JP2002321605A - Torque grade estimating device and anti-lock brake control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a torque grade estimating device capable of detecting the tendency of the locking and slipping of wheels at a high speed and stabilizing the estimation of the value of a torque grade. SOLUTION: A torque grade is estimated by means of a torque grade estimating means 13 by detecting the wheel speed at every specified sample time by means of a wheel speed detecting means 10, applying by-pass filter treatment to the time series data of the detected wheel speed by means of a treatment means 11, setting the value of a forgetting coefficient, which is a parameter showing a degree of eliminating past data by means of a torque grade increase and decrease determination means 12A, according to the increasing and decreasing state of the torque grade estimated by the torque grade estimating means 13, and applying an on-line system identification method, for which the forgetting coefficient is used, using a grade model, by which the time series data of the wheel speed, to which by-pass filter treatment is applied, a wheel driving model, and a brake torque or drive torque change in linear function relation as a peak of torque to a slipping speed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a torque gradient estimating device and an antilock brake control device, and more particularly, to a torque gradient estimating device and a torque gradient estimating device for estimating a torque gradient with respect to a slip speed from time-series data of wheel speeds. The present invention relates to an anti-lock brake control device that controls a braking force acting on a wheel based on a gradient of the vehicle.

[0002]

2. Description of the Related Art Conventionally, as a technique for estimating a braking torque gradient (gradient of a braking torque with respect to a slip speed), JP-A-10-114263 and JP-A-200200 have been proposed.
There is a technique described in Japanese Patent Application No. 0-118375.

[0003] Among them, in the technique described in Japanese Patent Application Laid-Open No. 2000-118375, data obtained by performing high-pass filtering on time-series data of wheel speed, a wheel motion model,
A gradient model in which the braking torque changes linearly according to the braking torque gradient with respect to the slip speed, and braking by applying an online system identification method using a forgetting factor that indicates the degree to which past data is removed. The torque gradient or the driving torque gradient was estimated. The torque gradient estimated in this way is used to detect the tendency of the wheels to lock or slip.

Here, since the forgetting coefficient is a parameter indicating the degree of removing past data, the responsiveness of the torque gradient estimation increases as this value is reduced, but the robustness against disturbance decreases. For this reason, the forgetting factor cannot be reduced unnecessarily, and the forgetting factor has conventionally been set to a constant value.

[0005]

However, in the technique described in Japanese Patent Application Laid-Open No. 2000-118375, the forgetting factor used in the online system identification method is set to a constant value as described above. It is difficult to increase the speed of detecting the tendency of locking and slipping of the wheels that require detection, and also stabilizes the estimated value of the torque gradient when driving on rough roads or running with a chain attached. There was a problem that it was difficult to make it.

[0006] In particular, when the performance of the antilock brake control device is to be improved by using the estimated value of the braking torque gradient in the antilock brake control device, the locking tendency of the wheels is required to secure the braking force or the lateral force. The detection is required to be faster, and the same is true at the time of driving. In a case where the traction control is used, the slip tendency is required to be detected faster, whereas the slip tendency is required. In a running scene in which a large vibration component is superimposed on a wheel speed signal such as when traveling on a road, stable estimation of the torque gradient is required, and thus the above problem is a serious problem.

By the way, as shown in FIG. 22, in an anti-lock brake control device which detects a tendency of the wheels to lock and reduces the hydraulic pressure of the wheel cylinder, and thereafter repeats a pattern of gradually increasing the pressure, the wheels are locked. When there is a tendency, the vibration of the wheel speed increases.

In the case where a technology for estimating a torque gradient by applying an on-line system identification method, such as the technology described in Japanese Patent Application Laid-Open No. 2000-118375, is used for such an anti-lock brake control device, The greater the vibration amplitude of the speed, the greater the weight of the wheel speed, which has a greater influence on the estimated value of the torque gradient later.

Therefore, even in the grip state where the wheel speed immediately after the pressure reduction of the wheel cylinder oil pressure is restored, the wheel speed in the past state of the slip state is reflected, and the estimated value of the torque gradient may be smaller than the actual value. There was a problem.

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above-mentioned problems, and it is an object of the present invention to realize both high speed detection of a tendency to lock and slip a wheel and stabilization of an estimated value of a torque gradient. It is a first object to provide a torque gradient estimating device capable of performing a torque gradient estimation, and a second object is to provide a torque gradient estimating device capable of improving the estimation accuracy of a torque gradient. A third object is to provide an anti-lock brake control device that can perform the control.

[0011]

In order to achieve the first object, a torque gradient estimating apparatus according to a first aspect of the present invention comprises a detecting means for detecting a wheel speed at a predetermined sampling time, and a detecting means for detecting a wheel speed. Processing means for performing high-pass filter processing on the time-series data of the detected wheel speeds; setting means for setting a value of a parameter indicating a degree of removing past data in accordance with the traveling state of the vehicle; Time-series data of the filtered wheel speed, a wheel motion model, and a gradient model in which the braking torque or the driving torque changes linearly as a gradient of the torque with respect to the slip speed as a gradient, using the above parameters online And a torque gradient estimating means for estimating a torque gradient by applying the system identification method.

[0012] According to the torque gradient estimating device of the first aspect, the time series data of the wheel speed detected by the detecting means for detecting the wheel speed at every predetermined sample time is used.
High-pass filtering is performed by the processing means.

According to the first aspect of the present invention, the value of a parameter (corresponding to the aforementioned forgetting factor) indicating the degree of removing past data is set by the setting means in accordance with the running state of the vehicle, and the torque gradient is set. The time series data of the wheel speed, the motion model of the wheel, and the gradient model in which the braking torque or the driving torque changes linearly with the gradient of the torque with respect to the slip speed as a gradient with respect to the slip speed. Is used, and the gradient of torque is estimated by applying an online system identification method using the above parameters.

