weight of a frequency band in which control performance is desired to be improved, and to reduce the weight of a frequency band in which control performance can be ignored. Further, after multiplying the evaluation output z _p and the control input u by the frequency weights W _s (s) and W _u (s), the functions a ₁ (x) and a ₂ (x) of the state quantity x are obtained as nonlinear weight functions. multiply. The non-linear weights a ₁ (x) and a ₂ (x) have characteristics defined by the following equations 42 and 43 in order to obtain a solution resulting from the Riccati equation.

[0093]

A ₁ (x)> 0, a ₂ (x)> 0

[0094]

A ₁ (o) = a ₂ (o) = 1 The nonlinear weight makes it possible to design a control system that more positively suppresses the L ₂ gain. The state space representation of this system is as shown in Equation 44 below.

[0095]

X _p '= A _p x _p + B _p1 w ₁ + B _p2 (x _p ) u Here, the state space representation of the frequency weight W _s (s) applied to the evaluation output z _p is expressed by the following equations 45 and 46. Express as follows.

[0096]

X _w '= A _w x _w + B _w z _p

[0097]

Equation 46] _{z w = C w x w +} D w z p Incidentally, x _w represents the state variable of the frequency weight W _{s (s),} z _w represents the output of the frequency weight W _{s (s),} A _w , B _w ,
C _w and D _w are constant matrices determined by control specifications. These constant matrices A _w , B _w , C _w , and D _w are used to improve the ride comfort (roughness) of the occupant, and to increase the sprung acceleration x _b ″
Is reduced in the frequency range of about 3 to 8 Hz (FIG. 12A), and the gain for the sprung speed x _b ′ is reduced in the frequency range of about 0.5 to 1.5 Hz in order to suppress the resonance of the vehicle body BD. decrease (FIG. 12 (B)), 10~ the gain for relative velocity x _w '-x _b' in order to avoid resonance of the wheel WH
The frequency is determined to be lowered in a frequency range of about 14 Hz (FIG. 12C). Then, the frequency ranges for lowering these gains do not overlap, that is, do not interfere with each other, so that the sprung acceleration x _b ″, the sprung speed x _b ′, and the relative speed x _w ′ − that constitute the evaluation output z _p. Each element of x _b ′ is controlled independently.

The frequency weight W applied to the control input u
The state space representation of _u (s) is represented as in the following Expressions 47 and 48.

[0099]

X _u '= A _u x _u + B _u u

[0100]

Equation 48 Note _{z u = C u x u +} D u u, x u represents the state of the frequency weight W _u (s), z _u represents the output of the frequency weight W _u (s), A _u , B _u ,
_Cu and _Du are constant matrices according to the control specifications. These constant matrices A _u , B _u , C _u , and D _u are used as control inputs u to take into account the response of the electric actuator that controls the damping coefficient.
Is determined so as to be suppressed in a high frequency region in accordance with the frequency characteristics of the actuator (FIG. 1).
2 (D)).

At this time, the state space representation of the generalized plant in the nonlinear H∞ state feedback control system is as shown in the following Expressions 49 to 51.

[0102]

X ′ = Ax + B ₁ w ₁ + B ₂ (x) u

[0103]

Z ₁ = a ₁ (x) (C ₁₁ x + D ₁₂₁ (x) u)

[0104]

Where z ₂ = a ₂ (x) (C ₁₂ x + D ₁₂₂ u) where x, A, B ₁ , B ₂ (x),
C ₁₁ , D ₁₂₁ (x), C ₁₂ , and D ₁₂₂ are as shown in the following Expressions 52 to 59.

[0105]

(Equation 52)

[0106]

(Equation 53)

[0107]

(Equation 54)

[0108]

[Equation 55]

[0109]

C ₁₁ = [D _w C _p1 C _w o]

[0110]

D ₁₂₁ (x) = [D _w D _p12 (x _p )]

[0111]

[Number 58] _{C 12 = [o o C u} ]

[0112]

D ₁₂₂ = D _u Next, in order to obtain a solution based on the Riccati equation, under the conditions defined by the following equation 60,
Rewriting the state space representation of the generalized plant represented by 51 gives the following equations 60-63.

[0113]

D _w D _p12 (x) = 0

[0114]

X ′ = Ax + B ₁ w + B ₂ (x) u

[0115]

[Equation 62] z ₁ = a ₁ (x) C _11x

[0116]

[Equation 63] z_Two= A_Two(x) C₁₂x + a_Two(x) D₁₂₂u where A is a stable matrix representing the damping force control system
Reported that the generalized plant
The internal index is stable "and" evaluated from road surface disturbance w1
L up to force z_TwoGain is less than a certain positive constant γ ”
Add the nonlinear H∞ state feedback control law u = k (x)
Try to design.

The above-mentioned nonlinear H∞ state feedback control law u = k (x) can be obtained if the following condition is satisfied. That is, when D ₁₂₂ ⁻¹ exists and a certain positive constant γ is given,
If there exists a positive definite symmetric solution P that satisfies the Riccati equation of the following equation 64 for this positive constant γ, and the nonlinear weights a ₁ (x) and a ₂ (x) satisfy the constraint of the following equation 65, a closed loop the system was internally stable and one control law u = k (x) of the L ₂ gain and less γ is given by the following Expression 66.

[0118]

[Equation 64]

[0119]

[Equation 65]

[0120]

[Equation 66]

Here, the nonlinear weights a ₁ (x) and a ₂ (x) satisfying the constraint condition of the above equation (66) are exemplified in the following equations (67) and (68).

[0122]

[Equation 67]

[0123]

[Equation 68]

Note that m ₁ (x) in the above equations 67 and 68 is an arbitrary positive definite function. As a result of the calculation by the computer, the positive definite symmetric solution P as described above could be obtained. Then, using the above equation (68), the equation (66) is transformed into the following equation (69).

[0125]

[Equation 69]

[0126] These facts indicate that in order to design a control system using the nonlinear H∞ control theory, Hamilton's equation is generally used.
Although the partial differential inequality called the Jacobi inequality must be solved, the Hamilton-Jacobi inequality is solved by giving the above-mentioned constraint condition to the nonlinear weights a ₁ (x) and a ₂ (x) as described above. It shows that the control law can be designed by solving the Riccati inequality instead. Since the solution of the Riccati inequality can be easily obtained by using known software such as Matlab, a positive definite symmetric solution P can be easily found by this method, and the control law u = k (x) can also be derived.

Further, the above-mentioned D ₁₂₂ does not appear in the Riccati inequality, but relates only to the constraint condition and the control law for the nonlinear weight. This adjustment of the control law using D _122, indicates that it is possible to some extent without re solving Riccati inequalities. That is, the adjustment of the control law means that scaling is performed on the control input u,
When the scaling of the control input u is increased by a factor of 10, D ₁₂₂ becomes 1
/ 10 times, and the B ₂ (x) term in Equation 66 is 100 times,
The term C ₁₂ is multiplied by ten.

Next, in order to confirm the role of the non-linear weight, a generalized plant not using the non-linear weight of the bilinear system is assumed and compared with the above-mentioned generalized plant using the non-linear weight. . That is, the nonlinear weights a ₁ (x) and a ₂ (x) are calculated as a ₁ (x) = 1 and a ₂ (x), respectively.
= 1, and C ₁₂ = o, D ₁₂₂ = I, assuming that the orthogonality condition is satisfied for simplicity. The state space expressions represented by the above equations 61 to 63 are as shown in the following equations 70 to 72.

[0129]

X ′ = Ax + B ₁ w + B ₂ (x) u

[0130]

Equation 71] z ₁ = C ₁₁ x

[0131]

Z ₂ = u Accordingly, the control law u = k (x) of the generalized plant is expressed by the following equation 73.

[0132]

U = B ₂ ^T (x) Px where P is a positive definite symmetric solution that satisfies the Riccati inequality of the following equation 74.

[0133]

[Equation 74]

On the other hand, the linear approximation system in the vicinity of the origin of the generalized plant represented by the above equations 70 to 72 is as shown in the following equations 75 to 77.

