JP3454090B2 - Anti-lock brake control device, torque gradient estimating device, and braking torque gradient estimating device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an antilock brake control device, a torque gradient estimating device, and a braking torque gradient estimating device, and more particularly, to a braking torque or a driving force with respect to a slip speed based on time series data of wheel speeds. A torque gradient estimator that estimates a torque gradient with a small number of parameters, a braking torque gradient estimator that estimates a braking torque gradient with respect to a slip speed with a small number of parameters from time series data of wheel deceleration and brake torque, and an estimated The present invention relates to an antilock brake control device that controls a braking force that acts on a wheel based on a gradient of a braking torque.

[0002]

2. Description of the Related Art A conventional anti-lock brake control device creates a vehicle speed, a vehicle acceleration / deceleration, or a speed signal close to the vehicle speed based on a signal from a wheel speed sensor,
Based on these comparisons, the anti-lock braking operation is performed by controlling the braking force.

That is, in Japanese Unexamined Patent Publication No. 61-196853, a comparison is made between an estimated estimated vehicle speed and a reference speed obtained from the wheel speed or the like to determine whether or not the wheel may be locked, An anti-lock brake controller is described that reduces the braking force when there is a possibility of locking. In this antilock brake control apparatus, the estimated vehicle speed v _v is obtained by connecting the troughs of the velocity v _w determined from the wheel speed as shown in FIG. 16 at a constant rate, the estimated vehicle speed v _v and the actual vehicle body It can be understood that there is a deviation from the velocity v _{v *} .

Further, in this anti-lock brake control device, in order to prevent the estimated vehicle speed v _v by a change in the wheel ground load during running on a rough road becomes larger actual vehicle speed v _{v *,} the estimated vehicle speed When the wheel speed changes more than the change, the increase rate of the estimated vehicle speed is suppressed.

When the vehicle is traveling at a certain speed,
As you apply the brakes, slippage will occur between the wheels and the road surface.
Although it occurs, the friction coefficient μ between the wheel and the road surface is expressed by the following equation.
It is possible to change the slip ratio S as shown in FIG.
Is known. Note that v _{v *}Is the actual vehicle speed, v_wIs a car
Wheel speed.

[0006] In _{S = (v v * -v w} ) / v v * The mu-S characteristic, the coefficient of friction mu is to a peak value at a certain slip ratio (A2 region of FIG. 17). If the slip ratio that takes this peak value is known in advance, the slip ratio can be controlled by obtaining the slip ratio from the vehicle body speed and the wheel speed.

Therefore, in the anti-lock brake control device of Japanese Patent Laid-Open No. 1-249559, the slip ratio is calculated from the approximate value of the vehicle speed and the wheel speed, and the calculated slip ratio is compared with the set slip ratio. It controls the braking force. In this anti-lock brake control system, in order to prevent the long a state of no-brake by the deviation between the estimated vehicle speed v _v and the actual vehicle speed v _{v *,} a long time brake pressure than necessary in a reduced pressure state I try not to.

As shown in FIG. 18, these conventional anti-lock brake control devices include a vehicle body speed estimating section 2 for estimating a vehicle body estimated speed v _v from a wheel speed ω _w and a vehicle body acceleration dv _v / dt, and a wheel. Speed ω _w and estimated vehicle speed v
_v and a braking force control unit 3 that detects the locked state of the wheels and controls the braking force P _b for the driving system 1 of the vehicle. Further, in the braking force control unit 3,
The so-called PID control or the like is used to control the braking force for all four wheels or for each wheel.

[0009]

However, such a conventional antilock brake control device has the following problems.

That is, since the vehicle body speed estimating section is required, as shown in FIG. 16 for estimating the vehicle body speed, the speed v _w obtained from the wheel speed and the actual vehicle speed v _{v *} are equal or close to each other. It was necessary to return the braking force until it became, and for that purpose, it was necessary to repeatedly increase and decrease the braking force of the wheels at a relatively low frequency. In addition, since the vehicle body speed compared with the reference speed is an approximate value obtained from the wheel speed, vehicle body acceleration / deceleration, etc., it may differ greatly from the actual vehicle body speed, and in some cases the wheels may be locked for a long time, There was a problem that the braking force was extremely reduced for the restoration. As a result, the behavior of the vehicle may be significantly affected, resulting in an increase in braking distance and unpleasant vibration.

Further, in an anti-lock brake control device that controls the braking force by the slip ratio, it is easy to predict that the slip ratio, which is the maximum friction coefficient, differs depending on the road surface condition on which the vehicle is running. It is necessary to detect and estimate the condition and prepare a plurality of reference slip ratios according to the road surface condition, or change the reference slip ratio according to the road condition.

US Pat. No. 4,794,538 (Dec. 2)
7,1988), the friction coefficient μ between the road surface and the wheel is estimated from the wheel cylinder pressure and the wheel speed.
A method of controlling a braking force acting on a wheel based on a value is disclosed. In this technology, three parameters (p ₂ , p ₁ , c ₁ ) are identified by applying an online identification method based on a mathematical model of wheel and vehicle body speed from time series data of wheel cylinder pressure and wheel speed. Based on the determined parameter p ₂ , the road surface μ gradient is calculated,
The μ value is calculated from the gradient. According to this technique, the friction coefficient μ and the μ gradient for each road surface can be obtained without estimating the vehicle speed, so that some of the above problems can be solved.

However, in general, when the system identification method is applied, a calculation amount proportional to the square of the number of parameters to be identified is required, and the identification accuracy also deteriorates as the number of parameters increases. As a result, 3 This conventional method, in which two parameters must be identified at the same time, requires a large amount of calculation and has a problem in identification accuracy.

Further, any of the ABS control systems described above is a system having non-linear characteristics with strong tire characteristics.
Braking force P _b for all four wheels or PID independently for each wheel
In the conventional ABS control means that controls using control, etc.,
There is also a problem in that it is not possible to perform fine ABS control because the interference of the four wheels is not taken into consideration.

The present invention has been made in order to solve the above-mentioned problems of the prior art. Instead of detecting the locked state of the wheel from the comparison of the wheel speed or the vehicle speed or the comparison of the slip ratio, the braking torque with respect to the slip speed is obtained. And an antilock brake control device capable of stably and comfortably performing an antilock brake operation irrespective of the condition of the road surface on the basis of the gradient of the braking torque based on the gradient of the braking torque. It is an object of the present invention to provide an antilock brake control device capable of performing fine ABS control in consideration of wheel interference and the like.

Another object of the present invention is to provide a torque gradient estimating device and a braking torque gradient estimating device capable of highly accurately estimating the gradient of the braking torque or the driving torque with respect to the slip speed with a small number of parameters. It is in.

[0017]

(Means for Solving the Antilock Brake Control Device of the Present Invention) In order to achieve the above object, the invention of claim 1 is a wheel for detecting a wheel speed at every predetermined sample time. and speed detecting means, using only time-series data of the detected wheel speed by the wheel speed detecting means as a detection target, the torque gradient estimating means to estimate the gradient of the braking torque with respect to the slip speed, the torque gradient estimating means The control means for controlling the braking force acting on the wheels so that the gradient of the braking torque estimated by means of becomes a value in a predetermined range including the reference value.

According to a second aspect of the present invention, the torque gradient estimating means according to the first aspect calculates a physical quantity related to a change in the wheel speed and a physical quantity related to a change in the wheel speed based on the detected time-series data of the wheel speed. Based on the first computing means for computing, the physical quantity relating to the change in the wheel speed and the physical quantity relating to the change in the wheel speed calculated by the first computing means, the history of the physical quantity relating to the change in the wheel speed and the wheel speed And a second calculation means for calculating a physical quantity representing a history of the physical quantity related to the change and estimating the gradient of the braking torque from the physical quantity.

According to a third aspect of the present invention, the second computing means of the second aspect determines the braking torque and the motion state of the wheel when the braking torque acts, and the braking torque is a gradient of the braking torque with respect to the slip speed. While approximating by a gradient model that changes linearly according to, the approximated motion state, the gradient of the braking torque with respect to the slip speed which is a parameter to be identified, the physical quantity relating to the change of the wheel speed and the change of the wheel speed. The physical quantity relating to the change is converted in advance, and the physical quantity relating to the change in the wheel speed calculated by the first computing means and the physical quantity relating to the change in the wheel speed are sequentially applied to each of the above-mentioned relations. It is characterized by estimating the gradient of braking torque with respect to slip speed by applying the system identification method of That.

According to a fourth aspect of the invention, the first calculating means of the second aspect is the sampling time k at the wheel with the wheel number i.
When the time series data of the wheel speed detected at (k = 1, 2, ...) Is ω _i [k], the sample time is τ, and the wheel inertia is J, As a physical quantity,

[0021]

[Equation 5]

[Mathematical formula-see original document] y _i [k] = − ω _i [k] + 2ω _i [k−1] −ω _i [k−2] is calculated as a physical quantity relating to a change in wheel speed. The second computing means sets a physical quantity θ _i representing a history of physical quantities related to changes in wheel speeds and a history of physical quantities related to changes in changes in wheel speeds to a forgetting coefficient λ and transposition of a matrix as “ ^T ”.

[0023]

[Equation 6]

It is characterized in that the first element of the matrix of the estimated value θ _i is obtained as the gradient of the braking torque with respect to the slip speed by estimating from the recurrence formula of ^.

Note that θ _i with ∧ means a matrix having estimated elements of the first element (first component of matrix) and the second element (second component of matrix) in the matrix θ _i as elements.

According to a fifth aspect of the present invention, time-series data of wheel deceleration detected at each predetermined sample time, and brake torque detected at each predetermined sample time or a time series of a physical quantity related to the brake torque. Based on the data
A torque gradient estimating means for estimating a gradient of the braking torque with respect to the slip speed, and a braking force acting on the wheels so that the gradient of the braking torque estimated by the torque gradient estimating means falls within a predetermined range including a reference value. In the antilock brake control device having the control means, the torque gradient estimating means changes the wheel motion state when the braking torque and the braking torque act to the gradient of the braking torque with respect to the slip speed. In addition to being approximated by a gradient model that changes in a linear function, the approximated motion state is represented by the gradient of the braking torque with respect to the slip speed, which is a parameter to be identified, and the braking torque and the wheel deceleration, respectively. The relationship between the physical quantity related to the change in braking torque and the physical quantity related to the change in slip speed. Is converted in advance to time-series data of the detected wheel deceleration and time-series data of the detected brake torque or the physical quantity related to the brake torque, and the data is sequentially applied to the above-mentioned relationship. It is characterized in that the gradient of the braking torque with respect to the slip speed is estimated by applying the identification method.