That is, in the present invention, the value of the parameter indicating the degree of removing the past data is set by the setting means in accordance with the running state of the vehicle. The values of the above parameters can be appropriately switched between a driving scene in which a large vibration component is superimposed on a wheel speed signal when the vehicle is traveling, and a driving scene in which it is desired to speed up the detection of the locking tendency and the slipping tendency of the wheels. As a result, both the speed of detecting the tendency of the wheels to lock and the tendency of slipping can be increased and the estimated value of the torque gradient can be stabilized.

As described above, according to the torque gradient estimating apparatus of the first aspect, the high-pass filter processing is performed on the time series data of the wheel speed detected by the detecting means for detecting the wheel speed at every predetermined sampling time. At the same time, the value of a parameter indicating the degree of removing past data is set according to the traveling state of the vehicle, and the time series data of the wheel speed subjected to the high-pass filter, the wheel motion model, and the braking torque or the driving torque slip. Using a gradient model that changes linearly with the gradient of the torque with respect to the speed as a gradient, and applying the on-line system identification method using the above parameters, the gradient of the torque is estimated. A running scene in which a large vibration component is superimposed on the wheel speed signal when running with the chain attached, etc. It is possible to appropriately switch the values of the above parameters depending on the driving scene in which the detection of the tendency to slip and the tendency to slip is desired to be speeded up. As a result, the detection speed of the tendency to lock and slip the wheel and the torque can be increased. It is possible to realize both stabilization of the estimated value of the gradient.

Incidentally, the torque gradient has a characteristic that when the wheel is in a locking tendency or a slipping tendency, the torque gradient changes to a decreasing side, and when the wheel is in a gripping tendency, the torque gradient changes to an increasing side.

In addition, the quick detection of the tendency of the wheels to be locked or slipped is related to the improvement of vehicle stability and the reduction of oil consumption by optimizing the ABS pressure reduction control. Fast detection is desired.

Therefore, a torque gradient estimating device according to a second aspect of the present invention provides the torque gradient estimating device according to the first aspect,
When the torque gradient estimated by the torque gradient estimating means changes to an increasing side, the value of the parameter is set to a maximum value or a value near the maximum value allowed as the value of the parameter, When the torque gradient estimated by the above changes to the side where the torque gradient decreases, the value of the parameter is set to a minimum value or a value in the vicinity of the minimum value allowed as the value of the parameter.

According to the torque gradient estimating device of the second aspect, when the setting means changes the torque gradient estimated by the torque gradient estimating means to the increasing side, the value of the parameter is set as the value of the parameter. When the torque gradient estimated by the torque gradient estimating means changes to the maximum value or a value near the maximum value of the allowable value on the side where the torque gradient decreases, the value of the parameter is set to the minimum value of the allowable value of the parameter. The value is set to a value or a value near the minimum value.

When the value of the parameter is reduced, the degree of removal of the past data increases, so that faster estimation becomes possible. When the value of the parameter is increased, the degree of removal of the past data decreases,
The estimated value is stabilized, and more accurate estimation is possible.

As described above, according to the torque gradient estimating apparatus according to the second aspect, when the estimated torque gradient changes to an increasing side, that is, when the tire has a tendency to grip, the past data is used. If the value of the parameter indicating the degree of removal of the parameter changes to the maximum value of the value allowed as the value of the parameter or a value near the maximum value, the estimated torque gradient decreases, that is, the wheel tends to lock or slip. If there is a tendency, the value of the parameter is set to a minimum value or a value in the vicinity of the minimum value allowed as the value of the parameter, whereby the parameter is appropriately switched according to the driving scene. Therefore, both the speed of detecting the tendency of wheel lock and the tendency of slip can be increased and the estimated value of the torque gradient can be stabilized. Faster determination of a locking tendency or slip tendency of an effect wheel reduction of stability and ABS oil consumption can be achieved.

As described above, a large vibration component is superimposed on the wheel speed signal when traveling on a rough road or traveling with a chain attached.

According to a third aspect of the present invention, in the torque gradient estimating apparatus according to the first aspect, the setting means includes:
When the vibration level of the wheel speed detected by the detection means is equal to or higher than a predetermined level, the value of the parameter is set to a maximum value or a value near the maximum value allowed as the value of the parameter.

According to the third aspect of the present invention, when the vibration level of the wheel speed detected by the detecting means is equal to or higher than a predetermined level, the value of the parameter is set as the value of the parameter. It is set to the maximum value of the allowable value or a value near the maximum value. Note that, when the vibration level is equal to or higher than the value indicated by the level, the predetermined level is a value that can be regarded as running on a rough road or running with a chain attached. Alternatively, values obtained in advance by computer simulation or the like can be applied.

As described above, according to the torque gradient estimating device of the third aspect, when the vibration level of the wheel speed is equal to or higher than the predetermined level, the value of the parameter indicating the degree of removing the past data is changed to the value of the parameter. Because the value is set to the maximum value of the value allowed as a value or a value near the maximum value,
It is possible to stabilize the estimated value of the torque gradient at the time of traveling on a rough road or traveling with a chain attached.

Further, in the torque gradient estimating apparatus according to a fourth aspect of the present invention, in the first aspect of the present invention, when the vibration level of the wheel speed detected by the detecting means is equal to or higher than a predetermined level, The parameter value is set to the maximum value or a value near the maximum value of the value allowed as the parameter value, and when the vibration level is less than the predetermined level, the torque gradient estimated by the torque gradient estimating means is set. When the value changes to the increasing side, the value of the parameter is set to the maximum value or a value near the maximum value of the value allowed as the value of the parameter, and the value of the parameter is reduced to the value near the maximum value. When it changes, the value of the parameter is set to the minimum value or a value near the minimum value of the value permitted as the value of the parameter.