[0135]

X '= Ax + B ₁ w

[0136]

[Expression 76] z ₁ = C ₁₁ x

[0137]

Z ₂ = u The Riccati inequality in equation 74 above means that the closed-loop system is internally stable for this generalized plant and the L ₂ gain is less than or equal to γ. That, L ₂ gain of bilinear system is from being determined by the value at the origin (x = o) shown in FIG. 13. This is because the bilinear system has B ₂ (o) = o at the origin, and the control input u does not work, so that the L ₂ gain cannot be improved near the origin. Also, the generalized plant (Equations 70 to 72) in which the control input u is u = o matches the generalized plant (Equations 75 to 77) obtained by linear approximation. even if the control input u is u = o against 70 to 72), closed-loop system is internally stable, which means that L ₂ gain is less than gamma. That is, even if the state quantity x becomes large and the control input u becomes effective, the control system is designed for a generalized plant (equations 70 to 72) in which the control output is represented by the following equations 78 and 79. If, but only L ₂ gain even over a control input u is to ensure that no greater than g _0.

[0138]

[Expression 78] z ₁ = C ₁₁ x

[0139]

Z ₂ = u That is, when the control output is expressed as in the above equations 78 and 79, the control performance may have been improved by using the control input u. It may not be different. Therefore, the non-linear weights a ₁ (x) and a ₂ (x) are used for the control outputs z ₁ and z ₂ to obtain the following Expressions 80 and 81. as represented in the _first line, the control system stifle L ₂ gain of the plant so as to asymptotically to the X-axis representing the origin level becomes possible design.

[0140]

[Expression 80] z ₁ = a ₁ (x) C ₁₁ x

[0141]

Equation 81] _{z 2 = a 2 (x)} u Further, in this control, divides the damping coefficient C _s of the damper 12 to the linear damping coefficient C _s0 and nonlinear damping coefficient C _v, the same non-linear damping coefficient C _v A control system was designed as the control input u.
Then, as shown in FIG. 14 (A), the linear damping coefficient _{Cs0 is set} to the minimum damping force characteristic line of the damper 12 (corresponding to the maximum orifice opening) and the maximum damping force characteristic line (corresponding to the minimum orifice opening). And the control input u
And the gain is controlled in accordance with the frequency, damping coefficient C _s is with changes at both sides of the linear damping coefficient C _s0, the maximum damping force characteristic damping force and the minimum damping force characteristic line according to the attenuation coefficient It fits between the lines. Therefore, the nonlinear damping coefficient C _v can be easily determined in accordance with the design specification of the damper 12, and the damping force control can be performed within a range achievable by the actual damper 12.
The intended damping force control can be performed. For comparison, FIG. 14B shows a Lissajous waveform diagram when the damping coefficient of the damper 12 is controlled based on the skyhook theory. In this case, the Lissajous waveform can be realized by the actual damper 12. Control cannot be performed within a proper range, and the intended damping force cannot be controlled.

[0142] When the damping force of the damper 12 (damping coefficient) is configured to switch stepwise to one of a plurality of stages, in setting the linear damping coefficient C _s0, the linear damping coefficient C _s0 The linear damping coefficient C _s0 is set so that the damping force determined by the following equation is approximately equal to the damping force generated by a predetermined one of the plurality of stages of the damper 12 within a small range of the damping force. I do. In this type of suspension device, the linearity of the damping force with respect to the relative speed is strong in a small range of the damping force, that is, there is a high possibility that the calculated nonlinear damping coefficient C _v is “0”. Therefore, in the damper 12, the possibility that the damper 12 is maintained at the predetermined one stage is high, and the frequency of switching of the damping coefficient is reduced, so that the durability of the damper 12 is kept high.

B2. Specific Calculation Example Next, a specific calculation example of the first target damping force Fd using the above-described nonlinear H∞ state feedback control law will be described.

In this case, the electric control unit 20 includes sprung acceleration sensors 21a and 21a as shown by broken lines in FIG.
b, 21c, 21d, relative displacement sensors 22a, 22
b, 22c, 22d, pitch angular velocity sensor 23 and roll angular velocity sensor 24, as well as each wheel FW1, FW2, R
Tire displacement sensors 25a, 25 provided for each of W1, RW2
5b, 25c, 25d and unsprung acceleration sensor 26a,
26b, 26c, 26d are connected. However, in the following description, an example is shown in which the first target damping force Fd of only one wheel is calculated on behalf of each wheel FW1, FW2, RW1, RW2, so that the tire displacement sensors 25a, 25b, 25
c and 25d are simply described as the tire displacement amount sensor 25, and the unsprung acceleration sensors 26a, 26b, 26
c and 26d will be simply described as the unsprung acceleration sensor 26. Further, the sprung acceleration sensors 21a, 21b, 21
c, 21d and relative displacement sensors 22a, 22b, 22
c and 22d are also simply described as a sprung acceleration sensor 21 and a relative displacement amount sensor 22, respectively.

The tire displacement sensor 25 detects a road surface displacement x _pr
And detects the displacement amount x _pr -x _pw tire TR is the relative displacement of the unsprung displacement x _pw, for example the strain sensor for detecting a deformation of the tire, such as a pressure sensor for detecting the air pressure of the tire Based on the output, the tire displacement x _pr -x
Detect _pw . The unsprung acceleration sensor 26 is fixed to the wheel WH and detects an unsprung acceleration x _pw ″ that is an acceleration in the vertical direction of the wheel WH. The microcomputer in the electric control device 20 calculates the first target damping force Fd by executing the first damping force calculation routine of FIG. 15 every predetermined short time by a built-in timer.

The execution of the first damping force calculation routine is started in step 200, and in step 202, the tire displacement amount sensor 25, the relative displacement amount sensor 22, the sprung acceleration sensor 21, and the unsprung acceleration sensor 26 Each detection signal representing the tire displacement _{xpr- xpw} , the relative displacement _{xpw- xpb} , the sprung acceleration _xpb "and the unsprung acceleration _xpw " is input. Then, in step 204, the sprung acceleration x
By calculating the sprung speed x _pb ′ and the unsprung speed x _pw ′ by time-integrating the _pb ″ and the unsprung acceleration x _pw ″, respectively, by time-differentiating the relative displacement x _pw −x _pb , Calculate the relative speed x _pw '-x _pb '.

Next, in step 206, the following equation 8 which is the same as the above equations 32 and 38 using the relative velocity x _pw '-x _pb '
Calculating B _p2 (x _p ) and D _p12 (x _p ) by the operations of _2, 83, the following equation 84 which is the same as the above equation 55 using these B _p2 (x _p ) and D _p12 (x _p ) B ₂ (x) is calculated by the following calculation.

[0148]

(Equation 82)

[0149]

[Equation 83]

[0150]

[Equation 84]

Note that M _w and M _b in the above equations 82 and 83 are
These are the mass of the wheel WH and the mass of the vehicle body BD, respectively. Further, B _w and B _u in Equation 84 are coefficient matrices related to the frequency weights W _s (s) and W _u (s) set in Equations 45 and 47, and are stored in the microcomputer in advance. Is a constant matrix.

After the processing of step 206, step 2
At 08, the control target of this modification input by the processing of step 202 or calculated by the processing of step 204, the evaluation output z _p defined by the above equation 35
(Spring speed x _pb ′, sprung acceleration x _pb ″, and relative speed x _pw ′ −x _pb ′)
A frequency weight state variable _xw is calculated based on the following equation.

[0153]

X _w ′ = A _w x _w + B _w z _p where A _w and B _w in the above equation 85 are coefficient matrices related to the frequency weight W _s (s) set in the above equation 45. Is a constant matrix stored in advance in the microcomputer.

Next, at step 210, the above equation 47,
Using the same numbers 86 to 88 as in 52 and 69, the state variable x _u of the frequency weight for the control input _u , the expanded state quantity x, and the control input u are calculated.