The invention of claim 6 is the same as that of claim 5.
The torque gradient estimating means determines whether the wheel with the wheel number i is
Time series data of wheel deceleration at pull time j is y
_i[J], the time series data of the brake torque is T
_bi[J], the predetermined sample time is τ, and the wheel inertia is
J, wheel radius is R_c, The vehicle mass is M, and the
T is a vector that has time series data of torque
_b[J], time series data of wheel deceleration of each wheel
Y [j], the identity matrix is I, and the diagonal component is
{(J / MR_c ²) +1} and the off-diagonal component is J / MR_c ²
Where A is the matrix of
The physical quantity φ related to the change of the quantity f and the slip speed is

[0028]

[Equation 7]

The motion state approximated by
A matrix in which the gradient of the braking torque for each wheel, which is a parameter to be identified, is included in the diagonal component and the non-diagonal component is 0, is converted into a relational expression of K.phi. = F in advance and detected. Wheel deceleration time series data y _i [j] (j = 1,2,3, ...) And detected brake torque time series data T _bi [j] (j =
1,2,3, .....) is applied to each of the above relational expressions in sequence, and an online system identification method is applied to each data to estimate the braking torque gradient for each wheel. And

The invention of claim 7 is from claim 1 to claim 6.
The control means according to any one of the items 1 to 5 uses the slip speed at which the braking torque generated in each wheel is the maximum as an equilibrium point of each wheel, and the operation amount of the braking force acting around the braking torque and the equilibrium point is The motion state of each wheel when acting on each wheel, the motion state of the vehicle body when the braking torque generated in each wheel acts on the entire vehicle body, and each wheel against the disturbance of the slip speed of each wheel around the equilibrium point A non-linear variation of the braking torque of the vehicle with a linear variation that fluctuates within a first range with respect to the disturbance of the slip speed of each wheel, and the disturbance of the slip speed of each wheel around the equilibrium point A second model in which the nonlinear variation of the braking torque gradient of each wheel is represented by a linear variation that fluctuates within the second range with respect to the disturbance of the slip speed of each wheel, based on the first range. as well as The torque gradient estimating means estimates the braking torque gradient of the second model designed such that the second range is within the predetermined allowable range and the second range is within the predetermined allowable range. An operation amount of the braking force of each wheel that matches the gradient of the braking torque is calculated, and the braking force acting on each wheel is controlled based on the calculated operation amount of the braking force of each wheel. And (Solution Means of Torque Gradient Estimating Device of the Present Invention) The invention of claim 8 is a wheel speed detecting means for detecting a wheel speed at every predetermined sample time, and a time series of wheel speeds detected by the wheel speed detecting means. Only data is detected
With Te, and output means for outputting the torque gradient estimating means to estimate the gradient of the braking torque or driving torque with respect to the slip speed, the estimated value of the slope of the estimated braking torque or the driving torque by the torque gradient estimating means, It is configured to include.

According to a ninth aspect of the present invention, the torque gradient estimating means of the eighth aspect calculates a physical quantity related to a change in wheel speed and a physical quantity related to a change in wheel speed based on the detected time-series data of the wheel speed. Based on the first computing means for computing, the physical quantity relating to the change in the wheel speed and the physical quantity relating to the change in the wheel speed calculated by the first computing means, the history of the physical quantity relating to the change in the wheel speed and the wheel speed Second calculation means for calculating a physical quantity representing a history of the physical quantity related to the change and estimating the gradient of the braking torque or the driving torque from the physical quantity is provided. (Means for Solving the Braking Torque Gradient Estimating Device of the Present Invention) The invention of claim 10 is the time series data of wheel deceleration detected at every predetermined sample time, and the brake torque detected at every predetermined sample time, or Based on time-series data of the physical quantity related to the brake torque, a torque gradient estimating means for estimating the gradient of the braking torque with respect to the slip speed, and an estimated value of the gradient of the braking torque estimated by the torque gradient estimating means are output. In a braking torque gradient estimating device having an output means, the torque gradient estimating means determines a braking torque and a motion state of a wheel when the braking torque acts on the braking torque gradient estimating means according to a gradient of the braking torque with respect to a slip speed. And a linear model that changes in a linear function. The slope of the braking torque with respect to the slip speed, the relationship between the physical quantity relating to the change of the braking torque and the physical quantity relating to the change of the slip speed, which are respectively represented by the braking torque and the wheel deceleration, are converted in advance and detected. Slips by applying an online system identification method to each data in which the time series data of the wheel deceleration and the detected brake torque or the time series data of the physical quantity related to the brake torque are sequentially applied to the relationship. It is characterized in that the gradient of the braking torque with respect to the speed is estimated.

The invention of claim 11 is based on the invention of claim 10.
The torque gradient estimating means is
The time series data of wheel deceleration at sample time j is y
_i[J], the time series data of the brake torque is T
_bi[J], the predetermined sample time is τ, and the wheel inertia is
J, wheel radius is R_c, The vehicle mass is M, and the
T is a vector that has time series data of torque
_b[J], time series data of wheel deceleration of each wheel
Y [j], the identity matrix is I, and the diagonal component is
{(J / MR_c ²) +1} and the off-diagonal component is J / MR_c ²
Where A is the matrix of
The physical quantity φ related to the change of the quantity f and the slip speed is

[0033]

[Equation 8]

The motion state approximated by
A matrix in which the gradient of the braking torque for each wheel, which is a parameter to be identified, is included in the diagonal component and the non-diagonal component is 0, is converted into a relational expression of K.phi. = F in advance and detected. Wheel deceleration time series data y _i [j] (j = 1,2,3, ...) And detected brake torque time series data T _bi [j] (j =
1,2,3, .....) is applied to each of the above relational expressions in sequence, and an online system identification method is applied to each data to estimate the braking torque gradient for each wheel. And

( Principle of Antilock Brake Control of Claims 1 to 7) The braking force acts on the road surface via the surface of the tread of the tire that is in contact with the road surface. It acts on the vehicle body as a reaction force (braking torque) from the road surface through the frictional force between the wheels. When the vehicle body is traveling at a certain speed, slipping occurs between the wheels and the road surface when the braking force is applied.The braking torque that acts as a reaction force from the road surface at this time is expressed by the following equation. The slip speed ω s (converted into angular speed) changes as shown in FIG.

Ω _s = ω _v −ω _i where ω _v is the vehicle speed (equivalently expressed as an angular velocity) and ω _i is the i-th wheel (i is the wheel number, i = 1,2,3 ,. ....
) Is the wheel speed converted into angular velocity.

As shown in FIG. 5, the braking torque initially increases with the increase of the slip speed, reaches the maximum value f _i0 at the slip speed ω ₀ , and decreases with the increase of the slip speed at the slip speed higher than ω _0. . The slip speed ω ₀ corresponds to the slip speed when the friction coefficient between the wheel and the road surface has the maximum value (peak μ; corresponding to peak μ in FIG. 17).

[0038] Therefore, as is clear from FIG. 5, the gradient of the braking torque with respect to the slip speed (hereinafter referred to as "braking torque gradient") is, omega _s <positive at _{ω 0 (> 0), ω} s = ω 0
0, and negative (<0) when ω _s > ω ₀ . That is, when the braking torque gradient is positive, the wheels are gripping on the road surface, when the braking torque gradient is 0, the peak μ is reached, and when the braking torque gradient is negative, the wheels are locked. Moreover, the dynamic characteristics of the wheel motion change according to the braking torque gradient.

According to the present invention, the vehicle body speed is not estimated, the current braking torque gradient is estimated only from the time series data of the wheel speed, and the estimated braking torque gradient is within a predetermined range including the reference value. The braking force acting on the wheels is controlled so that the value becomes.

In the inventions of claims 5 and 6,
Without estimating the vehicle body speed, the current braking torque gradient is estimated from the wheel deceleration time series data and the brake torque time series data so that the estimated braking torque gradient falls within a predetermined range including the reference value. Controls the braking force that acts on the wheels. Instead of the brake torque, a physical quantity related to the brake torque, such as a wheel cylinder pressure, may be used.

Therefore, according to the present invention, the state of wheel motion corresponding to the braking torque gradient in the predetermined range including the reference value can be maintained. Further, if the reference value is set near 0, which corresponds to the peak μ, even if the slip speed that reaches the peak μ changes depending on the road surface state on which the vehicle is traveling, the braking torque gradient remains near 0 at the peak μ. Therefore, if the braking torque gradient is controlled to be near 0, the peak μ
Follow-up becomes possible. Further, since the vehicle body speed estimation unit is unnecessary, it is not necessary to repeatedly increase and decrease the braking force, and stable traveling is possible. (Principle of Estimating Braking Torque or Driving Torque Gradient of the Inventions of Claims 1 to 4, 8 and 9) The wheel motion and vehicle motion of each wheel are described by the following equation of motion. In the following, it is assumed that the number of wheels is four,
The present invention is not limited to this.

[0042]

[Equation 9]

Where F _i 'is the braking force generated on the i-th wheel, T _bi is the brake torque applied to the i-th wheel in response to the pedaling force, M is the vehicle mass, R _c is the effective radius of the wheel, J is the wheel inertia, and v is the vehicle speed (see FIG. 11). In addition,
Indicates the derivative with respect to time. In equation (1) and equation (2),
F _i ′ is shown as a function of slip speed (v / R _c −ω _i ).

Here, the vehicle body speed is represented by an equivalent angular velocity ω _v of the vehicle body, and the braking torque R _c F _i 'is described as a linear function of the slip velocity (slope k _i , y intercept T _i ).

[0045]       v = R_cω_v                                           (3)       R_cF_i’(Ω_v−ω_i) = K_i× (ω_v−ω_i) + T_i       (Four) Furthermore, by substituting Eqs. (3) and (4) into Eqs. (1) and (2), the wheel speed
Degree ω_iAnd vehicle speed ω _vDiscretized at every sample time τ
Time series data ω_i[k], ω_v[k] (k is sample
Sample time in units of τ, k = 1,2, .....)
We get the following expression.