According to a fourth aspect of the present invention, when the vibration level of the wheel speed detected by the detecting means is equal to or higher than the predetermined level, the value of the parameter is set as the value of the parameter. When the vibration level is set to the maximum value or a value near the maximum value of the allowable value and the vibration level is less than the predetermined level, the torque gradient estimated by the torque gradient estimating means changes to the increasing side. When the value of the parameter changes to the maximum value or a value near the maximum value of the value allowed as the value of the parameter to the side where the torque gradient estimated by the torque gradient estimating means decreases, the value of the parameter is changed to the value of the parameter. Are set to the minimum value or a value in the vicinity of the minimum value that is allowed as the value of.

As described above, according to the torque gradient estimating device of the fourth aspect, when the vibration level of the wheel speed is equal to or higher than the predetermined level, the value of the parameter indicating the degree of removing the past data is changed to the value of the parameter. When the vibration level is less than the predetermined level, the value of the parameter is set when the estimated torque gradient changes to an increasing side. When the estimated torque gradient changes to the maximum value or a value near the maximum value of the value allowed as the value of the parameter and the estimated torque gradient decreases, the value of the parameter is set to the minimum value of the value allowed as the value of the parameter. By setting the values to values or values near the minimum value, the above parameters are appropriately switched according to the driving scene. Both can be achieved in the stabilization of the estimates of the speed and torque gradient of the detected speed.

On the other hand, in order to achieve the second object,
The torque gradient estimating device according to claim 5, wherein a detecting means for detecting a wheel speed at every predetermined sample time; a processing means for performing a high-pass filter process on the time-series data of the wheel speed detected by the detecting means; The time series data of the wheel speed subjected to the high-pass filter processing by the processing means, the wheel motion model, and a gradient model in which the braking torque or the driving torque changes linearly with the gradient of the torque with respect to the slip speed as a gradient, A torque gradient estimating means for estimating a torque gradient by applying an online system identification method using a parameter indicating a degree of removing data, a state determining means for determining a locked state of a wheel, and the state determining means. The torque gradient estimating means is controlled by the torque gradient estimating means based on the determination result. It comprises a constant control means.

According to the torque gradient estimating device of the fifth aspect, the time series data of the wheel speed detected by the detecting means for detecting the wheel speed at every predetermined sample time is obtained.
High-pass filtered by the processing means, the time-series data of the high-pass filtered wheel speed, the wheel motion model, and the gradient in which the braking torque or the driving torque changes linearly with the gradient of the torque with respect to the slip speed. The torque gradient is estimated by the torque gradient estimating means by applying an online system identification method using a model and a parameter indicating a degree of removing past data.

According to the fifth aspect of the present invention, the locked state of the wheels is determined by the state determining means, and the estimation operation of the torque gradient by the torque gradient estimating means is controlled by the estimation control means based on the determination result. You.

That is, according to the present invention, the torque gradient estimating operation is controlled by the torque gradient estimating means in accordance with the locked state of the wheel. By controlling the torque gradient estimating operation such that the influence of the wheel speed on the torque gradient estimated value in the locked state becomes small, the accuracy of the torque gradient estimation can be improved.

As described above, according to the torque gradient estimating device of the fifth aspect, the operation of estimating the torque gradient by the torque gradient estimating means is controlled based on the determination result of the locked state of the wheel. The torque gradient estimation operation is controlled by controlling the torque gradient estimation operation such that when the vehicle is released from the state in which the vehicle tends to be locked and the wheel speed in the past state of the lock state has less effect on the estimated value of the torque gradient. Accuracy can be improved.

According to a sixth aspect of the present invention, in the torque gradient estimating device according to the fifth aspect, the estimation control means determines that the torque gradient is high when the state determination means determines that the wheels are in a locking tendency. The estimation operation by the torque gradient estimating means is controlled so that the estimation of the torque gradient by the estimating means is interrupted, and the estimation of the torque gradient by the torque gradient estimating means is resumed when it is determined that the locking tendency has been released. Or if the state determination means determines that the wheel has a tendency to lock, the value of the parameter is smaller than a value when it is determined that the parameter has no tendency to lock for a predetermined period from the time when the wheel is released from the locked state. It is controlled so as to be a value.

According to the torque gradient estimating device of the sixth aspect, the estimation gradient means suspends the estimation of the torque gradient by the torque gradient estimating means when the state judging means judges that the wheels have a tendency to lock. The estimation operation by the torque gradient estimating means is controlled so that the estimation of the torque gradient by the torque gradient estimating means is resumed when it is determined that the wheel is released from the locking tendency, or the wheels are locked by the state determining means. When it is determined that there is a tendency, the value of the parameter is controlled to be smaller than the value when it is determined that there is no tendency to lock for a predetermined period from the time when the lock state is released.

As described above, according to the torque gradient estimating apparatus of the sixth aspect, when it is determined that the wheels have a tendency to lock, the estimation of the torque gradient is interrupted, and it is determined that the wheels are released from the locking tendency. The estimation operation is controlled so that the estimation of the torque gradient is resumed at the point in time, or the degree of removal of past data for a predetermined period from the point in time when the wheel is released from the locked state when it is determined that the wheel is in a locked state. Since the value of the parameter indicating is controlled to be smaller than the value when it is determined that there is no tendency to lock,
The influence of the wheel speed on the estimated value of the torque gradient in the locked state, which is the past state, can be reduced, and the estimation accuracy of the torque gradient can be improved.

On the other hand, in order to achieve the third object,
According to a seventh aspect of the present invention, there is provided an antilock brake control device, wherein the gradient of the braking torque estimated by the torque gradient estimating device according to any one of the first to sixth aspects includes a reference value. Control means for controlling the braking force acting on the wheels so as to be within a predetermined range.