[0155]

X _u '= A _u x _u + B _u u

[0156]

[Equation 87]

[0157]

[Equation 88]

A _u and B _{u in} Equation 86 are coefficient matrices related to the frequency weight W _u (s) set in Equation 47, and are constant matrices stored in advance in the microcomputer. In addition, D ₁₂₂ in the above 88 is the above 59
And the frequency weight W set in Equation 48 above
_{This is} a coefficient matrix relating to _u (s), which is a constant matrix stored in advance in the microcomputer. m ₁ (x)
Is an arbitrary positive constant function, and an algorithm relating to the function is stored in the microcomputer in advance. Note that this positive constant function m ₁ (x) is a positive constant,
For example, it may be set to “1.0”. C ₁₁ is defined by the above Expressions 37 and 56, that is, the mass M of the vehicle body
and _w and the vehicle body mass M _b, a spring constant K _s of the spring 11, and the linear damping coefficient C _s0 of the damper 12, the frequency weight was set at the number 46 W _{s (s)} the coefficient matrix for C _w, D _w And a constant matrix previously stored in the microcomputer. B ₂ (x) is calculated in step 206
Is a matrix calculated by. P is a positive definite symmetric solution satisfying the above equations 64 and 65, and is a constant matrix stored in the microcomputer in advance. C _12, the above number 58
The frequency weight W defined by
_This is a constant matrix stored in the microcomputer in advance, including a coefficient matrix _Cu for _u (s).

In the calculation of the state variable x _u relating to the frequency weight of the control input u, the expanded state quantity x and the control input u in step 210, each value is given an initial value, and each value x _u , x , U until convergence
Are repeatedly performed to determine each value x _u , x, u.

After the processing of step 210, step 2
At 12, the control input u is equal to the non-linear damping coefficient C _v, overall target damping coefficient of the following equation 89 the damper 12 by the operation of adding the linear damping coefficient C _s0 and the control input u C
Calculate _s . Then, in step 216, the execution of the first damping force calculation routine ends.

[0161]

C _s = C _s0 + C _v = C _s0 + u Next, at step 214, the calculated target damping coefficient C is multiplied by the relative velocity x _pw '−x _pb ′ calculated by the processing at step 204. The first target damping force Fd is calculated by the following equation (90).

[0162]

Fd = C _s (x _pw '−x _pb ') c. Second modified example c1. Design Example of Nonlinear H∞ Output Feedback Control System Next, the design of the nonlinear H∞ state feedback control system is advanced by one step, and the state quantity x _p (tire displacement x _pr −x _pw ,
A part of the relative displacement amount x _pw -x _pb , unsprung speed x _pw ′, sprung acceleration x _pb ″) (for example, tire displacement amount x _pr −x _pw and unsprung speed x _pw ′, or tire displacement amount) x _pr −x _pw , relative displacement x _pw −x _pb and unsprung speed x _pw ′) are estimated by an observer including the control system, and using the estimated values, a non-linear H∞ output feedback is performed as a control system. Attempt to design a control system. In this case, a generalized plant of an output feedback control system as shown in FIG. 16 in which a frequency weight is added to the evaluation output z _p and the control input u is assumed. In this case, after the evaluation output z _p is multiplied by the frequency weight W _s (s), a ₁ (x, xｘ) is multiplied as a nonlinear weight function, and the frequency weight W _u (s) is multiplied by the control input u. Then, a ₂ (x, xｘ) is multiplied as a nonlinear weight function. This nonlinear weight function a ₁ (x,
x ^), a ₂ (x, x ^) have the characteristics shown in the following equations (91) and (92). This enables more aggressive design of the control system, such as to suppress L ₂ gains. Note that x ^ represents a state quantity partially including an estimated value as described above.

[0163]

A ₁ (x, x ^)> 0, a ₂ (x, x ^)> 0

[0164]

A ₁ (o, o) = a ₂ (o, o) = 1 Both the state space expression of this system and the state space expression of the frequency weight W _s (s) applied to the evaluation output z _p are controlled by the control input. The state space representation of the frequency weight W _u (s) applied to u is also the same as in the state feedback control system described above,
97.

[0165]

X _p ′ = A _p x _p + B _p1 w ₁ + B _p2 (x _p ) u

[0166]

X _w '= A _w x _w + B _w z _p

[0167]

[Number 95] _{z w = C w x w +} D w z p

[0168]

X _u '= A _u x _u + B _u u

[0169]

Equation 97 Note _{z u = C u x u +} D u u, the state variable x _w, evaluation function z _w, the constant matrix A _w, B _w,
C _w and D _w are the same as in the case of the state feedback control system.

However, the state space representation of the generalized plant in this nonlinear H∞ output feedback control system is as shown in the following equations (98) to (101).

[0171]

X ′ = Ax + B ₁ w + B ₂ (x) u

[0172]

Z ₁ = a ₁ (x, x ^) (C ₁₁ x + D ₁₂₁ (x) u)

[0173]

[Equation 100] z ₂ = a ₂ (x, x ^) (C ₁₂ x + D ₁₂₂ u)

[0174]

Y = C ₂ x + D ₂₁ w + D ₂₂ (x) u where x, w, A, B ₁ , B _{2 in the} above equations 98 to 101
(x), C ₁₁ , D ₁₂₁ (x), C ₁₂ , D ₁₂₂ , C ₂ , D ₂₁ , D
₂₂ (x) is as shown in the following Expressions 102 to 113.

[0175]

[Equation 102]

[0176]

[Equation 103]

[0177]

[Equation 104]

[0178]

[Equation 105]

[0179]

[Equation 106]

[0180]

C ₁₁ = [D _w C _p1 C _w o]

[0181]

D ₁₂₁ (x) = [D _w D _p12 (x _p )]

[0182]

[Number 109] _{C 12 = [o o C u} ]

[0183]

D ₁₂₂ = D _u

[0184]

[Equation 111] C ₂ = [C _p2 o o]

[0185]

D ₂₁ = [o D _p21 ]

[0186]

D ₂₂ (x) = D _p22 (x _p ) Next, in order to obtain a solution based on the Riccati equation, the above equation 98 is obtained under the conditions defined by the following equation 114.
Rewriting the state space representation of the generalized plant represented by to 101 gives the following equations 115 to 118.

[0187]

D _w D _p12 (x) = o

[0188]

X ′ = Ax + B ₁ w + B ₂ (x) u

[0189]

Z ₁ = a ₁ (x, x ^) C ₁₁ x

[0190]

[Equation 117] z ₂ = a ₂ (x, x ^) C ₁₂ x + a ₂ (x, x ^)
D ₁₂₂ u

[0191]

Y = C_Twox + D_{twenty one}w + D_{twenty two}(x) u Here, as in the case of the state feedback control system described above.
In addition, for the generalized plant,
The internal index is stable "and" L from w to z _TwoGay
H∞ output that satisfies “is less than or equal to some positive constant γ”
Attempt to design feedback control law u = k (y)
You. Furthermore, this nonlinear H∞ output feedback control
In the following, it is divided into the following first to third types and explained respectively.
I do.

C1-1) Example of design of control system of first type This first type is based on B ₂ (x) and Eq.
3 D ₂₂ (x) are known functions ie at least the relative speed x
_{pw'- xpb} 'is observable, and the observer gain L is a constant matrix.

The above-mentioned nonlinear H∞ output feedback control rule u = k (y) can be obtained if the following condition is satisfied. That, D ₁₂₂ ^-1 exists, gamma ₁ are ^{_{γ 1 2 I-D 21 T}} Θ T ΘD 12> o
Is a positive constant that satisfies the following equation, and γ ₂ is> 1, and a Ricchiat inequality for designing an observer (observer gain) of the following expression 119 and a Ricchiat inequality for designing a controller (controller) of the following expression 120 are provided. There exist positive definite symmetric matrices P and Q and positive definite matrix を 満 た す that satisfy the following equation, and the nonlinear weights a ₁ (x, x ^) and a ₂ (x, x ^)
If the constraint conditions of 1,122 are satisfied, one of the control rules represented by the following expression 123 is given by the following expressions 124,125.

[0194]

[Equation 119]

[0195]

[Equation 120]

[0196]

Γ ₂ ² −a ₁ (x, x ^) ² > 0, γ ₂ ² −a
₂ (x, x ^) ² > 0

[0197]

[Equation 122]

[0198]

[Equation 123]

[0199]

X′１２４ = (A + LC ₂ ) x ^ + (B ₂ (x) + LD ₂₂
(x)) u-Ly

[0200]

[Equation 125]

Here, the observer gain L is expressed by the following equation (12).
6 is represented.