[0046]

[Equation 10]

Here, if the equations (5) and (6) are combined and the equivalent angular velocity ω _{v of the} vehicle body is deleted,

[0048]

[Equation 11]

To obtain By the way, slip speed 3rad / s
Assuming that the maximum braking torque of R _c Mg / 4 (g is gravitational acceleration) is generated under the following conditions,

[0050]

[Equation 12]

To obtain. Here, as a concrete constant, τ
= 0.005 (sec), R _c = 0.3 (m), and M = 1000 (kg), max (k _i ) = 245. Therefore,

[0052]

[Equation 13]

Then, the equation (7) can be approximated by the following equation.

[0054]

[Equation 14]

It is By organizing like this,
The equation (8) can be described in a linear form with respect to the unknown coefficients k _i and f _i. By applying the online parameter identification method to the equation (8), the braking torque gradient k _i with respect to the slip speed can be calculated. Can be estimated.

Not only when the braking torque is acting, but also when the driving torque is acting, by applying the online system identification method to the equation (8), the driving for the slip speed is similarly performed. A torque gradient (hereinafter referred to as "driving torque gradient") can be obtained.

For example, time series data of detected wheel speeds are obtained by repeating the following step 1 and the following step 2 derived based on the least squares method which is one of the online system identification methods. The braking torque gradient or the driving torque gradient can be estimated from ω _i [k].

[0058]

[Equation 15]

Let us say that. Note that the first element of the matrix φ _i [k] in equation (9) is a physical quantity related to the change in wheel speed at one sample time, and equation (10) is one sample of the change in wheel speed at one sample time. It is a physical quantity related to changes over time.
This shows that Eq. (8) is the equation of motion for wheel (deceleration) motion, and it is clear that the braking torque gradient is proportional to the characteristic root that expresses the dynamic characteristics of wheel deceleration. . That is, the identification of the braking torque gradient can be interpreted as the identification of the characteristic root of the wheel (deceleration) motion.

[0060]

[Equation 16]

[0061] As the gradient of. Here, λ is a forgetting factor (for example, λ = 0.98) indicating the degree of removing the past data, and “ ^T ” indicates the transposition of the matrix.

Note that θ _{i on} the left side of the equation (11) is a physical quantity representing a history of physical quantities relating to changes in wheel speed and a history of physical quantities relating to changes in wheel speed. (Principle of estimation of braking torque gradient in claims 5, 6, 10, and 11) (1), (2) Formula, braking torque F _i (= F _i '· R _c ), angular velocity conversion Vehicle speed ω _v (= v
/ R _c ),

[0063]

[Equation 17]

It becomes Further, from the equation (12), the wheel deceleration y _i (= −dω _i / dt) of the _i -th wheel is

[0065]

[Equation 18]

It is represented as Here, by replacing the slip speed (ω _v −ω _i ) of the i-th wheel with x _i and rearranging equations (12) to (14),

[0067]

[Formula 19]

It becomes Here, the braking torque F _{i of} the i-th wheel
Is a non-linear function of the slip speed (see FIG. 5), the braking torque F near a certain slip speed x _i
(X _i ) is approximated by a straight line as in the following equation. That is, a gradient model in which the braking torque F (x _i ) changes in a linear function according to the braking torque gradient k _i with respect to the slip speed x _i is applied.

F _i (x _i ) = k _i x _i + μ _i (17) Here, regarding the i-th wheel (i = 1,2,3,4), the time series data of the slip speed is x _i [j]. , T _bi [j] for time series data of brake torque, and y for time series data of wheel deceleration.
_{Let i} [j] (j = 0,1,2, ...). However, each time series data is assumed to be sampled at every predetermined sampling time τ.

Substituting the equation (17) into the equations (15) and (16) and discretizing the obtained equation using the time series data for each sampling time τ,

[0071]

[Equation 20]

It becomes That is, x [j], y [j],
T _b [j] is a vector that has slip speed, wheel deceleration, and brake torque for each wheel in each component.

From the equation (19), the wheel deceleration y [j + 1] after one sample is

[0074]

[Equation 21]

It becomes From equations (19) and (20), K · (x [j + 1] −x [j]) = − J (y [j + 1] −y [j]) + T _b [j + 1] −T _b [j] (21 ) Is obtained.

[0076] In (21), φ = x [j + 1 ] -x [j] (22) f = -J (y [j + 1] -y [j]) + T b [j + 1] -T b [j] ( Putting it as 23), K · φ = f (24).

Here, considering the meaning of φ, it can be seen that it indicates the difference in slip speed between adjacent samples, that is, the physical quantity relating to the change in slip speed.

The equations (18) and (19) are combined into (Kx [j] +
If the terms of μ) are deleted and rearranged, from equation (22),

[0079]

[Equation 22]

Is obtained. Further, the time series data of the braking torque is F [j] (the time series data F of the braking torque of the i-th wheel F
_{When the} equation (14) is discretized and arranged as a vector having _i [j] as a component, F [j] = − Jy [j] + T _b [j] (26) is obtained.

Applying the equation (26) to the equation (23),     f = F [j + 1] -F [j] (27) Becomes

From equation (27), it can be seen that f represents the difference in braking torque between adjacent samples, that is, the physical quantity relating to the change in braking torque.

From the above, the motion state of the wheel shown in the equations (12) to (14) is approximated by the gradient model of the equation (17) as in the equations (18) and (19), and this approximation is performed. It was shown that the motion state can be converted into the relation of Eq. (24). That is, the wheel motion state is a physical quantity related to the change in the braking torque expressed by the equations (23) and (25) by the gradient of the braking torque with respect to the slip speed which is the parameter to be identified, and the braking torque and the wheel deceleration, respectively. And the relationship between physical quantities relating to changes in slip speed.

As a result, the number of identification parameters can be set to 1, and the calculation accuracy can be greatly improved and the calculation time can be shortened as compared with the above-mentioned conventional technique in which the number of identification parameters is 3.

Here, when the equation (24) is shown for the i-th wheel, k _i · φ _i = f _i (28) However, f and φ in the equation (24) are set as f = [f ₁ f ₂ f ₃ f ₄ ] ^T φ = [φ ₁ φ ₂ φ ₃ φ ₄ ] ^T.

In the present invention, based on the time-series data y _i [j] of the wheel deceleration of the i-th wheel and the time-series data T _bi [j] of the braking torque of the i-th wheel, f _i , φ of the i-th wheel are calculated. _i to (2
3), by applying the online system identification method to each data obtained by substituting the calculated f _i , φ _i into Eq. The braking torque gradient k _i can be estimated and calculated. (Principle of ABS control of invention of claim 7) In the inventions of claims 1 to 6, there is a certain reference value (complete peak μ for the braking torque gradient with respect to the slip speed estimated as described above).
When the follow-up is performed, the ABS control is performed so as to follow 0). The control system for feedback-controlling the control torque gradient may be designed for each wheel by PID control or the like, but it is also possible to systematically design it as an integrated system of four wheels by applying modern control theory. in this case,
Since the interference of the four wheels is also taken into consideration in the design, finer control can be realized.

By the way, the ABS control system is a system having a non-linear characteristic having a strong tire characteristic, and the modern control theory cannot simply be applied. Therefore, in the invention of claim 6, the non-linear characteristic can be regarded as an apparently equivalent plant variation, and a control system design that allows this plant variation is one of the modern control theories. This was achieved by applying the robust control theory, and a detailed control system design was performed in consideration of the interference of the four wheels. The details of the control system design are described below.

The brake torque T _bi 'corresponding to the pedaling force when the brake pedal is depressed just before the wheel is locked acts on the wheel, and in this state, the brake torque T _bi ′ is set so as to follow the peak μ without locking the wheel. The wheel motion and the vehicle body motion of each wheel when (operation amount) u _bi acts are described as follows from the equations (12) and (13).

[0089]

[Equation 23]

However, the equation (31) is an output equation showing that the braking torque gradient k _i of each wheel is a function of the slip speed.

By the way, F _i and G _i are as shown in FIG.
As shown in (b) respectively, at ωο peak and 0 respectively
Which are non-linear functions of the slip velocity, which can be represented by the forms of straight lines 20 and 23 shown by solid lines and fluctuations within a predetermined range. Here, assuming that the disturbance of the slip velocity from ωο is x _i , F _i = (f _i + W _{fi Δfi} ) x _i + f _i0 (32) G _i = (g _i + W _{gi Δgi} ) x _i (33) Can be represented.

Here, f _i represents the inclination of the straight line 20 in FIG. 6A, and g _i represents the inclination of the straight line 23 in FIG. 6B. Also,
W _fi and W _gi are weighting factors for standardizing the fluctuation,
The broken line 21 and the broken line 22 in FIG. 6A and the broken lines 24 and 25 in FIG.
_This corresponds to setting _fi and _Δgi to ± 1.

That is, the equation (32) shows the nonlinear fluctuation of the braking torque of each wheel with respect to the disturbance x _i around the equilibrium point ω ₀ as shown in FIG.
It is a linear model represented by a variation within a range from a broken line 21 to a broken line 22 including the straight line 20 of (a). Further, the equation (33) shows the nonlinear variation of the braking torque gradient of each wheel with respect to the disturbance x _i around the equilibrium point ω ₀ by the broken line 2 including the straight line 23 of FIG. 6 (b).
It is a linear model represented by the variation within the range of 4 to the broken line 25.

Furthermore, equations (32) and (33) are converted into equations (29), (30) and (31).
Substituting into the equation and describing it as a state equation around the equilibrium point (ωο), we obtain the following equation.

[0095]

[Equation 24]

Also,

[0097]

[Equation 25]

It is Here, x is the wheel slip speed disturbance around omega _0, y braking torque gradient of each wheel around omega _0, u is equivalent to u _bi of each wheel operation amount ((29) around omega ₀ ) Is represented.

Here, an arbitrary Δ (-having the structure of equation (36) is
By designing a control system that allows 1 ≦ Δ _fi and Δ _gi ≦ 1), it is possible to design an ABS control system that takes into account the interference of four wheels. This design is a robust control technique μ
This can be easily done by applying the design method.

That is, an arbitrary Δ having the structure of equation (36)
The following controller is derived by designing a control system that allows (−1 ≦ Δ _fi , Δ _gi ≦ 1) using the so-called μ design method.

[0101]

[Equation 26]

Where xc is the controller state, Ac
, Bc, Cc, Dc represent the coefficient matrix of the designed controller, and y represents the braking torque gradient of the designed control system. Then, by substituting the state value of the controller into xc of the equation (39) and the estimated braking torque gradient value into y of the equation, the manipulated variable u of the ABS control is obtained. (Principle of the Invention to be Referenced ) A vibration phenomenon at a wheel when a vehicle equipped with a vehicle body having a weight Wv is traveling at a vehicle body speed ωv, that is, a vibration phenomenon of a vibration system constituted by the vehicle body, the wheel, and the road surface is described. , Consideration will be made with reference to the model shown in FIG. 12, which is equivalently modeled by the wheel rotation axis.