According to a seventh aspect of the present invention, the torque gradient is estimated by the torque gradient estimating device according to any one of the first to sixth aspects. The braking force acting on the wheels is controlled by the control means so that the gradient has a value within a predetermined range including the reference value.

As described above, according to the antilock brake control device of the present invention, the torque gradient estimating device according to the present invention is provided, and the gradient of the braking torque estimated by the torque gradient estimating device is set to the reference value. Since the braking force acting on the wheels is controlled so as to be a value in a predetermined range including the above, the antilock brake control can be accurately performed.

(Principle of Estimating Torque Gradient of the Inventions of Claims 1 to 6) Here, the principle of estimating the torque gradient of the inventions of claims 1 to 6 will be described.

The wheel motion and the vehicle motion of each wheel are described by the following equations of motion.

[0042]

(Equation 1)

Where F _i ′ is the braking force generated on the i-th wheel, T _bi is the braking torque applied to the i-th wheel corresponding 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. 20). In addition,
Indicates the derivative with respect to time. In the equations (1) and (2), F _i ′ is shown as a function of the slip speed (v / R _c −ω _i ).

Here, the vehicle speed is equivalent to the angular speed of the vehicle.
ω_vAnd the braking torque R_cF _i’Is the slip speed
Linear function (slope k_i, Y section T_i). You
That is, as shown in FIG._cF_i’Pickpocket
Top speed (ω_v−ω_i) To the braking torque gradient k_iTilt
A gradient model that changes like a linear function is applied.

[0045] _{v = R c ω v (3} ) R c F i '(ω v -ω i) = k i × (ω v -ω i) + T i (4) In addition, (3), (4) Is substituted into the equations (1) and (2), and the wheel speed ω _i and the body speed ω _v are discretized for each sample time τ, and the time series data ω _i [k], ω _v [k] (k is a sample Sample time in units of time τ, k = 1,
When expressed as (2, ....), the following equation is obtained.

[0046]

(Equation 2)

Here, when equations (5) and (6) are simultaneously set and the equivalent angular velocity ω _{v of the} vehicle body is eliminated,

[0048]

(Equation 3)

Is obtained.

Assuming that a maximum braking torque of R _c Mg / 4 (g is gravitational acceleration) is generated under the condition of a slip speed of 3 rad / s,

[0051]

(Equation 4)

Is obtained. Here, as a specific constant, τ
= 0.005 (sec), R _c = 0.3 (m), M = 1
Considering 000 (kg), max (k _i ) = 245
Becomes Therefore,

[0053]

(Equation 5)

Equation (7) can be approximated as equation (8).

[0055]

(Equation 6)

By rearranging in this way, the equation (8) can be described in a linear form with respect to the unknown coefficients k _i and f _i . That is, equation (8) is a wheel and vehicle body motion equation in which the braking torque is approximated by a linear function of the slip speed. Then, by applying the online parameter identification method to the equation (8), the braking torque gradient k _i with respect to the slip speed can also be estimated.

By the way, when the above equation (8) is arranged by focusing on the acceleration of the wheel,

[0058]

(Equation 7)

Is as follows. From equation (9), the identification of k _i and f _i is as follows.
It can also be interpreted that the characteristic root -k _i / j and offset -f _i / τ of the wheel acceleration are estimated.

By the way, the wheel speed signal, an acceleration stationary component of the wheel is processed to be removed, i.e., put the wheel speed signal to the high-pass filter (the high-pass filtering), the offset term (-f _i / tau ) Can be zero. For example, when braking is performed at a constant deceleration, the offset term can be omitted by inserting a first-order or higher-order high-pass filter (a high-pass filter that allows a steady value of the wheel acceleration to be regarded as an offset value). Therefore, it is possible to assume that _fi = 0 by putting the wheel speed signal into a high-pass filter. Therefore, when the wheel gradient signal is input to the high-pass filter to estimate the torque gradient, the equation (8) can be modified as follows. That is, it is possible to transform into a relational expression including, as an unknown, a torque gradient obtained from the equations of motion of the wheels and the vehicle body and obtained by omitting the stationary term of the wheel acceleration from the equations of motion using the acceleration of the wheels.

[0061]

(Equation 8)

Where ω _hi [k] is the wheel speed after high-pass filtering.

Then, the following steps 1 and 2
Is repeated, the time series data of the braking torque gradient can be estimated from the time series data ω _i [k] of the detected wheel speed.

[0064]

(Equation 9)

Here, Note that φ _i [k] in equation (10)
Is a physical quantity related to a change in wheel speed during one sample time, and equation (12) is a physical quantity related to a change in wheel speed during one sample time.

[0066]

(Equation 10)

From the recurrence equation

[0068]

[Equation 11]

That is, the gradient of the braking torque is estimated. Here, λ is a forgetting coefficient indicating a degree of removing past data, and “ ^T ” indicates transposition of a matrix.

The left side of the equation (13) is a physical quantity representing the history of the physical quantity relating to the change of the wheel speed and the history of the physical quantity relating to the change of the wheel speed.

Here, in the torque gradient estimating apparatus according to the first to fourth aspects, the forgetting coefficient λ which is a parameter indicating the degree of removing the past data according to the running state of the vehicle.
Thus, both the speed of detecting the locking tendency and the tendency of slipping of the wheels are increased, and the estimated value of the torque gradient is stabilized.

In the torque gradient estimating device according to the fifth and sixth aspects, the torque gradient estimating operation is controlled based on the determination result by the state determining means for determining the locked state of the wheel. The accuracy of torque gradient estimation is improved.