[0202]

[Number 126] L = -QC ₂ ^T Θ ^T Θ In addition, "|| ||" represents the Euclidean norm, "||
‖ ₂ ”represents the norm on the square integrable function space L ₂ , and is defined by the following expression 127 with respect to f (t) ∈L ₂ .

[0203]

[Equation 127]

Θ is a positive definite matrix having Θ ⁻¹ , and the observer gain L can be adjusted by using Θ. As in the case of the state feedback control law, the gain L of the controller can be adjusted using _D122 . Furthermore, gamma ₁ is L ₂ gain of the observer, gamma ₂ is L ₂ gain controllers, L ₂ gain of the closed loop system is determined as the product of gamma ₁ and gamma _2. Therefore, it must be determined L ₂ gain of the system to successfully adjust the observer and controller.

Here, non-linear weights a ₁ (x, x ^) and a ₂ (x, x ^) satisfying the constraints of the above equations 121 and 122 are illustrated in the following equations 128 and 129.

[0206]

[Equation 128]

[0207]

[Equation 129]

In addition, m ₁ (x,
x ^) is an arbitrary positive definite function, and ε is ε <1 and εγ ₂ ² >
It is a positive constant of 1. As a result of the calculation by the computer, the positive definite symmetric solution P as described above could be obtained. Using the above equations 128 and 129, the equations 124 and 125 are transformed into the following equations 130 and 131.

[0209]

X ′ ^ = (A + LC ₂ ) x ^ + (B ₂ (x) + LD ₂₂
(x)) u-Ly

[0210]

[Equation 131]

As a result, in this case as well, similarly to the case of the state feedback control system, the solution can be easily obtained by using known software. P can be found, and the estimated state quantity x ′ ^ and the control law u =
k (y) can also be derived.

C1-2) Specific Calculation Example of First Type Next, the first target damping force F using the type 1 control law
A specific calculation example of d will be described. In this case, the tire displacement sensor 25 and the unsprung acceleration sensor 26 (in FIG. 1, the tire displacement sensors 25a, 25b, 25c,
25d and unsprung acceleration sensors 26a, 26b, 26
c, 26d) are omitted, and the microcomputer executes the first damping force calculation routine of FIG. 17 instead of the first damping force calculation routine of FIG. The other parts are the same as in the first modification.

In this case as well, the execution of the first damping force calculation routine shown in FIG.
02a, the relative displacement amount x _pw -x _pb and the sprung acceleration x from the relative displacement sensor 22 and the sprung acceleration sensor 21
Each detection signal representing _pb ″ is input, and a relative speed x _pw ′ −x _pb ′ and a sprung speed x _pb ′ are calculated in step 204a as in the first modification.

Next, at step 206a, the relative speed x
_{The following equation} 13 which is the same as the above equations 32 and 38 using _pw '−x _pb '
In addition to calculating B _p2 (x _p ) and D _p12 (x _p ) by the operations of 2,133, the following equation 134 which is the same as the above equation 106 using these B _p2 (x _p ) and D _p12 (x _p ) By the calculation of B
₂ (x) is calculated, and D ₂₂ (x) is calculated by the calculation of the following Expressions 135 and 136, which are the same as Expressions 41 and 113 above, using the relative speed x _pw '−x _pb '.

[0215]

(Equation 132)

[0216]

[Equation 133]

[0219]

[Equation 134]

[0218]

[Equation 135]

[0219]

D ₂₂ (x) = D _p22 (x _p ) where M _w , M _b , B _w , B
_u is the same value or constant matrix as in the first modification.

After the processing in step 206a, in step 210a, the same number 137,
138, the estimated state quantity x ^ and the control input u are calculated in the same manner as in the first modification.

[0221]

X '^ = (A + LC ₂ ) x ^ + (B ₂ (x) + LD ₂₂
(x)) u-Ly

[0222]

138

A in the above equation (137) is stored in advance in the microcomputer, and
Is a constant matrix determined by L is a constant matrix pre-stored in the microcomputer and defined by the expression 126, and is a positive definite symmetric matrix Q;
9, 111, a constant matrix C ₂ and a positive definite matrix Θ
Is the gain of the observer determined by C _{2 is} also the constant matrix previously stored in the microcomputer. B ₂ (x) and D ₂₂ (x) correspond to step 206a.
Is the matrix calculated in. Also, y is an observed value,
In the first type, the relative displacement amount x _pw -x _pb input in the process of the step 202a and the step 204
The sprung speed x _pb ′ calculated by the processing of a.

In addition, D in the above equation 138₁₂₂Is the above number
The frequency defined by 110 and set by the above equation (48)
Weight W_ua coefficient matrix related to (s),
It is a constant matrix stored in the computer. γ_Two
Is the γ_Two> 1. m₁(x, x ^)
Is an arbitrary positive constant function, and the algorithm
Is stored in the microcomputer in advance.
It is. Note that this positive constant function m₁(x) is a positive constant,
For example, it may be set to “1.0”. C₁₁Is on
Defined by the notation 37,107, ie
Mass M_wAnd the mass M of the body BD_bAnd the spring constant of the spring 11
K_sAnd the linear damping coefficient C of the damper 12_s0And the above Equation 46
The frequency weight W set in_scoefficient matrix C for (s)_w,
D_wIs defined in advance in the microcomputer.
It is a stored constant matrix. B _Two(x)
This is the matrix calculated in step 206a. P is the above number
A positive definite symmetric solution that satisfies 119 and 120
It is a constant matrix stored in the computer. C
₁₂Is defined by the above equation (109) and set by the above equation (48).
The specified frequency weight W_ucoefficient matrix C for (s)_uIncluding
With a constant matrix stored in the microcomputer
is there.

After the processing of step 210a, the first
By the processing of steps 212 and 214 similar to the modification,
The first target damping force Fd calculated with calculating the overall target damping coefficient C _s of the damper 12, and ends the execution of the first damping force calculating routine at step 216.

C2-1) Example of design of control system of second type This second type is composed of B ₂ (x) and Eq.
3 D ₂₂ (x) is an unknown function or relative speed x _pw '-x _pb'
Is unknown, and the observer gain L is a constant matrix.

In this kind of bilinear system, B ₂ (x), D
₂₂ (x) is a linear function of x.
When the generalized plants represented by 15 to 118 are rewritten, the following equations 139 to 142 are obtained. Where B
₂₀ , D ₂₂₀ and d ₁₂₂ are constant matrices.

[0228]

[Number 139] _{x '= Ax + B 1 w} + B 20 xu

[0229]

[Equation 140] z ₁ = a ₁ (x ^) C ₁₁ x

[0230]

(141) z ₂ = a ₂ (x ^) C ₁₂ x + a ₂ (x) d ₁₂₂ u

[0231]

Y = C ₂ x + D ₂₁ w + D ₂₂₀ xu For this generalized plant, we try to design a nonlinear H∞ output feedback control law. When the observer gain L is a constant matrix, an output feedback control law can be designed by the following theorem. That, gamma ₁ is a positive constant satisfying the ^{_{γ 1 2 I-D 21 T}} Θ T ΘD 12> o, and the gamma ₂ is a positive constant is γ _2> 1, and epsilon ₁ ²
When there is a positive constant ε satisfying −u ² > 0,
There are a positive definite symmetric matrix P, Q and a positive definite matrix を 満 た す satisfying the Riccati inequality for designing the observer (observer gain) of 43 and the Riccati inequality for designing the controller of the following Expression 144, and further, the nonlinear weight a ₁ ( x, x ^), a ₂ (x, x ^)
If the constraints of 5,146 are satisfied, one of the control rules represented by the following equation 147 is given by the following equations 148,149.

[0232]

[Equation 143]

[0233]

[Equation 144]

[0234]

[Equation 145] γ ₂ ² −a ₁ (x, x)) ² > 0, γ ₂ ² −a
₂ (x, x ^) ² > 0

[0235]

[Equation 146]

[0236]

[Equation 147] {[z ₁ ^T z ₂ ^T ] ^T ‖ ₂ ≦ γ ₁ γ ₂ ‖w‖ ₂

[0237]

X ′ ^ = (A + L (u) C ₂ ) x ^ + (B ₂₀ + L)
(u) D ₂₂₀ ) x ^ u-L (u) y

[0238]

149

However, the observer gain L (u) is expressed by the following equation (150).