In the model of FIG. 12, the braking force is
Although it acts on the road surface via the surface of the tread 115 of the tire that is in contact with the road surface, this braking force actually acts on the vehicle body as a reaction (braking force) from the road surface. The wheel 1 through the friction element 116 (road surface μ) between the tire tread and the road surface.
It will be connected to the opposite side of 13. This is similar to the case where a chassis dynamo device can simulate the weight of a vehicle body with a large inertia under a wheel, that is, a mass on the side opposite to the wheel.

In FIG. 12, the inertia of the wheel 113 including the tire rim is J _w , the spring element 1 between the rim and the tread 115.
The spring constant of 14 is K, the wheel radius is R, the inertia of the tread 115 is J _t , and the friction element 1 between the tread 115 and the road surface is 1.
Assuming that the friction coefficient of 16 is μ and the inertia of the equivalent model 117 of the vehicle body weight W _v converted to the rotation axis is J _V , the brake torque T _b 'generated by the wheel cylinder pressure is used to determine the wheel speed ω _w.
The transfer characteristics up to

[0105]

[Equation 27]

It becomes Note that s is a Laplace transform operator. When the tire grips the road surface, assuming that the tread 115 and the vehicle body equivalent model 117 are directly connected, the sum inertia of the vehicle body equivalent model 117 and the tread 115 and the inertia of the wheel 113 resonate. . That is, this vibration system can be regarded as a wheel resonance system including wheels, a vehicle body, and a road surface. The resonance frequency ω ∞ of the wheel resonance system at this time is expressed by ω ∞ = √ {(J _w + J _t + J _v ) K / J _w (J _t + J _v )} / 2π (41 ). In FIG. 17, this state corresponds to the area A1 before shifting to the vicinity of the peak μ.

On the other hand, when the friction coefficient μ of the tire approaches the peak μ, the friction coefficient μ of the tire surface becomes difficult to change with respect to the slip ratio, and the component accompanying the vibration of the inertia of the tread 115 is the vehicle body equivalent model. 117 is no longer affected. That is, the tread 115 and the vehicle body equivalent model 117 are equivalently separated, and the tread 115 and the wheel 113 resonate. It can be considered that the wheel resonance system at this time is composed of the wheels and the road surface, and the resonance frequency ω ∞ ′ is equal to the vehicle body equivalent inertia J _v in the equation (41).
It will be equal to That is, ω ∞ ′ = √ {(J _w + J _t ) K / J _w J _t )} / 2π (42). This state corresponds to the area A2 near the peak μ in FIG.

Comparing equations (41) and (42), the body equivalent inertia J
_vIs the wheel inertia J_w, Tread inertia J _tAssume greater than
Then, the resonance frequency ω∞ ′ of the wheel resonance system in the case of the formula (42)
Is to shift to the higher frequency side than ω∞ than Eq. (41)
become. Therefore, the change in the resonance frequency of the wheel resonance system is reflected.
Determine the limit of braking torque characteristics based on the physical quantity
It is possible to

Therefore, in the present invention, the following minute gain G _d is introduced as a limit determination amount as a physical amount that reflects such a change in resonance frequency.

First, the micro-excitation means of the present invention uses the resonance frequency ω ∞ (equation (41)) of the vibration system including the wheels, the vehicle body, and the road surface.
When the braking force (here, the braking pressure P _b ) is minutely excited, the wheel speed ω _w also minutely vibrates around the average wheel speed at the resonance frequency ω ∞.

Here, the limit determining means of the present invention, when the minute amplitude of the resonance frequency ω ∞ of the brake pressure P _b at this time is P _v and the minute amplitude of the resonance frequency ω ∞ of the wheel speed is ω _wv , The minute gain G _d is calculated as G _d = ω _wv / P _v (43). Note that this minute gain G _d is regarded as a vibration component of the resonance frequency ω ∞ of the ratio (ω _w / P _b ) of the wheel speed ω _w to the brake pressure P _b , and G _d = ((ω _w / P _b ) | It can also be expressed as s = jω∞) (44).

This minute gain G_dAs shown in equation (44)
To (ω_w/ P_b) Of the resonance frequency ω ∞ of
Then, the wheel motion reached the limit region A2 of the braking torque characteristic.
Then, the resonance frequency shifts to ω∞ '
To do. Therefore, the small gain G _dMoved to the limit area A2
Reference gain G preset as a value when_sAnd minute
Gain G_dAnd a small gain G_dIs the reference gain G
_sIt is judged as the limit of the braking torque characteristic when
be able to.

Next, it will be explained that the minute gain G _d is a physical quantity equivalent to the braking torque gradient.

As shown in FIG. 13, the slip speed Δω
It is known that a functional relationship in which the friction coefficient μ has a peak at a certain slip speed is established between and the wheel-road surface friction coefficient μ, similarly to the relationship in FIG. The friction characteristic of FIG. 13 corresponds to the braking torque characteristic of FIG.

By the way, when the brake pressure is slightly excited by the minute excitation means, the wheel speed is minutely excited, so that the slip ratio also vibrates slightly around the slip ratio. Now, let us consider a change in the friction coefficient μ with respect to the slip speed Δω when the vehicle vibrates slightly around a certain slip ratio on the road surface having the characteristics shown in FIG.

At this time, the friction coefficient μ of the road surface can be approximated as μ = μ ₀ + αRΔω (45). That is, since the change of the slip speed due to the minute vibration is small, it can be approximated by the straight line of the inclination αR.

Here, by substituting the equation (45) for the braking torque T _b = μWR generated by the friction coefficient μ between the tire and the road surface, T _b = μWR = μ ₀ WR + αR ² ΔωW (46) Here, W is the wheel load. Δ on both sides of equation (46)
Differentiating the first order with ω,

[0118]

[Equation 28]

To obtain Therefore, the expression (47) shows that the braking torque gradient (dT _b / Δω) is equal to αR ² W.

On the other hand, since the brake torque T _b ′ is proportional to the brake pressure P _b , the minute gain G _d is the ratio of the wheel speed ω _{w to} the brake torque T _b ′ (ω _w /
It has a proportional relationship with the vibration component of the resonance frequency ω ∞ of T _b '). Therefore, due to the transfer characteristics of equation (40), a small gain G _d
Is represented by the following equation.

[0121]

[Equation 29]

Generally, in the equation (50),       │A│ = 0.012 << │B│ = 0.1 (51) Therefore, from equations (47) and (48),

[0123]

[Equation 30]

To obtain That is, the gradient of the braking torque T _b with respect to the slip speed Δω is proportional to the minute gain G _d .

The [0125] above, micro-gain G _d is shown to be braking torque gradient equivalent physical quantity, the braking torque gradient can be seen that it is possible to estimate the on the basis of the micro-gain G _d. Since the minute gain G _d is a parameter that sensitively reflects the vibration characteristics that fluctuate depending on the frictional state between the wheel and the road surface, the braking torque gradient can be estimated extremely accurately regardless of the road surface state.

[0126]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The ABS control device according to each embodiment of the present invention will be described in detail below with reference to the drawings. (First Embodiment) FIG. 1 shows the configuration of an ABS control apparatus according to a first embodiment of the present invention.

As shown in FIG. 1, the ABS control apparatus according to the first embodiment detects a wheel speed detecting means 10 for detecting a wheel speed at every predetermined sample time τ, and the wheel speed detecting means 10 detects the wheel speed. Based on the braking torque gradient estimated by the torque gradient estimating means 12, a torque gradient estimating means 12 for estimating a braking torque gradient from the time series data of the wheel speeds thus obtained, and an operation signal for each wheel for ABS control are provided. An ABS control means 14 for calculating and an ABS control valve 16 for performing ABS control by operating a brake pressure for each wheel based on an operation signal calculated by the ABS control means 14.
It consists of and. The wheel speed detecting means 10 and the torque gradient estimating means 12 among them constitute a braking torque gradient estimating device 8 which outputs the estimated braking torque gradient value.

The wheel speed detecting means 10 of FIG. 1 can be realized by the structure of FIG. 7A, for example. As shown in FIG. 7A, the wheel speed detecting means 10 includes a signal rotor 30 having a predetermined number of teeth cut at equal intervals and mounted so as to rotate with the wheel, and a pickup coil 32 fixed to the vehicle body. And a permanent magnet 34 arranged to penetrate a magnetic flux inside the pickup coil 32, and a frequency of an AC voltage generated in the pickup coil 32 that is connected to the pickup coil 32 and is detected at each sampling time τ. And a frequency detector 36 for outputting the output.

When the signal rotor 30 rotates as the wheels rotate, the signal rotor 30 and the pickup coil 3 rotate.
The air gap between the two changes at a cycle according to the rotation speed. Therefore, the magnetic flux of the permanent magnet 34 penetrating the pickup coil 32 changes, and an AC voltage is generated in the pickup coil 32. Here, FIG. 7B shows a temporal change in the AC voltage generated in the pickup coil 32.

As shown in FIG. 7B, the AC voltage generated in the pickup coil 32 has a low frequency when the rotation speed of the signal rotor 30 is low and a high frequency when the rotation speed of the signal rotor 30 is high. Since the frequency of this AC voltage is proportional to the rotation speed of the signal rotor 30, that is, the wheel speed, the output signal of the frequency detector 36 is proportional to the wheel speed for each sample time τ.

The wheel speed detecting means 10 of FIG. 7 (a) is attached to all the first to fourth wheels, and the i-th wheel (i is the wheel number) is output from the output signal of the frequency detector 36 for each wheel. ,
i = 1,2,3,4) wheel speed time series data ω _i [k] (k
Is the sample time; k = 1, 2, ...) Are detected.

Next, the structure of the ABS control valve 16 will be described with reference to FIG. As shown in FIG. 8, the ABS control valve 16
Is a control solenoid valve 132 for the right front wheel (hereinafter,
"Valve SFR"), control solenoid valve 134 for the left front wheel (hereinafter "valve SFL"), control solenoid valve 140 for the right rear wheel (hereinafter "valve SRR")
And the control solenoid valve 142 for the left rear wheel (hereinafter,
“Valve SRL”).