As described above, the braking torque gradient is estimated.
Coefficient f which was conventionally required _iMay be omitted
The amount of calculation for estimating the braking torque gradient is small.
Can be eliminated. Therefore, the braking torque gradient is estimated.
The accuracy can be expected to improve. Note that the braking torque
Alternatively, a driving torque can be applied. Shown here
The estimating method uses the least squares method,
It is also possible to use other online identification methods such as numerical methods.
You.

(Principle of the ABS control according to the seventh aspect of the invention) The braking force acts on the road surface via the surface of the tread of the tire in contact with the road surface. In practice, however, this braking force is applied between the road surface and the wheels. It acts on the vehicle body as a reaction force (braking torque) from the road surface through the frictional force between the two. When the vehicle is running at a certain speed, a slip occurs between the wheels and the road surface when the braking force is applied. At this time, the braking torque acting as a reaction force from the road surface is expressed by the following equation. It changes as shown in FIG. 2 with respect to the slip speed ω _s (converted to angular speed).

Ω _s = ω _v −ω _i where ω _v is the vehicle speed (equivalently expressed by angular velocity), ω _i is the i-th wheel (i is the wheel number, i = 1, 2, 3, 4)・ ・
A wheel speed converted into an angular speed of (for a four-wheeled vehicle, i = 1, 2, 3, 4 (hereinafter, a four-wheeled vehicle is taken as an example)).

As shown in FIG. 2, the braking torque initially increases as the slip speed increases, reaches a maximum value f _i0 at the slip speed ω ₀ , and decreases as the slip speed increases at a slip speed greater than ω _0. . Note that the slip speed ω ₀ corresponds to the slip speed when the friction coefficient between the wheel and the road surface is the maximum value.

Therefore, as is apparent from FIG.
The gradient of the braking torque with respect to the
Is called ω)_s<Ω₀Is positive (> 0), ω_s= Ω₀so
0, ω _s> Ω₀Becomes negative (<0). That is, the braking torque
When the slope is positive, the wheels are gripping the road,
When the braking torque gradient is 0, the peak μ state, braking torque
When the slope is negative, the wheels are locked.
Dynamic characteristic of wheel motion changes according to braking torque gradient
You.

According to the seventh aspect of the present invention, the current braking torque gradient is estimated from only the time-series data of the wheel speed without estimating the vehicle body speed, and the estimated braking torque gradient is set to the reference value. The braking force acting on the wheels is controlled so as to have a value in a predetermined range including the predetermined range. Thereby, the state of the wheel motion corresponding to the braking torque gradient in the predetermined range including the reference value can be maintained. In addition, if the reference value is set to 0 corresponding to the peak μ, the braking torque gradient becomes 0 at the peak μ even if the slip speed at which the peak μ changes due to the road surface condition on which the vehicle travels. ,
If the braking torque gradient is controlled to be zero, it is possible to completely follow the peak μ. In addition, since the vehicle speed estimating unit is not required, it is not necessary to repeatedly increase and decrease the braking force, and stable driving can be performed.

The control system for feedback-controlling the control torque gradient may be designed for each wheel by PID control or the like, but may be systematically designed as an integrated system for all wheels by applying modern control theory. . In this case, more detailed control can be realized because the interference of all the wheels is considered in the design.

By the way, the ABS control system is a system having a strong non-linear characteristic of tire characteristics, and cannot simply apply modern control theory. Focusing on the fact that this nonlinear characteristic can be regarded as an apparently equivalent plant variation, a control system design that allows this plant variation has been achieved by applying robust control theory, one of modern control theories. In addition, a detailed control system design that takes into consideration the interference of the four wheels in the design was performed. The details of the control system design are described below.

The wheel motion and the vehicle motion of each wheel are described by the following equations of motion.

[0082]

(Equation 12)

Here, F _i is the braking torque generated on the i-th wheel and is shown as a function of the slip speed (ω _v −ω _i ). T _bi ′ is a brake torque corresponding to the depression force when the brake pedal is depressed immediately before the wheel is locked, u
_bi is a brake torque (operation amount) applied to the wheel so that the wheel follows the peak μ without falling into the locked state in a state where the brake torque is applied. Further, M vehicle mass, Rc is the effective radius of the wheel, the omega _v is the vehicle speed (which was expressed by equivalently angular velocity). Equation (16) is an output equation indicating 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 shown in FIGS.
(B) shows a peak and a zero at ωο, respectively.
These are non-linear functions of the slip speed, which can be expressed in the form of straight lines 20, 23 indicated by solid lines and fluctuations within a predetermined range. Here, assuming that the disturbance of the slip speed from ωο is x _i , F _i = (f _i + W _{fi Δfi} ) x _i + f _i0 (17) G _i = (g _i + W _{gi Δgi} ) x _i (18) Can be represented.

Here, f _i indicates the inclination of the straight line 20 in FIG. 3A, and g _i indicates the inclination of the straight line 23 in FIG. 3B. Also,
W _fi and W _gi are weighting factors for normalizing the fluctuation,
The broken lines 21 and 22 in FIG. 3A and the broken lines 24 and 25 in FIG. 3B represent the upper and lower limits of the nonlinear fluctuation, respectively.
_This corresponds to ± 1 for _fi and △ _gi .

That is, equation (17) shows the non-linear fluctuation of the braking torque of each wheel with respect to the disturbance x _i around the equilibrium point ω ₀ .
3A 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 in FIG. Also, (1
Equation 8) is a linear model that represents the non-linear variation of the braking torque gradient of each wheel with respect to the disturbance x _i around the equilibrium point ω _{0 by} the variation within the range from the broken line 24 including the straight line 23 to the broken line 25 in FIG. It is.

Further, the expressions (17) and (18) are
4), (15) and (16), and the equilibrium point (ω ₀ )
If described as the surrounding equation of state, the following equation is obtained.