[0240]

Equation 150] L (u) = - The QC ₂ ^T Θ ^T Θ, Θ is a positive definite matrix exists theta ^-1, it is possible to adjust the observer gain L (u) by using the theta . Further, similarly to the case of the above state feedback control law, the gain L of the controller can be adjusted using _d122 .

Here, non-linear weights a ₁ (x, x ^) and a ₂ (x, x を 満 た す) satisfying the constraints of the above equations 145 and 146 are exemplified in the following equations 151 and 152.

[0242]

[Equation 151]

[0243]

[Equation 152]

Incidentally, m ₁ (x,
x ^) is an arbitrary positive definite function, and ε is ε <1 and εγ ₂ ² >
It is a positive constant of 1. Then, as a result of the calculation by the computer, the positive definite symmetric solution P as described above could be obtained. Using the equations 151 and 152, the equations 148 and 149 are transformed into the following equations 153 and 154.

[0245]

Equation 153] x '^ = (A + L (u) C 2) x ^ + (B 20 + L
(u) D ₂₂₀ ) x ^ u-L (u) y

[0246]

[Equation 154]

As a result, in this case, similarly to the case of the state feedback control system described above, a solution can be easily obtained by using known software. P can be found, and the estimated state quantity x ′ ^ and the control law u =
k (y) can also be derived.

C2-2) Specific Calculation Example of Second Type Next, a description will be given of a calculation example of the first target damping force Fd using the control law of the second type. In this case, the relative displacement sensor 22 in FIG. 6 (the relative displacement sensors 22a, 22b, 22c, and 22d in FIG. 1) in the case of the first type is omitted, and the relative displacement in steps 202a and 204a in FIG. Relative displacement x _pw -x _pb from sensor 22
And the calculation of the relative speed x _pw '-x _pb ' and the calculation process in step 206a are omitted, and the calculation according to the second type of control law is executed.

Also in this case, the first damping force calculation routine is started in step 200 of FIG.
The sprung acceleration x _pb ″ is input at step 204a, the sprung velocity x _pb ′ is calculated at step 204a, and the relative speed x _pw ′ is calculated at step 210a using the same numbers 155 and 156 as the above equations 153 and 154. Calculate the estimated state quantity x ′ ^ including the estimate of −x _pb ′ and the control input u.

[0250]

X ′ ^ = (A + L (u) C ₂ ) x ^ + (B ₂₀ + L)
(u) D ₂₂₀ ) x ^ u-L (u) y

[0251]

[Equation 156]

A, L, C ₂ ,
γ ₂ , m ₁ (x, x ^), C ₁₁ , P, C ₁₂ are the same as those of the first type. B ₂₀ , D ₂₂₀ , d
_{Reference numeral 122} denotes an appropriate constant matrix stored in the microcomputer in advance and described above. In this case, y is an observed value, and is the sprung speed x _pb ′ calculated by the processing in step 204a.

After the processing in step 210a, the first
By the processing of steps 212 and 214 similar to the type,
Calculating a first target damping force Fd with calculating the overall target damping coefficient C _s of the damper 12. However, in this case, in the calculation of the first target damping force Fd in step 214,
Estimated relative speed x _pw '^ -x calculated in step 210a
_{Use p b} '^. c3-1) Example of Design of Control System of Third Type This third type also has the B ₂ (x) and the eleven
3 D ₂₂ (x) is an unknown function or relative speed x _pw '-x _pb'
Is unknown, and the observer gain L is a function matrix.

Also in the third type, an attempt is made to design a non-linear H フ ィ ー ド バ ッ ク output feedback control law for the generalized plant represented by the above-mentioned second type by Expressions 139 to 142. When the observer gain L is a function of the control input u, an output feedback control law can be designed by the following theorem. That, gamma ₁ is a positive constant satisfying the ^{_{γ 1 2 I-D 21 T}} Θ T ΘD 12> o, and the gamma ₂ is a positive constant is γ _2> 1, and epsilon ₁ ²
When there is a positive constant ε satisfying −u ² > 0,
There are a positive definite symmetric matrix P, Q and a positive definite matrix を 満 た す satisfying the Riccati inequality for designing the observer (observer gain) of 57 and the Riccati inequality for designing the controller of the following Expression 158, and further, the nonlinear weight a ₁ ( x, x ^), a ₂ (x, x ^)
If the constraint conditions of 9,160 are satisfied, one of the control rules represented by the following equation 161 is given by the following equations 162,163.

[0255]

[Equation 157]

[0256]

[Equation 158]

[0257]

Γ ₂ ² −a ₁ (x, x ^) ² > 0, γ ₂ ² −a
₂ (x, x ^) ² > 0

[0258]

[Equation 160]

[0259]

[Equation 161] {[z ₁ ^T z ₂ ^T ] ^T ‖ ₂ ≦ γ ₁ γ ₂ ‖w‖ ₂

[0260]

X ′ ^ = (A + L ₁ C ₂ ) x ^ + (B ₂₀ + L)
_{2 D 220) x ^ u- (} L 1 + L 2 u) y

[0261]

[Equation 163]

Note that the observer gain L (u) is expressed by the following equation (164).

[0263]

[Number 164] _{L (u) = - QC 2} T Θ T Θ-uQD 220 T Θ T Θ
= L ₁ + uL ₂ L ₁ and L ₂ in the above formula 164 are the following formulas 165 and 16
6 is represented.

[0264]

[Number 165] _{^{L 1 = -QC 2 T Θ T}} Θ

[0265]

L ₂ = −QD ₂₂₀ ^T Θ ^T Θ Θ is a positive definite matrix in which ^{-1 −1} exists, and the observer gain L (u) can be adjusted using Θ. Further, similarly to the case of the above state feedback control law, the gain L of the controller can be adjusted using _d122 .

Here, the non-linear weights a ₁ (x, x ，) and a ₂ (x, x ^) satisfying the constraints of the above equations 159 and 160 are exemplified in the following equations 167 and 168.

[0267]

167

[0268]

168

Note that m ₁ (x,
x ^) is an arbitrary positive definite function, and ε is ε <1 and εγ ₂ ² >
It is a positive constant of 1. As a result of the calculation by the computer, the positive definite symmetric solution P as described above could be obtained. Then, when the above equations 167 and 168 are used, the equations 162 and 163 are transformed into the following equations 169 and 170.

[0270]

X '^ = (A + L ₁ C ₂ ) x ^ + (B ₂ x ^ + L ₂ D
₂₂ x ^) u-L (u) y

[0271]

[Equation 170]

As a result, in this case as well, similarly to the case of the above-mentioned state feedback control system, the solution can be easily obtained by using known software. P can be found, and the estimated state quantity x ′ ^ and the control law u =
k (y) can also be derived. c3-2) Specific Calculation Example of Third Type Next, a specific calculation example of the first target damping force Fd using this third type control law will be described. The configuration in this case is the same as that of the second type.

Also in this case, the execution of the first damping force calculation routine is started in step 200 of FIG. 17, and after the processing of steps 202a and 204a similar to the first type, the above equations 169 and 170 are obtained in step 210a. The same number 17 as
The estimated state quantity x '^ and the control input u are calculated in the same manner as in the case of the second type using 1,172.

[0274]

X '^ = (A + L ₁ C ₂ ) x ^ + (B ₂ x ^ + L ₂ D
₂₂ x ^) u-L (u) y

[0275]

[Equation 172]

A, C ₂ ,
_{B 20 ,, D 220, γ 2} , m 1 (x, x ^), C 1 1, d 122,
P, with respect to C ₁₂ is the same as that of the second type. L, L ₁ , and L ₂ are gains defined by Equations 164 to 166. Further, also in this case, y is an observed value, and is the sprung speed x _pb ′ calculated by the processing in step 204a.