Valve SFR, valve SFL, valve SR
R and valve SRL are pressure increasing valves 132a and 132a, respectively.
34a, 140a, 142a and pressure reducing valve 132
b, 134b, 140b, 142b, and are connected to the front wheel cylinders 144, 146 and the rear wheel cylinders 148, 150, respectively.

Pressure increasing valves 132a, 134a, 140
a, 142a and pressure reducing valves 132b, 134b, 1
Reference numerals 40b and 142b denote an SFR controller 131 and an SFL controller 13 that control opening and closing of valves, respectively.
3, SRR controller 139, SRL controller 1
41 is connected.

The SFR controller 131, the SFL controller 133, the SRR controller 139, and the SRL controller 141, based on the operation signal for each wheel sent from the ABS control means 14, increase the pressure of the control solenoid valve and decrease the pressure of the control solenoid valve. Controls opening and closing of valves.

The structure of the system hydraulic circuit including the ABS control valve 16 will be described in detail with reference to FIG.

As shown in FIG. 9, the system hydraulic circuit is provided with a reservoir 100 for storing the brake fluid of the master cylinder system and the power supply system. The reservoir 100 is provided with a level warning switch 102 for detecting a decrease in the liquid level of the brake fluid stored inside, and a relief valve 104 for relieving the brake fluid to the reservoir 100 when the power supply system has an abnormally high pressure. There is.

Further, a pipe arranged from the relief valve 104 side of the reservoir 100 is provided with a pump 106 for drawing up brake fluid from the reservoir 100 and discharging high hydraulic fluid, and further on the fluid discharge side, a pump 106 is provided. Hydraulic pressure generated by pump (power supply system)
An accumulator 108 for accumulating pressure and a pressure sensor 110 for detecting the oil pressure of the accumulator 108 are provided. This pressure sensor 110 outputs a control signal for the pump 106 based on the hydraulic pressure of the accumulator 110, and outputs a warning signal (ABS, T when the pressure is low).
The RC control prohibition signal) is output.

Further, the control signal of the pump 106 is output to the piping on the high hydraulic side of the accumulator 108 when the hydraulic pressure of the accumulator 110 is low, and the warning signal (ABS, TRC control prohibition signal) is output when the hydraulic pressure is low. A pressure switch 112 is provided.

Further, a master cylinder 114 for generating a hydraulic pressure according to the pedal effort applied to the brake pedal 118 is connected to another pipe extending from the reservoir 100. A brake booster 116 is arranged between the master cylinder 114 and the brake pedal 118 to adjust / introduce the high hydraulic pressure of the accumulator 110 into a hydraulic pressure corresponding to the pedaling force to generate a brake assisting force.

The brake booster 116 is connected to a pipe on the high hydraulic side of the accumulator and a pipe extending directly from the reservoir 100. When the depression amount of the brake pedal 118 is less than a certain value, the reservoir 100 From the accumulator 108 is introduced when the depression amount exceeds a certain value.

Further, a front master pressure pipe 164 and a rear master pressure pipe 166 are respectively provided from the master cylinder 114 for supplying the hydraulic pressure (master pressure) of the master cylinder to the front and rear wheels. Then, the front master pressure pipe 164 and the rear master pressure pipe 16
A P & B valve 120 for adjusting the brake hydraulic pressure of the rear system is interposed at 6 so as to properly distribute the braking force to the front and rear wheels. The P & B valve 120 stops the pressure regulation of the rear system when the front system is lost.

Further, in the front master pressure pipe 164 extending from the P & B valve 120, when the oil pressure of the power supply system decreases, the pressure intensifying device 122 for increasing the front wheel cylinder oil pressure to secure a high braking force. Is provided. A booster pipe 168 connected to the booster chamber of the brake booster 116 is connected to the pressure booster 122.
8 and the pressure intensifier 122, a pressure limiter 124
And the differential pressure switch 126 is interposed.

The pressure limiter 124 closes the path to the booster chamber so that the pressure booster 122 and the differential pressure switch 126 are not operated when an input exceeding the assisting force limit of the brake booster 116 is applied when the system is normal. Further, the differential pressure switch 126 detects a hydraulic pressure difference between the master cylinder 114 and the booster chamber.

In the booster pipe 168, the control solenoid valve 132 ("valve SF" for the right front wheel described above is used.
The R ”) pressure increasing valve 132a is connected to the pressure increasing valve 134a of the left front wheel control solenoid valve 134 (“ valve SFL ”). Further, the pressure reducing valve 132b of the valve SFR and the pressure reducing valve 1 of the valve SFL.
A low-pressure pipe 162 extending directly from the reservoir 100 is connected to 34b.

A switching solenoid valve 136 (hereinafter "valve SA1") and a switching solenoid valve 138 (hereinafter "valve SA2") are connected to the pressure supply side pipes of the valve SFR and the valve SFL, respectively. The pressure-increasing side pipe of the pressure-increasing device 122 is further connected to the valves SA1 and SA2. And
The pipe on the pressure supply side of the valve SA1 is connected to the front wheel cylinder 144 that applies brake pressure to the brake disc 152 of the left front wheel, and the valve SA2 is
It is connected to a front wheel cylinder 146 that applies brake pressure to the brake disc 154 of the right front wheel.

The valves SA1 and SA2 have the pressure from the pressure intensifying device 122 in the normal brake mode.
The valves are switched so as to be applied to the front wheel cylinders 144 and 146, respectively, and in the ABS control mode, the valves are switched so that the pressures from the valve SFR and the valve SFL are applied to the front wheel cylinders 144 and 146, respectively. That is, for the front wheels, the switching between the normal brake mode and the ABS control mode can be independently performed for each of the left and right wheels.

The control solenoid valve 140 for the right rear wheel described above is connected to the booster pipe 168 via a switching solenoid valve 130 (hereinafter, "SA3").
("Valve SRR") pressure increasing side valve 140a and left rear wheel control solenoid valve 142 ("Valve SR"
L ") pressure increasing valve 140b is connected. Further, the pressure reducing valve 140b and the valve SR of the valve SRR
A low pressure pipe 162 extending directly from the reservoir 100 is connected to the L pressure reducing valve 142b.

The pipe on the pressure supply side of the valve SRR is connected to a rear wheel cylinder 148 which applies a brake pressure to the brake disc 156 of the right rear wheel, and the valve S
The RL is connected to a rear wheel cylinder 150 that applies brake pressure to a brake disc 158 for the left rear wheel.

The valve SA3 switches the valve so that the master pressure from the rear master pressure pipe 166 is applied to the valve SRL and the valve SRR in the normal brake mode, and the high hydraulic pressure of the booster pipe 168 is set in the ABS control mode. The valves are switched so as to cover the valves SRL and SRR. That is, for the rear wheels, the switching between the normal brake mode and the ABS control mode is performed together on the left and right.

Next, the operation of this embodiment will be described. In the ABS mode, the valves SA1 and SA2 in FIG. 9 close the valves on the pressure booster 122 side and open the valves SFR and SFL side. The valve SA3 closes the valve on the rear master pressure pipe 166 side and opens the valve on the booster pipe 168 side.

First, the wheel speed detecting means 10 detects the wheel speed of each of the first to fourth wheels at each sample time τ and outputs the time series data ω _i [k] of the wheel speed of each wheel. To do.

Next, the torque gradient estimating means 12 calculates the equations (9) and (10) based on ω _i [k] in the above step 1, and then, for example, an online system identification method. The braking torque gradient is estimated from the recurrence formula of the formula (11) in the above step 2 derived based on the least squares method. By repeating step 1 and step 2 sequentially, time series data of the estimated braking torque gradient is obtained.

Then, the ABS control means 14 performs the processing according to the flow of the flowchart of FIG. As shown in FIG. 10, the ABS control unit 14 uses the braking torque gradient at each sample time estimated by the torque gradient estimating unit 12 to manipulate the operation amount u (u _i : i = 1 of each wheel at each sample time).
2, 3, 4) are calculated (step 200).

That is, from equations (29) to (33) to equation (34), (3
Deriving the equation of state of Eq. (5) and expressing any of the formulas (34) and (35), which has the structure of Eq. (36), any Δ (-1 ≤ △ _fi , △ _gi ≤
The controller of equations (38) and (39) is derived by designing a control system that allows 1) using the so-called μ design method. Then, by substituting the state value of the controller into x _c of the equation (39) and substituting the value of the braking torque gradient estimated by the torque gradient estimating means 12 into y of the equation, the ABS control valve 16
To obtain the manipulated variable u.

Next, the wheel number i is set to 1 (step 202), and it is determined whether or not the operation amount u _i of the i-th wheel is larger than the positive reference value + e (step 204). Manipulated variable u _i
Is larger than the positive reference value + e (Yes at Step 204), the operation signal of the ABS control valve of the i-th wheel is set to the pressure increase signal (Step 206).

When the manipulated variable u _i is not larger than the positive reference value + e (NO at step 204), it is determined whether the manipulated variable u _i is smaller than the negative reference value −e (step 20).
8). When the manipulated variable u _i is smaller than the negative reference value −e (Yes in step 208), the operation signal of the ABS control valve of the i-th wheel is set to the pressure reduction signal (step 210).

If the manipulated variable u _i is not smaller than the negative reference value −e (No at step 208), that is, if the manipulated variable u _i is greater than or equal to the negative reference value −e and is positive reference value + e.
In the following cases, the operation signal of the ABS control valve for the i-th wheel is
The hold signal is set (step 212).

When the operation signal for the operation amount u ₁ of the first wheel is set in this manner, the wheel number i is incremented by 1 (step 214), and it is then determined whether i exceeds 4 (step 214). Step 216). If i does not exceed 4 (No in step 216), step 204
Then, the operation signal is set for the operation amount u _i of the wheel number i which is similarly incremented.

When the wheel number i exceeds 4 (step 2
16 affirmative judgment), that is, A of all the first to fourth wheels
When the operation signal of the BS control valve is set, the set operation signal is sent to the ABS control valve 16 (step 21).
8). The operation signal setting and the operation signal transmission as described above are performed at each sample time.

When the operation signal for each wheel is transmitted in this way, in the ABS control valve 16, the SFR controller 131, the SFL controller 133, the SRR controller 139, and the SRL controller 141 in FIG. 8 respond to each operation signal. Valve SFR, valve SFL, valve SRR,
The opening / closing of the valve SRL is controlled.