[0088]

(Equation 13)

However,

[0090]

[Equation 14]

Also,

[0092]

(Equation 15)

Is as follows. Here, x is omega ₀ each wheel slip speed disturbance around, y is omega ₀ braking torque gradient of each wheel around, u is the manipulated variable of each wheel around ω ₀ ((14) equation u
(equivalent to _bi ).

Here, an arbitrary △ having the structure of equation (21)
By performing a control system design that allows (−1 ≦ △ _fi , △ _gi ≦ 1), it is possible to design an ABS control system in consideration of interference of four wheels. This design can be easily performed by applying the μ design method, which is a method of robust control.

That is, the following controller is designed by designing a control system having a structure of the formula (21) and allowing any Δ (−1 ≦ △ _fi , △ _gi ≦ 1) using a so-called μ design method. Derive.

[0096]

(Equation 16)

[0097]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an ABS control apparatus according to an embodiment of the present invention will be described in detail with reference to the drawings.

[First Embodiment] AB according to the present embodiment
FIG. 1 shows the configuration of the S control device.

As shown in FIG. 1, A according to the present embodiment
The BS control device comprises: a wheel speed detecting means 10 for detecting a wheel speed at every predetermined sample time τ;
Processing means (first-order or higher (first-order in the present embodiment) high-pass filter) 11 for processing (high-pass filter processing) the wheel speed detected by the above so that a steady component with the wheel acceleration is removed. Torque gradient increase / decrease determination means 1 for setting the value of forgetting factor λ in accordance with the increase / decrease in torque gradient estimated value
2A; a torque gradient estimating unit 13 for estimating a braking torque gradient based on the forgetting factor λ set by the torque gradient increase / decrease determining unit 12A and the time-series data of the wheel speed subjected to the high-pass filtering by the processing unit 11; ABS control means 14 for calculating an operation signal for each wheel for ABS control based on the braking torque gradient estimated by the gradient estimating means 13, and each wheel based on the operation signal calculated by the ABS control means 14. ABS control valve 16 that performs ABS control by operating the brake pressure every time
And

The wheel speed detecting means 10 shown in FIG. 1 can be realized, for example, by the configuration shown in FIG. As shown in FIG. 4A, the wheel speed detecting means 10 includes a signal rotor 30 having a predetermined number of teeth cut at equal intervals and attached to rotate with the wheels, and a pickup coil 32 fixed to the vehicle body. And a permanent magnet 34 arranged to penetrate magnetic flux inside the pickup coil 32, and a frequency of an AC voltage generated in the pickup coil 32 which is connected to the pickup coil 32 and is generated at every sampling time τ. And a frequency detector 36 for outputting the output.

When the signal rotor 30 rotates with the rotation of the wheels, the signal rotor 30 and the pickup coil 3
The air gap between the two changes at a cycle corresponding to the rotation speed. Therefore, the magnetic flux of the permanent magnet 34 passing through the pickup coil 32 changes, and an AC voltage is generated in the pickup coil 32. Here, FIG. 4B shows a temporal change of the AC voltage generated in the pickup coil 32.

As shown in FIG. 4B, the frequency of the AC voltage generated in the pickup coil 32 decreases when the rotation speed of the signal rotor 30 is low, and increases when the rotation speed of the signal rotor 30 is high. Since the frequency of the 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 shown in FIG. 4A is attached to all of the first to fourth wheels, and for each wheel, the i-th wheel (i is the wheel number) from the output signal of the frequency detector 36. ,
i = 1,2,3,4) time series data of wheel speed ω _i [k]
(K is a sample time; k = 1, 2,...) Is detected.

Next, the configuration of the ABS control valve 16 will be described with reference to FIG.

As shown in FIG. 5, the ABS control valve 16
A control solenoid valve 132 for the right front wheel (hereinafter, “valve SFR”) and a control solenoid valve 13 for the left front wheel
4 (hereinafter, "valve SFL"), a control solenoid valve 140 for right rear wheel (hereinafter, "valve SRR"), and a control solenoid valve 142 for left rear wheel (hereinafter, "valve SRL"). It is comprised including.

Valve SFR, valve SFL, valve SR
R and the valve SRL are pressure-increasing valves 132a, 1
34a, 140a, 142a and pressure reducing valve 132
b, 134b, 140b and 142b, and are connected to front wheel cylinders 144 and 146 and rear wheel cylinders 148 and 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 for controlling opening and closing of a valve, respectively.
3, SRR controller 139, SRL controller 1
41.

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 reduce the pressure of the control solenoid valve. Controls opening and closing of valves.

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

As shown in FIG. 6, the system hydraulic circuit is provided with a reservoir 100 for storing the brake fluid for 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 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. I have.

Further, a pump 106 for pumping up brake fluid from the reservoir 100 and discharging high-pressure hydraulic fluid is provided in a pipe provided from the relief valve 104 side of the reservoir 100. 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. The pressure sensor 110 outputs a control signal for the pump 106 based on the oil pressure of the accumulator 108, and outputs a warning signal (ABS, T
(A prohibition signal of the RC control).

The control signal for the pump 106 is output to the high-pressure side pipe of the accumulator 108 when the hydraulic pressure of the accumulator 108 is low, and a warning signal (ABS, TRC control inhibition signal) is output when the hydraulic pressure is low. A pressure switch 112 is provided.

A master cylinder 114 for generating a hydraulic pressure according to the depression force applied to the brake pedal 118 is connected to another pipe extending from the reservoir 100. A brake booster 116 is provided between the master cylinder 114 and the brake pedal 118 to adjust and introduce the high oil pressure of the accumulator 108 to an oil pressure corresponding to the pedaling force to generate a brake assisting force.