After the processing in step 210a, the second
By the processing of steps 212 and 214 similar to the type,
The first target damping force Fd calculated with calculating the overall target damping coefficient C _s of the damper 12, and ends the execution of the first damping force calculating routine at step 216.

D. Third modified example d1. Design of Kalman Filter Based Nonlinear H∞ Control System For the above model a, the bilinear terms B _p2 (x _p ) and D _p2 (x
_{On the condition that p} ) is known, that is, the relative velocity x _pw '−x _pb ' can be observed, an attempt is made to design an output feedback system when a Kalman filter is used as an observer.

[0279] The reference numerals in the third modification are the same as those in the second modification, and the suffix p is added to the coefficients and variables relating to the plant. The state space expression of the suspension device is expressed by the following equations 173 and 174.
Is represented by

[0280]

X _p '= A _p x _p + B _p1 w ₁ + B _p2 (x _p ) u

[0281]

[Number 174] in the case of _{y p = C p x p +} D p1 w 2 + D p2 (x p) u t → ∞ the Kalman filter, in the case of D _p1 = I,
It is expressed as in the following Expression 175.

[0282]

Equation 175] _{x o '= A p x o} + B p2 u + K (C p x o + D
_p2 however _{(x p) u-y)} , x o, x o ' is an estimated state quantity in the Kalman filter, the filter gain K is as follows number 176.

[0283]

K =-{C _p ^T W ^-1 } The estimated error covariance Σ is a positive definite symmetric solution of the Riccati equation of the following formula 177.

[0284]

数 A _p Σ + pA _p ^T + B _p1 VB _p1 ^T -ΣC _p ^T W ^-1
C _p Σ = o where V is the covariance matrix of w ₁ and W is the covariance matrix of w ₂ .

A block diagram of a generalized plant of this system is as shown in FIG. 18. In this case, the frequency weight W (s) is added to the "estimated state quantity _xo " which is the output of the observer.
And "frequency weight W _u (s) applied to control input u."
Is used as the evaluation output z. That is, here, the control system is designed so that the output of the Kalman filter is reduced using the Kalman filter as a detector. Although this point is different from the first and second modifications, it is considered that if the state estimation is performed well, the same performance as that of the first and second modifications is obtained. The state space representation of the system given by the block diagram of FIG.

[0286]

X _p ′ = A _p x _p + B _p1 w ₁ + B _p2 (x _p ) u

[0287]

[Number 179] _{x o '= A p x o} + B p2 (x p) u + L (C 2 x o +
D _p2 (x _p ) u-y)

[0288]

Equation 180] _{y = C p x p + D} p1 w 2 + D p2 (x p) u

[0289]

[Number 181] _{x w '= A w x w} + B w C s x o

[0290]

182 = z ₁ = a ₁ (x _p , x _o , x _w , x _u ) (C _w x _w +
D _w C _s x _o)

[0291]

X _u '= A _u x _u + B _u u

[0292]

[Number 184] _{z 2 = a 2 (x p} , x o, x w, x u) (C u x u +
D _u u) where x _p is the state quantity of the system, Equation 178 is the state space representation of the system, x _o is the estimated state quantity, Equation 179 is the state space representation of the observer, y is the observation output, and x _w is the frequency weight. Indicates a state. The evaluation outputs z ₁ and z ₂ are given non-linear weights to be designed later.

[0293] For this system, "internal index closed loop system stable" and control law u = k for feedback controlling the state of the observer satisfying "positive constant γ below is L ₂ gain to z from w" (x _o ) design. This system is characterized in that the input for the frequency weight W _s (s) is x _o , as shown in Expression 185 below.

[0294]

[Equation 185]

First, when an error variable is defined as in the following equation 186, the error system is as in the following equations 187 and 188.

[0296]

186 = x _e = x _p- x _o

[0297]

[Number 187] _{x e '= (A p +} LC p) x e + B p1 w 1 + LD p
w ₂

[0298]

188

Further, y _e is multiplied by a constant matrix Θ (scaling matrix) for which an inverse matrix exists, and the above equation 187,
When the error system represented by 188 is modified, the modified system is represented by the following equations 189 and 190.

[0300]

[Number 189] _{x pe '= (A p +} LC p) x pe + B p1 w 1 + LD
_p w ₂

[0301]

Equation 190] _{y e ~ = ΘC p x e} + ΘD p1 w 2 with respect to the deformed error system, the disturbance input w = positive constant gamma ₁ exists in [w ₁ ^T w ₂ ^T] from y _e ~ L up to
Given _{that 2} gain is designed to become like the observer gain L gamma ₁ or less (‖y _e ~ ‖ ₂ ≦ γ _{1 ‖w‖ 2).}

Here, γ₁Is γ₁ID_p1 ^TΘ^TΘD_p1> O
正 y_e~ ‖ _Two≤γ₁{W}_TwoTona
L is given by Equation 191 below.

[0303]

[Number 191] L = -QC _p ^T Θ ^T Θ However, Q is a positive definite symmetric matrix that satisfies the Riccati equation of the following number 192.

[0304]

[Equation 192]

It should be noted that the Riccati equation of Equation 192 to be solved here is the order of the plant, and is smaller than the order of the generalized plant of the first and second modifications.

Next, when the above equation 179 relating to the observer is rewritten, the following equation 193 is obtained.

[0307]

193

The observer represented by the equation (193) is
The positive constant γ used_TwoExists and the observer error y_e^
L to the evaluation output z_TwoGain is γ_TwoThe following ({z}_Two≤γ_Two
‖Y _e~ ‖_Two) To design the controller
think of. Here, the ob-
Variable x for the frequency weight using the_w, X_uTogether
When a generalized plant is configured, the state of the plant
The spatial representation is as shown in the following equations 194 to 196.

[0309]

X _k '= Ax _k + B ₂ (x _p ) u + L ₁ ^{-1 -1} y _e

[0310]

195: z ₁ = a ₁ (x _p , x _k ) C ₁₁ x _k

[0311]

196 = z ₂ = a ₂ (x _p , x _k ) C ₁₂ x _k + a ₂ (x _p , x
_k ) D ₁₂ u However, the variable matrices and the constant matrices in the above equations 194 to 196 are the cages in the following equations 197 to 204.

[0312]

[Equation 197]

[0313]

[Equation 198]

[0314]

[Equation 199]

[0315]

[Equation 200]

[0316]

(Equation 201)

[0317]

[Number 202] _{C 11 = [D w C s} C w o]

[0318]

[Number 203] _{C 12 = [o o C u} ]

[0319]

D ₁₂ = D _u Note that the state quantity x _k defined here does not include the state quantity x _p .

At this time, assuming that D ₁₂ ^-1 exists, there exists a positive definite symmetric solution P of the Riccati inequality of the following Expression 205, and
Furthermore, if the nonlinear weights a ₁ (x _p , x _k ) and a ₂ (x _p , x _k ) satisfy the following expression 206, there exists a certain positive constant γ ₂ and {z‖ ₂ ≦
The controller that satisfies γ ₂ ‖y _e ‖ ₂ is given by Expression 207 below.

[0321]

[Equation 205]

[0322]

[Equation 206]

[0323]

[Equation 207]

Therefore, an observer and a controller satisfying the following Expressions 208 and 209 can be designed.

[0325]

数 y _e ~ ‖ ₂ ≤γ ₁ ‖wγ ₂

[0326]

From Equation 209] ‖z‖ ₂ ≦ γ ₂ ‖y _e ~ ‖ ₂ these things, positive definite symmetric matrix Q satisfying Riccati equation and the following inequality number 210 and 211, it can be seen that P is present.

[0327]

[Equation 210]

[0328]

[Equation 211]

Then, the nonlinear weights a ₁ (x _p , x _k ) and a ₂ (x
Assuming that ( _p , x _k ) satisfies the constraint condition of the following expression 212, the control law of the following expression 213 is given by the following expressions 214 and 215.