That is, when the pressure increasing signal is received, the pressure increasing valve is opened and the pressure reducing valve is closed. As a result, the high hydraulic pressure of the booster pipe 168 of FIG. 9 is applied to the corresponding wheel cylinder to increase the braking force. On the contrary, when the signal is the pressure reducing signal, the pressure increasing side valve is closed and the pressure reducing side valve is opened. As a result, the low hydraulic pressure of the low pressure pipe 162 of FIG. 9 is applied to the corresponding wheel cylinder, and the braking force is reduced. Also,
When the signal is a hold signal, the pressure increasing valve and the pressure reducing valve are closed at the same time. As a result, the hydraulic pressure applied to the corresponding wheel cylinder is retained and the braking force is retained.

As described above, in the present embodiment, the braking torque gradient is estimated only from the time-series data of the wheel speed, and the ABS control is performed so that the braking torque gradient becomes zero.
The ABS control can be stably performed even if the slip speed at which the peak μ changes depending on the road surface state on which the vehicle is traveling.

Further, in the present embodiment, it is only necessary to identify two parameters, that is, a physical quantity history relating to a change in wheel speed and a physical quantity history relating to a change in wheel speed. Therefore, three parameters have to be identified. The calculation time can be reduced and the calculation accuracy can be improved as compared with the above-mentioned conventional technology (US patent) which does not occur. Therefore, highly accurate anti-lock brake control becomes possible.

Further, in this conventional technique, the wheel cylinder pressure must be detected in addition to the wheel speed, whereas in the present embodiment, only the wheel speed can be detected without using an expensive pressure sensor or the like. Therefore, the cost of the device can be reduced and the device can be simplified.

Further, in the present embodiment, since it is not necessary to estimate the vehicle body speed, the speed v _w obtained from the wheel speed and the actual vehicle body speed v _{v *} for estimating the vehicle body speed as in the conventional case _.
The wheel is locked for a long time when the braking force is repeatedly increased and decreased at a relatively low frequency until the value becomes close to or close to, or when the vehicle speed compared with the reference speed is significantly different from the actual vehicle speed. It is possible to avoid problems such as the braking force being extremely reduced due to falling or recovery, and it is possible to realize comfortable ABS control.

Furthermore, in the present embodiment, the modern control theory is not simply applied to a system having a strong non-linear characteristic of the tire, but the non-linear characteristic can be regarded as an apparently equivalent plant fluctuation. Focusing on what can be done, the ABS control system design that allows the plant fluctuation is achieved by applying the robust control theory, so that it is possible to realize the fine ABS control in consideration of the interference of four wheels.

The braking torque gradient estimating device 8 is not only applied to the ABS device as in the above example, but also a warning device for issuing a warning regarding braking to the driver according to the value of the estimated braking torque gradient. Etc. are also applicable.

The invention according to the present embodiment can be applied not only to the braking force but also to the driving force control device.
In this case, the braking torque gradient estimation device 8
Can estimate the drive torque gradient. (Second Embodiment) A according to the second embodiment
The BS controller will be described with reference to FIG. The same components as those in the first embodiment are designated by the same reference numerals and detailed description thereof will be omitted.

As shown in FIG. 2, the ABS control device according to the second embodiment has a brake torque detecting means 51 for detecting the brake torque T _b at every predetermined sample time τ.
, Wheel deceleration detecting means 52 for detecting wheel deceleration y at every predetermined sample time τ, and time series data T _b [j] (j = 1,2,3, ...) Of detected brake torque. ..) and wheel deceleration time-series data y [j] (j = 1,2,3, ..) based on the torque gradient estimating means 53 for estimating the braking torque gradient, and the estimated braking. ABS control means 15 for calculating an operation signal for each wheel for ABS control based on the torque gradient;
An ABS control valve 16 that performs ABS control by operating the brake pressure for each wheel based on the operation signal calculated by the BS control means 15. The braking torque detecting means 51, the wheel deceleration detecting means 52, and the torque gradient estimating means 53 among them output the estimated braking torque gradient value.
Configure 0.

Here, the brake torque detecting means 51 is
A pressure sensor that detects the wheel cylinder pressure of each wheel,
And a multiplier for calculating and outputting the brake torque of each wheel by multiplying the wheel cylinder pressure detected by the sensor by a predetermined constant.

Further, the wheel deceleration detecting means 52 is
Wheel speed sensor attached to (wheel speed detection means)
Wheel speed signal of the i-th wheel (i = 1,2,3,4) detected by
Issue ω _iBy reducing the wheel of the i-th wheel by
Speed y_iCan be realized as a filter for deriving.

[0173]

[Equation 31]

However, s is a Laplace transform operator. Note that the wheel deceleration detecting means 52 can also be configured by using a wheel deceleration sensor that directly detects the wheel deceleration regardless of the wheel speed.

The torque gradient estimating means 53 (23), based on the time series data y _i [j] of wheel deceleration of the i-th wheel and the time series data T _bi [j] of the brake torque of the i-th wheel, (2
F _i and φ _i of the i-th wheel are calculated by the equation (5), and the calculated f
_It is configured as an arithmetic unit for estimating and calculating the braking torque gradient k _i of the i-th wheel by applying, for example, an online system identification method to each data obtained by substituting _i and φ _i into the equation (28). be able to.

Next, the operation of the second embodiment will be described. First, the torque gradient estimating means 53 detects the i-th detected
Time series data of wheel brake torque T _bi [j] (j = 1,2,
3, ....) and time series data of wheel deceleration of the i-th wheel y
_{Based on i} [j] (j = 1,2,3, ...), the braking torque gradient of the i-th wheel is estimated and output for each wheel.

Next, the ABS control means 15 arithmetically outputs an operation signal for the i-th wheel so that the braking torque gradient of the i-th wheel does not become smaller than the reference value. And the ABS control valve 16
However, the brake pressure is controlled for each wheel based on this operation signal.

For example, when the calculated braking torque gradient becomes smaller than the reference value, the ABS control means 15 immediately outputs a brake pressure reduction command signal to the ABS control valve 16. Here, if this reference value is set to a positive value near 0, for example, the braking torque gradient is 0 in the region of the peak μ, so that when the brake is applied beyond the peak μ, the brake pressure is immediately increased. It is reduced to avoid tire locking.

Further, when the calculated braking torque gradient becomes larger than the reference value, the ABS control means 15 may immediately output the brake pressure increase command signal to the ABS control valve 16. Similarly to the above, when the reference value is set to a positive value near 0, when the area of μ is smaller than the peak μ, the braking force is immediately increased and returned to the vicinity of the peak μ. As a result, the most effective braking is possible, and the braking distance can be shortened.

As described above, according to the second embodiment, the control is performed based on the braking torque gradient obtained from the wheel deceleration and the braking torque. Therefore, as in the first embodiment,
It is not necessary to estimate the vehicle speed, and stable and comfortable anti-lock brake control can always be performed regardless of the road surface condition.

Further, in the second embodiment, since the method of directly identifying one parameter of the braking torque gradient is used, compared with the above-mentioned prior art (US patent) in which three parameters must be identified, the calculation is performed. It is possible to significantly reduce the time and greatly improve the calculation accuracy. Therefore, more accurate antilock brake control can be performed.

Furthermore, in the second embodiment as well, A
Instead of the BS control means 15, the ABS control means 14 according to the first embodiment can be applied. That is, it is possible to execute the processing of the flowchart of FIG. 10 in consideration of the interference of the four wheels. As a result, fine ABS control can be realized.

The braking torque gradient estimating device 50 is not only applied to the ABS device as in the above example, but also a warning device for issuing a warning regarding braking to the driver according to the value of the estimated braking torque gradient. Etc. are also applicable.

Further, the braking torque gradient estimating device 50 is
As shown in FIG. 3, it can be applied as a limit determination device 55 for determining the limit of the braking torque characteristic. The limit determining device 55 determines the limit of the braking torque characteristic based on the calculated braking torque gradient value at the output end of the braking torque gradient estimating device 50 of FIG.
Are connected.

Here, the braking torque characteristic means a characteristic of change in braking torque with respect to slip speed (see FIG. 5),
The limit of the braking torque characteristic means a boundary when the change characteristic changes from one state to another state. For example, in FIG. 5, the limit (point at which the braking torque gradient is a positive value close to 0) at the time of shifting from the region of slip speed smaller than the slip speed of peak μ to the region of slip speed near peak μ Can be mentioned. When detecting this limit, the limit determining unit 54 stores a positive value close to 0 as a reference value, and when the calculated braking torque gradient is equal to or greater than this reference value, determines that the limit is not reached, and the braking torque is determined. When the slope becomes smaller than the reference value, it is judged as the limit. Then, this limit determination result is output by an electric signal or the like.

The output signal of the limit judging device 55 may be inputted to the ABS control means 15 of FIG. In this case, when the braking torque characteristic is determined to be the limit, the ABS control means 15 outputs a brake pressure reduction command signal to the ABS control valve 16 to prevent tire lock.

Further, when the vehicle travels on a road surface where the control torque gradient changes rapidly near the peak μ, the servo control for following a certain target value may not function well. It is also possible to determine the limit at which the torque gradient changes abruptly, and perform control to change the target value of the control system according to the determination result. Thereby, better control performance can be obtained. Third Embodiment Next, as a third embodiment, a braking torque gradient estimating device as a reference invention will be described with reference to FIG. The same components as those in the first and second embodiments are designated by the same reference numerals and detailed description thereof will be omitted.

[0188] As shown in FIG. 4, the braking torque gradient estimating device 57 of the third embodiment, the micro-gain calculation section 22 for calculating a micro-gain G _d, computed fine gain G _d
And a braking torque gradient calculating unit 56 that performs a calculation for converting to a braking torque gradient.

Of these, the fine gain calculation unit 22 calculates the average gain.
A vibration composed of the vehicle body, wheels, and road surface around the rake pressure.
Exciting the brake pressure minutely with the resonance frequency of the dynamic system ω∞ (Equation (41))
Wheel speed signal ω when shaken_iAt the resonance frequency of ω ∞
Small amplitude (Wheel speed Small amplitude ω _wv) Wheel speed is small
The amplitude detection unit 40 and the brake pressure of the resonance frequency ω∞ are very small.
Amplitude P_vA brake pressure minute amplitude detection unit 42 for detecting
Detected wheel speed minute amplitude ω_wvBrake pressure small amplitude P
_vBy dividing by_dDivision that outputs
And a container 44. In addition, a small amount of brake pressure
The shaking means will be described later.