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

Further, from the master cylinder 114, a front master pressure pipe 164 and a rear master pressure pipe 166 for supplying the hydraulic pressure (master pressure) of the master cylinder to the front and rear wheels, respectively, are provided. The front master pressure pipe 164 and the rear master pressure pipe 16
6 is provided with a P & B valve 120 for adjusting the brake hydraulic pressure of the rear system so that an appropriate braking force is distributed between the front and rear wheels. Note that the P & B valve 120 stops adjusting the pressure in the rear system when the front system is lost.

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

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

The booster pipe 168 has a control solenoid valve 132 for the right front wheel (“valve SF”).
R)) and the pressure increasing valve 134a of the control solenoid valve 134 for the left front wheel (“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 piping on the pressure supply side of the valves SFR and SFL, respectively. The pressure increasing pipe of the pressure increasing device 122 is further connected to the valve SA1 and the valve SA2. And
The pipe on the pressure supply side of the valve SA1 is connected to a front wheel cylinder 144 that applies a brake pressure to the brake disk 152 of the left front wheel.
It is connected to a front wheel cylinder 146 that applies brake pressure to a brake disk 154 on the right front wheel.

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

Further, 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”) and a control solenoid valve 142 for the left rear wheel (“valve SR
L ") is connected. Further, the pressure reducing valve 140b of the valve SRR and the valve SR
A low-pressure pipe 162 extending directly from the reservoir 100 is connected to the L pressure reducing valve 142b.

The piping on the pressure supply side of the valve SRR is connected to a rear wheel cylinder 148 for applying a brake pressure to the brake disk 156 of the right rear wheel.
The RL is connected to a rear wheel cylinder 150 that applies a brake pressure to a brake disk 158 of the left rear wheel.

In the normal brake mode, 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 ABS control mode, the high hydraulic pressure of the booster pipe 168 is reduced. The valves are switched so as to cover the valves SRL and SRR. That is, in the rear wheels, switching between the normal brake mode and the ABS control mode is performed collectively on the left and right. The torque gradient increase / decrease determining means 12A corresponds to the setting means of the present invention (particularly, the setting means in the second aspect of the present invention).

Next, the operation of the present embodiment will be described. In the ABS mode, the valves SA1 and SA2 in FIG. 6 close the valve on the pressure increasing device 122 side and open the valves on the valve SFR and valve SFL side. Further, 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 for each wheel at each sampling time τ, and outputs time series data ω _i [k] of the wheel speed for each wheel. The processing means 11 generates time-series data ω of the wheel speed of each wheel.
_i [k] is high-pass filtered.

Next, the torque gradient estimating means 13 _{calculates the} equation (10) based on the time series data ω _hi [k] of the wheel speeds subjected to the high-pass filtering in step 1 described above.
2) Calculate the equation, and then in the above step 2 (13)
The braking torque gradient is estimated from the recurrence formula of the formula. Steps 1 and 2 are sequentially repeated to obtain time-series data of the estimated braking torque gradient.

When the torque gradient estimating means 13 according to the present embodiment first estimates the braking torque gradient, it sets the forgetting factor λ to a predetermined reference value λ ₀ (for example, λ ₀ = 0.
98), and in the second and subsequent estimations, the estimation is performed by applying the forgetting coefficient λ set by the torque gradient increase / decrease determination means 12A.

The torque gradient increase / decrease determining means 12A according to the first embodiment sets the forgetting factor λ in accordance with the flow of the flowchart shown in FIG.

That is, first, the estimated value of the braking torque gradient one sample time ago is obtained from the torque gradient estimating means 13 (step 300), and the obtained estimated value of the braking torque gradient and the high-pass filter processed by the processing means 11 are obtained. Based on the wheel speed signal, a determination value G is calculated by the following equation (25) (step 302).

[0130]

[Equation 17]

Next, the judgment value G obtained as described above is 0
It is determined whether or not the value is greater than 0 (Step 304). If the determination value G is greater than 0 (Yes at Step 304), the forgetting factor λ has a maximum value or a predetermined value λ in the vicinity of the maximum value that is allowed as the forgetting factor. _B (for example, λ _B = 0.99) is set (step 306).

On the other hand, when the judgment value G is not larger than 0,
That is, when the judgment value G is 0 or less (step 304)
It is determined whether the determination value G is smaller than 0 (step 308). When the determination value G is smaller than 0 (step 308, affirmative determination), the minimum value of the forgetting factor λ that is allowed as the forgetting factor λ is obtained. A value or a predetermined value λ _S near the minimum value (for example, λ _S = 0.95) is set (step 31).
0).

Further, when the judgment value G is not smaller than 0, that is, when the judgment value G is equal to 0 (No at Step 308), the reference value λ ₀ is set to the forgetting coefficient λ (Step 312). The setting of the forgetting coefficient λ as described above is performed at each sample time.

Then, the ABS control means 14 applies the braking torque gradient estimated by the torque gradient estimating means 13 using the forgetting coefficient λ set by the above processing, and performs the processing in the flow of the flowchart of FIG. .

[0135] As shown in FIG.
Calculates the operation amount u (u _i : i = 1, 2, 3, 4) of each wheel at each sample time using the braking torque gradient at each sample time estimated by the torque gradient estimating means 13 (step 200). ).

That is, the state equations (19) and (20) are derived from the equations (14) to (18), and (1)
By designing a control system that allows an arbitrary Δ (−1 ≦ す_{る fi} , △ _gi ≦ 1) having the structure of the expression (21) that appears in the expressions 9) and (20) using the so-called μ design method. , (23)
The controller of the equation (24) is derived. And
The state value of the controller is represented by x _c in equation (24), and y
The operation amount u of the ABS control valve 16 is obtained by substituting the value of the braking torque gradient estimated by the torque gradient estimating means 13 into.