[0330]

[Equation 212]

[0331]

[Equation 213] {z} ₂ ≦ γ ₁ γ ₂ ‖w‖ ₂

[0332]

X _k '= (A + L ₁ C ₂ ) x _k + (B ₂ (x _p ) + L ₁
D _p2 (x _p )) u-L ₁ y

[0333]

[Equation 215]

Here, the Riccati equation of the formula (177) used for designing the Kalman filter and the formula (19)
2, the covariance matrix V, W
Is defined as in the following Expressions 216 and 217, the positive definite symmetric solutions 両 and Q of both Riccati equations coincide with each other.

[0335]

[Equation 216] W ⁻¹ = { ^T }

[0336]

217

That is, using the covariance matrices V and W used when designing the Kalman filter,
If Θ and γ ₁ that satisfy 17 are selected, design here and
The observer represented by 18 matches the Kalman filter.

[0338]

X _o '= Ax _o + B ₂ (x _p ) u + L (C _{2 xo} + D
_p2 (x _p ) u−y) where the nonlinear weight a satisfying the constraint condition of the above equation 212
₁ (x _p , x _k ) and a ₂ (x _p , x _k ) are exemplified in the following equations 219 and 220.

[0339]

[Equation 219]

[0340]

[Equation 220]

In addition, m ₁ (x,
x ^) is an arbitrary positive definite function. As a result of the calculation by the computer, the positive definite symmetric solution P as described above could be obtained. Using the equations (219, 220), the equations (214, 215) are transformed into the following equations (221, 222).

[0342]

X _k '= (A + L ₁ C ₂ ) x _k + (B ₂ (x _p ) + L ₁
D _p2 (x _p )) u-L ₁ y

[0343]

(Equation 222)

As a result, also in this case, similarly to the case of the above-mentioned state feedback control system, the solution can be easily obtained by using known software. P can be found, and the state quantity x ′ and the control law u = k
(y) can also be derived.

D2. Specific Calculation Example Next, a specific calculation example of the first target damping force Fd using the Kalman filter-based control law will be described. The configuration in this case is also described in the second
This is the same as the case of the first type of the modified example.

Also in this case, the execution of the first damping force calculation routine is started in step 200 of FIG. 17, and steps 202a and 202 similar to the first type of the second modification are performed.
04a and 210a are executed. However, in this case, in step 210a, the following equations 223 and 224 are used, which are the same as equations 221 and 222, and the first equation of the second modification is used.
The state quantity x _k ′ and the control input u are calculated as in the case of the type.

[0347]

X _k '= (A + L ₁ C ₂ ) x _k + (B ₂ (x _p ) + L ₁
D _p2 (x _p )) u-L ₁ y

[0348]

224

A in the expression 223 is stored in advance in the microcomputer, and A in the expression 198, 185, 3
It is a constant matrix determined by 0,47. L ₁ is stored in advance in the microcomputer and
0, 191, 192, a positive definite symmetric matrix Q, a constant matrix C _p ,
A constant matrix C ₂ determined by 1 and a gain of the observer determined by a positive definite matrix Θ. C _{2 is} also the constant matrix previously stored in the microcomputer.
B ₂ (x _p ) is a constant matrix determined by the equations (199, 32, 47). D _p2 (x _p ) is a constant matrix determined by the equation (38). Further, y is an observed value, and the relative displacement amount x _pw input in the process of step 202a.
−x _pb and the sprung speed x _pb ′ calculated by the process of step 204a.

Further, D ₁₂ in the above formula 224 is the same as the above formula 2
This is a coefficient matrix related to the frequency weight W _u (s) defined by Expression 04 and set in Expression 48, and is a constant matrix stored in advance in the microcomputer. m ₁ (x
_p , x _k ) is an arbitrary positive constant function, and an algorithm relating to the function is stored in the microcomputer in advance. Note that this positive constant function m ₁ (x _p ,
x _k ) may be set to a positive constant, for example, “1.0”. C ₁₁ is defined by the above equation 202, is defined by coefficient matrices C _w , D _w , and C _s relating to the frequency weight W _s (s) set in the above equation 185, and is stored in the microcomputer in advance. It is a constant matrix. B ₂ (x _p ) is a constant matrix determined by the equations (199, 32, 47). P is a positive definite symmetric solution satisfying Equation 211 above,
This is a constant matrix stored in the microcomputer in advance. C ₁₂ is defined by the number 203, is a constant matrix stored in advance in the microcomputer including a frequency weight W _u (s) coefficient matrix for C _u set at the number 48.

After the processing in step 210a, the first
Step 2 similar to the first type of the modified example and the second modified example
The process of 12,214, and calculates a first target damping force Fd with calculating the overall target damping coefficient C _s of the damper 12, and ends the execution of the first damping force calculating routine at step 216.

F. Other Modifications In the first to third modifications, the tire displacement x _pr -x _pw , relative displacement x _pw -x _pb , unsprung speed as state quantities in the state space representation of the generalized plant. x _pw '
And the sprung speed x _pb ′, but other physical quantities related to the vertical movements of the vehicle body BD and the wheels WH can be used if the state space representation is possible. In the first type and the third modification of the second modification, the detection of the tire displacement x _pr -x _pw and the unsprung speed x _pw ′ is omitted and estimated. In the second and third types, the detection of the relative displacement amount x _pw -x _pb (relative velocity x _pw '-x _pb ') is also omitted, and the estimation is performed by slightly changing the control law. It is also possible to omit the detection of the state variables described above.

The evaluation output z of the first to third modified examples is
_{In p} , three kinds of physical quantities, that is, the sprung speed x _pb ′ that affects the resonance of the vehicle body BD,
Relative velocity x _pw '−x _pb ' as affecting the resonance of
Although the sprung acceleration x _pb ″ is used as an influence on the deterioration of the riding comfort of the vehicle (roughness), any one or two of these may be used as the evaluation output z _p. . Furthermore, instead of the sprung speed x _pb ′, a physical quantity closely related to the motion of the car body BD, such as a sprung acceleration x _pb ″ and a sprung displacement x _pb, is used as a factor that affects the resonance of the body BD. You may do so. Wheel W
As an influence on the resonance of H, the relative velocity x _pw '−x _pb '
Instead of the unsprung speed x _pw ′ and the tire displacement x _pr −x _pw
For example, a physical quantity closely related to the motion of the wheel WH may be used.

Further, in the above-mentioned various modified examples, the nonlinear H∞ control theory is applied as a control theory capable of giving a design specification in the frequency domain by handling a non-linear plant. Bilinear Matrix Inequality Bilinear Matrix Inequality that extends control theory
Control theory may be used.

According to the above-mentioned various modifications, the relative speed x of the wheel WH (unsprung member) with respect to the vehicle body BD (sprung member).
_pw a '-x _pb', the relative speed x _pw varies according to the '-x _pb' damping coefficient C _s and the first target damping force Fd = C _s given by the product (x _pw '-x _pb' In the bilinear control system that handles) as well, satisfactory damping that does not give a sense of incongruity to the control specifications (norm condition) given at the time of design and the control input (nonlinear damping coefficient C _v ) continuously changes. A force control device is realized. In addition, the above-described various modifications include the vehicle body B
The vertical velocity x _pb ′ of the vehicle body BD that affects the resonance of
_{Generalization} that the relative speed x _pw '−x _pb ′ of the wheel WH with respect to the vehicle body BD that affects the resonance of H, and the vertical acceleration x _pb ″ of the vehicle body BD that deteriorates the ride comfort (roughness) are evaluated outputs. A first target damping force Fd is calculated assuming a plant, and the vertical speed x _pb ′ and the relative speed x _pw ′ −
Since a predetermined frequency weight is _{assigned to} x _pb ′ and vertical acceleration x _pb ″, the same vertical speed x _pb ′, relative speed x _pw ′ −x _pb ′, and vertical acceleration x _pb ″ according to the frequency band. Can be controlled so as to better suppress the adverse effect on the vehicle. Thus, according to these various modified examples, the first target damping force Fd at which the running stability and the riding comfort of the vehicle are further improved is calculated.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of a damping force control device for a vehicle.

FIG. 2 is a flowchart of a program executed by the electric control device of FIG. 1;

FIG. 3 is a detailed flowchart of a first damping force calculation routine of FIG. 2;

FIG. 4 is a diagram of a single-wheel model of a vehicle.