Here, the small wheel speed amplitude detecting section 40 can be realized as a calculating section as shown in FIG. 14 which performs a filtering process for extracting the vibration component of the resonance frequency ω∞. For example, since the resonance frequency ω ∞ of this vibration system is approximately 40 [Hz], one cycle is 24 [ms], approximately 41.
The frequency is set to 7 [Hz], and the band pass filter 75 having this frequency as the center frequency is provided. By this filter, only the frequency component in the vicinity of about 41.7 [Hz] is extracted from the wheel speed signal ω _i . Further, this filter output is subjected to full-wave rectification and DC smoothing by the full-wave rectifier 76, and only the low-pass vibration component is passed from this DC smoothed signal by the low-pass filter 77 to output a small wheel speed amplitude ω _wv . To do.

An integer multiple of the cycle, for example, 24 of one cycle
The wheel speed minute amplitude detection unit 40 can also be obtained by continuously capturing time-series data of 48 [ms] for 2 [ms] and 2 cycles and obtaining a correlation with a unit sine wave and a cosine wave of 41.7 [Hz]. Can be realized.

By the way, as described above, each control solenoid valve (valve SFR, valve SFL, valve S
A desired braking force can be realized by controlling the pressure increasing / decreasing time of the RR and the valve SRL according to the master cylinder pressure (booster pressure). Then, the minute excitation of the brake pressure can be performed by simultaneously performing the pressure increase / decrease control of the control solenoid valve that realizes the average brake force and the pressure increase / decrease control in a cycle corresponding to the resonance frequency.

As a concrete content of the control, as shown in FIG. 15, the pressure increasing mode and the pressure decreasing mode are switched at every half cycle T / 2 of the minute excitation cycle (for example, 24 [ms]), and the valve is turned on. decrease pressure command, the moment between the pressure boosting t _i of mode switching, and outputs the respective just time period the pressure increase, pressure reduction command decompression time t _r, the rest of the time, and outputs the held command. The average braking force, as well as determined by the ratio of the pressure increasing time t _i corresponding to the master cylinder pressure (booster pressure) and decompression time t _r, a half cycle corresponding to the resonance frequency T /
By switching the pressure increasing / depressurizing mode for every two, a minute vibration is applied around the average braking force.

The brake pressure minute amplitude P_vIs the master
Cylinder pressure (booster pressure) of the valve shown in FIG.
Boosting time t_iLength and decompression time t_rDepending on the length of
Since it is determined by a predetermined relationship, the brake pressure micro amplitude detection in Fig. 4 is performed.
Outlet 42 is for master cylinder pressure (booster pressure), boosting pressure
Time t_iAnd decompression time t_rTo a small brake pressure amplitude P _v
Can be configured as a table for converting and outputting.

As already proved, the small gain G
_{Since d} is substantially proportional to the braking torque gradient,
The braking torque gradient calculator 56 can be configured as a multiplier that multiplies the calculated small gain G _{d by} an appropriate proportional coefficient. Since the minute gain G _d tends to increase as the vehicle speed decreases, by changing the proportional coefficient according to the vehicle speed, an accurate braking torque gradient can be calculated regardless of the vehicle speed.

Next, the operation of the third embodiment will be described. When the brake pressure is slightly excited at the resonance frequency ω∞,
The minute gain calculator 22 calculates the minute gain G _d , and the braking torque gradient calculator 56 converts and outputs the minute gain G _d to the braking torque gradient.

As described above, in the third embodiment, the braking torque gradient that accurately represents the dynamic characteristics of the wheel motion can be easily calculated, and therefore, the third embodiment can be applied to the technique of performing various controls according to the friction state.

For example, if the ABS control means 14 of FIG. 1 or the ABS control means 15 of FIG. 2 is configured to use the calculation result of the braking torque gradient estimating device 56, the first and second embodiments will be described. It is possible to achieve the same effect as the ABS control device.

Further, the braking torque gradient estimating device 57 is
Besides applying to the ABS device as in the above example, for example,
The present invention can also be applied to a warning device or the like that issues a warning regarding braking to a driver according to the value of the estimated braking torque gradient.

The embodiments of the present invention have been described above, but the present invention is not limited to the above examples, and various modifications can be made without departing from the gist of the present invention.

For example, the ABS control device of the above embodiment is designed to follow the peak μ so that the braking torque gradient becomes 0 or a value close to 0, but the braking torque gradient is controlled to a reference value other than 0. It can also be designed to

Further, the limit judgment determining device 55 of the second embodiment is the braking torque gradient estimating device 8 of the first embodiment and the braking torque gradient estimating device 57 of the third embodiment.
It is also possible to apply to the limit determination based on the braking torque gradient estimated by.

Further, in the third embodiment, the minute excitation means of the brake pressure is realized by the mode switching of the pressure increase / decrease of the control solenoid valve. However, the brake disc is expanded and contracted according to the excitation signal. The brake pressure micro-excitation may be directly applied to.

[0204]

As described above, the first to fourth aspects of the invention are described.
According to the invention, instead of detecting the locked state of the wheel from the comparison of the wheel speed or the vehicle speed or the comparison of the slip ratio, the braking torque gradient is estimated from the time-series data of the wheel speed, and based on this control torque gradient, Since the braking force is controlled, there is an excellent effect that stable and comfortable antilock brake control can be performed with high accuracy regardless of the state of the road surface on which the vehicle travels.

Further, according to the inventions of claims 2 to 4, a parameter for identifying a small number of physical quantity matrix elements representing the history of physical quantities relating to changes in wheel speed and the history of physical quantities relating to changes in change in wheel speed. Therefore, the calculation accuracy can be improved and the calculation time can be shortened, and the device can be simplified because only the wheel speed needs to be detected.

According to the fifth and sixth aspects of the present invention, instead of detecting the locked state of the wheel from the comparison of the wheel speed or the vehicle speed or the slip ratio, the wheel deceleration and the brake torque are detected. Since the braking torque gradient is estimated from the time-series data and the braking force is controlled based on this control torque gradient, stable and comfortable antilock brake control can be performed with high accuracy regardless of the state of the road surface on which the vehicle is traveling. , An excellent effect is obtained.

Further, according to the fifth and sixth aspects of the invention, the braking torque gradient, which is a parameter to identify the motion state approximated by the gradient model, the physical quantity relating to the variation of the braking torque, and the variation of the slip speed. Since the braking torque gradient is estimated by reducing the relationship between physical quantities and applying an online system identification method to this relationship, the number of identification parameters can be set to one and the calculation accuracy can be greatly improved. Further, there is a further effect that the calculation time can be shortened.

According to the seventh aspect of the present invention, the non-linear variation of the braking torque and the non-linear variation of the braking torque gradient with respect to the disturbance of the slip velocity around the equilibrium point, the non-linear variation of the braking torque gradient, and the motion state of each wheel are described. Based on the first and second models represented by linear fluctuations that fluctuate within the first and second ranges, and the first and second ranges fall within a predetermined allowable range,
In addition, the operation amount of the braking force is calculated so that the gradient of the braking torque of the second model matches the braking torque gradient estimated by the torque gradient estimating means, so that the interference between the four wheels is taken into consideration. The further effect that anti-lock brake control can be performed is obtained.

Further, according to the invention of claim 8, since the braking torque gradient is estimated from only the time-series data of the wheel speed with a small number of parameters, the calculation accuracy can be improved and the calculation time can be shortened. Since it suffices to detect only the wheel speed, the device can be simplified.

Further, according to the invention of claim 9, the braking torque gradient is used as a parameter to identify a small number of matrix elements of the physical quantity representing the history of the physical quantity related to the change of the wheel speed and the history of the physical quantity related to the change of the change of the wheel speed. Since it is possible to estimate, the calculation accuracy can be improved and the calculation time can be shortened, and further, the device can be simplified because only the wheel speed needs to be detected.

According to the tenth and eleventh aspects of the present invention, the braking torque gradient which is a parameter to be identified for the motion state approximated by the gradient model, the physical quantity relating to the variation of the braking torque, and the variation of the slip speed are related. Since the braking torque gradient is estimated by reducing the relationship between physical quantities and applying an online system identification method to this relationship, the number of identification parameters can be set to one and the calculation accuracy can be greatly improved. Further, the further effect that the calculation time can be shortened is obtained.

[0212]

[Brief description of drawings]

FIG. 1 is a block diagram showing configurations of an antilock brake control device and a braking torque gradient estimating device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an antilock brake control device and a braking torque gradient estimation device according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a limit determination device according to a second embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a braking torque gradient estimating device according to a third embodiment of the present invention.

FIG. 5 is a diagram showing a relationship between a slip speed, a braking torque and a braking torque gradient.

FIG. 6 shows changes in braking torque F _i and braking torque gradient G _i as a function of slip speed, FIG.
The upper and lower limits of the variation of the braking torque F _i, is a diagram showing the upper and lower limits of variations in (b) is a braking torque gradient G _i.

7A and 7B are views for explaining the configuration of the wheel speed detecting means according to the embodiment of the present invention, FIG. 7A is a configuration diagram of the wheel speed detecting means, and FIG. 7B is an alternating current generated in the pickup coil. It is a figure which shows the time change of a voltage.

FIG. 8 is a diagram showing a configuration of an ABS control valve according to an embodiment of the present invention.

FIG. 9 is a diagram showing a configuration of a system hydraulic circuit including an ABS control valve according to an embodiment of the present invention.

FIG. 10 is a flowchart showing a flow of ABS control according to the first embodiment of the present invention.

FIG. 11 is a diagram showing a dynamic model of a vehicle to which the ABS control according to the embodiment of the present invention is applied.

FIG. 12 is a diagram showing an equivalent model of a vibration system including wheels, a vehicle body, and a road surface.

FIG. 13 shows a change characteristic of a friction coefficient μ with respect to a slip speed and shows that a change in μ around a center of a minute vibration can be approximated by a straight line in order to explain that a minute gain is equivalent to a braking torque gradient. It is a figure.

FIG. 14 is a block diagram showing a configuration of a wheel speed minute amplitude detection section of a minute gain calculation section according to a third embodiment of the present invention.

FIG. 15 is a diagram showing a command to a control solenoid valve when minute excitation of brake pressure and control of average braking force are simultaneously performed.

FIG. 16 is a diagram showing an outline of a vehicle body speed estimation method used in a conventional antilock brake control device.

FIG. 17 is a diagram showing a characteristic of a friction coefficient μ between a tire and a road surface with respect to a slip ratio.

FIG. 18 is a block diagram of a conventional ABS control device using a vehicle body speed estimation unit.