FIG. 5 is a detailed flowchart of a second damping force calculation routine of FIG. 2;

FIG. 6 is a model diagram of front and rear wheels of a vehicle.

FIG. 7 is a detailed flowchart of a third damping force calculation routine of FIG. 2;

FIG. 8 is a left and right wheel model diagram of the vehicle.

FIG. 9 is a graph showing data characteristics in a relative speed-damping force table.

FIG. 10 is another single-wheel model diagram of a vehicle according to a modification.

FIG. 11 is a block diagram of a generalized plant of a nonlinear H∞ state feedback control system according to a first modification for calculating a first target damping force.

12 (A) to 12 (C) are graphs respectively showing frequency weights for sprung acceleration, sprung speed and relative speed as evaluation outputs, and FIG. 12 (D) is a frequency weight for a non-linear damping coefficient as a control input. It is a graph showing.

FIG. 13 is an image diagram showing the operation and effect of control based on the nonlinear H∞ control theory.

FIG. 14A is a graph showing damping force-relative speed (F-) in damping force control according to a modified example for calculating a first target damping force.
FIG. 5B is a Lissajous waveform diagram showing characteristics, and FIG. 6B is a graph showing damping force-relative speed (FV) in the conventional skyhook control.
It is a Lissajous waveform diagram showing characteristics.

FIG. 15 is a flowchart of a first damping force calculation routine according to a first modification for calculating a first target damping force.

FIG. 16 is a block diagram of a generalized plant of a non-linear H∞ output feedback control system according to a second modification for calculating the first target damping force.

FIG. 17 shows a second and a third calculation for calculating the first target damping force.
9 is a flowchart of a first damping force calculation routine according to a modification.

FIG. 18 is a block diagram of a generalized plant of a Kalman filter-based nonlinear H∞ control system according to a third modification for calculating the first target damping force.

[Explanation of symbols]

FW1, FW2: Left and right front wheels (unsprung members), RW1, R
W2: left and right rear wheels (unsprung member), BD: body (spring member), 10A to 10D: suspension device, 11, 1
1a to 11d, 11f, 11r, 11m, 11h ... springs, 12, 12a to 12d, 12f, 12r, 12m,
12h: damper, 20: electric control device (microcomputer), 21a to 21d: sprung acceleration sensor, 22
a to 22d: relative displacement amount sensor, 23: pitch angular velocity sensor, 24: roll angular velocity sensor, 25, 25a to 25
d: tire displacement amount sensor, 26, 26a to 26d: unsprung acceleration sensor.

Claims (9)

[Claims] 1. A damping force control device for controlling damping force of each damper provided between each wheel of a vehicle and a vehicle body, wherein the vibration of the vehicle body in a heave direction is suppressed based on a single wheel model of the vehicle. A first target damping force calculating means for calculating a first target damping force for each wheel, and a second target damping force for suppressing pitchwise vibration of the vehicle body based on a front and rear wheel model of the vehicle. A second target damping force calculating means for calculating each of the first and second target damping forces for each of the wheels, and a final target damping force for determining a final target damping force for each of the wheels based on the calculated first and second target damping forces for each of the wheels. Force determination means, and a control signal corresponding to the determined final target damping force is output to each of the dampers, and the damping force by each damper is set and controlled to the determined final target damping force. Output means for And a damping force control device. 2. A damping force control device for controlling a damping force of each damper provided between each wheel of a vehicle and a vehicle body, wherein the vibration of the vehicle body in a heave direction is suppressed based on a single wheel model of the vehicle. A first target damping force calculating means for calculating a first target damping force for each wheel, and a second target damping force for suppressing pitchwise vibration of the vehicle body based on a front and rear wheel model of the vehicle. A second target damping force calculating means for calculating each of the calculated first and second target damping forces, and a selecting means for selecting a larger one of the calculated first and second target damping forces for each wheel. Output means for outputting a control signal corresponding to each target damping force to each of said dampers and for setting and controlling the damping force of each of said dampers to said selected target damping force, respectively. Control device. 3. The first target damping force calculating means calculates a first target damping force according to a vertical motion state amount of the vehicle body, and the second target damping force calculating means calculates a first target damping force in the pitch direction of the vehicle body. 3. The damping force control device according to claim 1, wherein the second target damping force is calculated according to the amount of motion state of the damping force. 4. 4. A damping force control device for controlling a damping force of each damper provided between each wheel of a vehicle and a vehicle body, wherein a vibration in a heave direction of the vehicle body is suppressed based on a single wheel model of the vehicle. A first target damping force calculating means for calculating the first target damping force for each wheel, and a second target damping force for suppressing vibration in the roll direction of the vehicle body based on the left and right wheel models of the vehicle. A second target damping force calculating means for calculating each of the first and second target damping forces for each of the wheels, and a final target damping force for determining a final target damping force for each of the wheels based on the calculated first and second target damping forces for each of the wheels. Force determination means, and a control signal corresponding to the determined final target damping force is output to each of the dampers, and the damping force by each damper is set and controlled to the determined final target damping force. Output means for And a damping force control device. 5. A damping force control device for controlling a damping force of each damper provided between each wheel of a vehicle and a vehicle body, wherein a vibration in a heave direction of the vehicle body is suppressed based on a single wheel model of the vehicle. A first target damping force calculating means for calculating the first target damping force for each wheel, and a second target damping force for suppressing vibration in the roll direction of the vehicle body based on the left and right wheel models of the vehicle. A second target damping force calculating means for calculating each of the calculated first and second target damping forces, and a selecting means for selecting a larger one of the calculated first and second target damping forces for each wheel. Output means for outputting a control signal corresponding to each target damping force to each of said dampers and for setting and controlling the damping force of each of said dampers to said selected target damping force, respectively. Control device. 6. The first target damping force calculating means calculates the first target damping force according to the amount of vertical movement of the vehicle body, and the second target damping force calculating means calculates the first target damping force in the roll direction of the vehicle body. The damping force control device according to claim 4 or 5, wherein the second target damping force is calculated according to the amount of motion state of the second target damping force. 7. A damping force control device for controlling damping force of each damper provided between each wheel of a vehicle and a vehicle body, wherein the vibration in the heave direction of the vehicle body is suppressed based on a single wheel model of the vehicle. A first target damping force calculating means for calculating a first target damping force for each wheel, and a second target damping force for suppressing pitchwise vibration of the vehicle body based on a front and rear wheel model of the vehicle. A second target damping force calculating means for calculating each of the wheels, and a third target damping force for calculating, for each wheel, a third target damping force for suppressing vibration in the roll direction of the vehicle body based on the left and right wheel models of the vehicle. Force calculating means; final target damping force determining means for determining a final target damping force for each wheel based on the calculated first, second, and third target damping forces for each wheel; According to each final target damping force determined Damping force control apparatus characterized by comprising an output means for the setting controlled to each final target damping force of the damping force and the determination by the dampers respectively output to the respective dampers to the control signal. 8. A damping force control device for controlling a damping force of each damper provided between each wheel of a vehicle and a vehicle body, wherein the vibration in the heave direction of the vehicle body is suppressed based on a single wheel model of the vehicle. A first target damping force calculating means for calculating a first target damping force for each wheel, and a second target damping force for suppressing pitchwise vibration of the vehicle body based on a front and rear wheel model of the vehicle. A second target damping force calculating means for calculating each of the wheels, and a third target damping force for calculating, for each wheel, a third target damping force for suppressing vibration in the roll direction of the vehicle body based on the left and right wheel models of the vehicle. Force calculating means, selecting means for selecting the maximum target damping force among the calculated first, second, and third target damping forces for each wheel, respectively, and according to the selected target damping forces. Control signals are applied to each of the dampers. Damping force control apparatus characterized by by the force of the damping force by the respective dampers and output means for setting control the respective target damping force said selected. 9. The first target damping force calculating means calculates the first target damping force according to the amount of movement of the vehicle body in the vertical direction, and the second target damping force calculating means calculates the first target damping force in the pitch direction of the vehicle body. Calculating the second target damping force according to the amount of motion of the vehicle, and the third target damping force calculating means calculates the third target damping force according to the amount of motion of the vehicle body in the roll direction. The damping force control device according to claim 7 or 8.