[Explanation of symbols]

10 Wheel speed detection means 12 Torque gradient estimating means 14 ABS control means 15 ABS control means 16 ABS control valve 22 Micro gain calculator 40 Wheel speed minute amplitude detector 42 Brake pressure micro amplitude detector 50 Braking torque gradient estimating device 51 Brake torque detecting means 52 Wheel deceleration detection means 53 Torque gradient estimating means 54 Limit determination means 55 Limit determination device 56 Braking torque gradient calculator 57 Braking torque gradient estimating device

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Hiroyuki Yamaguchi Inventor Hiroyuki Nagakute, Aichi-gun, Aichi Nagalete 1 1 at 41 Yokomichi Toyota Central Research Institute Co., Ltd. (72) Ken Sugai, Aichi-gun Nagakute-cho, Aichi 1 Toyota Yokosuka Laboratories Ltd., Yokochi 41 (56) Reference JP-A-8-201235 (JP, A) JP-A-62-85751 (JP, A) JP-A-62-166152 (JP, A) Kaihei 5-39018 (JP, A) JP 4-224447 (JP, A) JP 8-324414 (JP, A) JP 59-224913 (JP, A) JP 58-201103 ( JP, A) JP-A-8-318842 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) B60T 8 / 00,8 / 32-8/96

Claims (11)

(57) [Claims] 1. A braking torque with respect to a slip speed using only wheel speed detection means for detecting a wheel speed at a predetermined sampling time and time series data of the wheel speed detected by the wheel speed detection means. a torque gradient estimating means to estimate the gradient of the control means the slope of the brake torque estimated by the torque-gradient estimating means controls the braking force acting on the wheel to a value in a predetermined range including a reference value , Including anti-lock brake control device. 2. The torque gradient estimating means calculates a physical quantity relating to a change in the wheel speed and a physical quantity relating to a change in the wheel speed, based on the detected time-series data of the wheel speed. A history of physical quantities related to changes in wheel speed and a history of physical quantities related to changes in wheel speed are represented based on physical quantities related to changes in wheel speeds and physical quantities related to changes in wheel speeds calculated by the first calculation means. The anti-lock brake control device according to claim 1, further comprising: a second calculation unit that calculates a physical quantity and estimates a braking torque gradient from the physical quantity. 3. The second computing means linearly changes the braking torque and the motion state of the wheel when the braking torque acts on the wheel according to the gradient of the braking torque with respect to the slip speed. While approximating with a gradient model, the approximated motion state is converted in advance into a relationship of a gradient of braking torque with respect to a parameter to be identified, that is, a gradient of braking torque, a physical quantity related to a change in wheel speed, and a physical quantity related to a change in change in wheel speed. By applying the online system identification method to each data in which the physical quantity relating to the change in the wheel speed and the physical quantity relating to the change in the wheel speed calculated by the first computing means are sequentially applied to the above relationship. 3. The antilock according to claim 2, wherein the gradient of the braking torque with respect to the slip speed is estimated. Brake control device. 4. The first calculating means calculates time series data of wheel speeds detected at a sample time k (k = 1, 2, ...) At a wheel with a wheel number i by ω i [k ], Where τ is the sample time and J is the wheel inertia, Is calculated, and yi [k] = − ωi [k] + 2ωi [k−1] −ωi [k−2] is calculated as a physical quantity related to the change of the wheel speed, and the second calculation means is , A physical quantity θi representing the history of physical quantities related to changes in wheel speed and a history of physical quantities related to changes in wheel speed, the forgetting factor λ, and the transpose of the matrix "T"
As shown below, The anti-lock brake control device according to claim 2, wherein the anti-lock brake control device is characterized in that the first element of the matrix of the estimated value θ _i is obtained as a gradient of the braking torque with respect to the slip speed, by estimating from a recurrence formula of ^. 5. Based on time series data of wheel deceleration detected for each predetermined sample time, and brake torque detected for each predetermined sample time or time series data of a physical quantity related to the brake torque. A torque gradient estimating means for estimating a gradient of the braking torque with respect to the slip speed, and a braking force acting on the wheels so that the gradient of the braking torque estimated by the torque gradient estimating means falls within a predetermined range including a reference value. An antilock brake control device having: a torque control means for controlling the braking torque and the wheel motion state when the braking torque is applied, wherein the braking torque is the braking torque with respect to the slip speed. Approximate by a gradient model that changes linearly according to the gradient, and identify the approximated motion state The slope of the braking torque with respect to the slip speed, which is a power parameter, is converted in advance into the relationship between the physical quantity related to the change of the braking torque and the physical quantity related to the change of the slip speed, which are represented by the braking torque and the wheel deceleration, respectively, and detected. By applying the online system identification method, to the respective data in which the time series data of the wheel deceleration and the detected brake torque or the time series data of the physical quantity related to the brake torque are sequentially applied to the relationship, An antilock brake control device characterized by estimating a gradient of a braking torque with respect to a slip speed. 6. The torque gradient estimating means sets yi [j] as time series data of wheel deceleration at a wheel of wheel number i at sample time j, and Tbi [j] as time series data of brake torque. The sample time of τ,
A wheel inertia is J, a wheel radius is Rc, a vehicle mass is M, a vector having time series data of brake torque of each wheel in each component is Tb [j], and time series data of wheel deceleration of each wheel is each component. Where y [j] is the vector, the identity matrix is I, and the matrix whose diagonal component is {(J / MRc 2) +1} and whose off-diagonal component is J / MRc 2 is A, the change in braking torque The physical quantity f and the physical quantity φ related to the change of the slip speed are given by Let K be a matrix having the gradient of the braking torque for each wheel, which is the parameter to be identified, as the diagonal component and the off-diagonal component being 0, and K is represented by K · φ = f It is converted into a relational expression in advance, and the detected wheel deceleration time series data yi [j] (j =
1,2,3, .....) and time series data Tbi [j] (j = 1,2,3, .....) of the detected braking torque are sequentially applied to the above relational expressions. The anti-lock brake control device according to claim 5 , wherein the gradient of the braking torque for each wheel is estimated by applying an online system identification method to the data. 7. The control means sets the slip speed at which the braking torque generated in each wheel is maximum as an equilibrium point of each wheel, and controls the braking torque and the operation amount of the braking force acting around the equilibrium point. The motion state of each wheel when it acts on the wheel, the motion state of the vehicle body when the braking torque generated on each wheel acts on the entire vehicle body, and the motion state of each wheel against the disturbance of the slip speed of each wheel around the equilibrium point. A first model in which the nonlinear fluctuation of the braking torque is represented by a linear fluctuation that fluctuates within a first range with respect to the disturbance of the slip speed of each wheel, and a model for each of the disturbance of the slip speed of each wheel around the equilibrium point A second model in which the nonlinear variation of the braking torque gradient of the wheels is represented by a linear variation that fluctuates within a second range with respect to the disturbance of the slip speed of each wheel, and The torque gradient estimating means estimates the braking torque gradient of the second model designed such that the second range is within the predetermined allowable range and the second range is within the predetermined allowable range. The operation amount of the braking force of each wheel that matches the gradient of the braking torque is calculated, and the braking force acting on each wheel is controlled based on the calculated operation amount of the braking force of each wheel. The antilock brake control device according to any one of claims 1 to 6. 8. A braking torque with respect to a slip speed, using only wheel speed detection means for detecting a wheel speed for each predetermined sample time and time series data of the wheel speed detected by the wheel speed detection means as a detection target. or a torque gradient estimating means to estimate the slope of the driving torque, the torque gradient estimating device comprising an output means for outputting the estimated value of the slope of the estimated braking torque or the driving torque by the torque gradient estimating means. 9. The torque gradient estimating means calculates a physical quantity related to a change in wheel speed and a physical quantity related to a change in wheel speed based on time series data of the detected wheel speed, and first calculating means. A history of physical quantities related to changes in wheel speed and a history of physical quantities related to changes in wheel speed are represented based on physical quantities related to changes in wheel speeds and physical quantities related to changes in wheel speeds calculated by the first calculation means. 9. A torque gradient estimating device according to claim 8, further comprising: a second computing means for computing a physical quantity and estimating a gradient of the braking torque or the driving torque from the physical quantity. 10. Based on time series data of wheel deceleration detected for each predetermined sample time, and brake torque detected for each predetermined sample time or time series data of a physical quantity related to the brake torque. A braking torque gradient estimating device comprising: a torque gradient estimating means for estimating a gradient of the braking torque with respect to a slip speed; and an output means for outputting an estimated value of the gradient of the braking torque estimated by the torque gradient estimating means, The torque gradient estimating means approximates the braking torque and the motion state of the wheel when the braking torque is applied by a gradient model in which the braking torque changes linearly according to the gradient of the braking torque with respect to the slip speed. , The gradient of the braking torque with respect to the slip speed, which is a parameter to identify the approximated motion state, The time series data of the detected wheel deceleration and the detected wheel deceleration are detected in advance and converted into the relationship between the physical quantity related to the change of the braking torque and the physical quantity related to the change of the slip speed, which are represented by the brake torque and the wheel deceleration, respectively. It is possible to estimate the gradient of the braking torque with respect to the slip speed by applying an online system identification method to each data obtained by sequentially applying time series data of the braking torque or the physical quantity related to the braking torque to the above relationship. A characteristic braking torque gradient estimating device. 11. The torque gradient estimating means sets yi [j] as time series data of wheel deceleration and Tbi [j] as time series data of brake torque at a wheel of wheel number i at sample time j, and the predetermined value. The sample time of τ,
A wheel inertia is J, a wheel radius is Rc, a vehicle mass is M, a vector having time series data of brake torque of each wheel in each component is Tb [j], and time series data of wheel deceleration of each wheel is each component. Where y [j] is the vector, the identity matrix is I, and the matrix whose diagonal component is {(J / MRc 2) +1} and whose off-diagonal component is J / MRc 2 is A, the change in braking torque The physical quantity f and the physical quantity φ related to the change of the slip speed are given by Let K be a matrix having the gradient of the braking torque for each wheel, which is the parameter to be identified, as the diagonal component and the non-diagonal component being 0, and K is represented by It is converted into a relational expression in advance, and the detected wheel deceleration time series data yi [j] (j =
1,2,3, .....) and time series data Tbi [j] (j = 1,2,3, .....) of the detected braking torque are sequentially applied to the above relational expressions. 11. The braking torque gradient estimating device according to claim 10, wherein the braking torque gradient for each wheel is estimated by applying an online system identification method to the data.