Next, the wheel number i is set to 1 (step 202), and it is determined whether or not the operation amount ui of the _i- th wheel is larger than a 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 a pressure increase signal (Step 206).

[0138] When the operation amount u _i is not greater than the positive reference value + e (step 204 negative determination), the operation amount u _i is determined whether the negative reference value -e smaller (Step 20
8). If the operation amount u _i is a negative reference value -e smaller (step 208 affirmative decision), an operation signal ABS control valve of the i wheel is set to decrease pressure signal (step 210).

[0139] When the operation amount u _i is not smaller than the negative reference value -e (step 208 negative decision), i.e., there is the operation amount u _i is a negative reference value -e above and the positive reference value + e
In the following cases, the operation signal of the ABS control valve for the i-th wheel is
It is set to a holding signal (step 212).

When the operation signal for the operation amount u ₁ of the first wheel is set as described above, the wheel number i is incremented by 1 (step 214), and then it is determined whether or not i exceeds 4 (step 214). Step 216). If i does not exceed 4 (step 216, negative determination), step 204
Return, to set the operation amount u _i the operation signals of the wheel number i is incremented in a similar manner.

When the wheel number i exceeds 4 (step 2
16 affirmative determination), that is, A for all of 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 setting of the operation signal and the transmission of the operation signal as described above are performed at each sample time.

When the operation signal for each wheel is transmitted in this manner, the ABS control valve 16 causes the SFR controller 131, SFL controller 133, SRR controller 139, and SRL controller 141 of FIG. 5 to respond to each operation signal. Valve SFR, valve SFL, valve SRR,
The opening and closing of the valve SRL is controlled.

That is, in the case of a pressure increase signal, the pressure increase side valve is opened and the pressure decrease side valve is closed. Thereby, the high oil pressure of the booster pipe 168 of FIG. 6 is applied to the corresponding wheel cylinder, and the braking force increases. Conversely, when the signal is a pressure reduction signal, the pressure increase valve is closed and the pressure reduction valve is opened. Thereby, the low oil pressure of the low pressure pipe 162 of FIG. 6 is applied to the corresponding wheel cylinder, and the braking force is reduced. Also,
At the time of the hold signal, the pressure increasing side valve and the pressure reducing side valve are simultaneously closed. As a result, the hydraulic pressure applied to the corresponding wheel cylinder is maintained, and the braking force is maintained.

(Embodiment) FIG. 9 shows the estimation result of the braking torque gradient during the operation of the antilock brake. The graph shown in the upper part of the figure shows the change in the wheel speed applied here, and the broken line in the graph shown in the lower part of the figure shows the braking torque estimated by the prior art. The gradient, that is, the transition of the braking torque gradient estimated by keeping the forgetting coefficient λ constant, and the solid line shows the braking torque gradient estimated by the torque gradient estimating means 13 of the present embodiment, ie, the braking torque. When the gradient changes to the increasing side, the forgetting factor λ
Is set to the maximum value of the allowable value or a value near the maximum value,
When the braking torque gradient changes to the decreasing side, the transition of the braking torque gradient estimated by setting the forgetting coefficient λ to the minimum value of the allowable value or a value near the minimum value is shown. In calculating the estimated value of the braking torque gradient, a low-pass filter is also performed after the high-pass filter processing by the processing unit 11 in order to remove a high-frequency vibration component of the wheel speed signal.

As shown in the drawing, the torque gradient estimating method according to the present embodiment has improved responsiveness of the estimation as compared with the conventional method, and the speed of detecting the tendency to lock the wheels is increased. You can see that it can be done.

On the other hand, FIG. 10 shows changes in wheel speed, brake pressure, and estimated value of the braking torque gradient when the torque gradient estimation method according to the present embodiment is applied to the antilock brake control device. I have.

As shown in the figure, when the estimated value of the braking torque gradient by the torque gradient estimating method shown in this embodiment is applied to the anti-lock brake control device, the tendency of the wheels to lock can be detected quickly (here, It is determined that the vehicle tends to be locked when the estimated value of the braking torque gradient becomes equal to or less than a predetermined value near 0), and when the vehicle approaches the limit, the pressure is maintained unnecessarily without increasing the pressure. Alternatively, the pressure can be reduced for a very short time.

The broken line in the figure is an image diagram when the pressure is reduced for a very short time, and it can be expected that a decrease in the wheel speed can be reduced by reducing the brake pressure by the pressure reduction. Therefore, the braking force or the lateral force can be secured. Also, the amount of oil drawn from the wheel cylinder during pressure reduction can be reduced,
The size of the actuator can be reduced by reducing the size of the reservoir.

As described above in detail, in the ABS control device according to the present embodiment, when the torque gradient estimated by the torque gradient estimating means 13 changes to an increasing side, that is, the tire tends to grip. If the value of the forgetting coefficient λ changes to the maximum value or a value near the maximum value of the allowable value, the estimated torque gradient changes to the side where the torque gradient decreases, that is, the wheels tend to lock or slip. In such a case, the forgetting factor λ is appropriately set according to the driving scene by setting the value of the forgetting factor λ to the minimum value or a value near the minimum value of the allowable value. , It is possible to realize both an increase in the detection speed and a stabilization of the estimated value of the torque gradient.

Further, in the ABS control device according to the present embodiment, the wheels are controlled so that the gradient of the braking torque estimated by the torque gradient estimating means 13 according to the present embodiment falls within a predetermined range including the reference value. Since the applied braking force is controlled, the antilock brake control can be accurately performed.

[Second Embodiment] In the first embodiment,
When the torque gradient estimated by the torque gradient estimating means 13 changes to the increasing side, the forgetting coefficient λ changes to the maximum value of the allowable value or a value near the maximum value, and the estimated torque gradient decreases to the decreasing value. In this case, the embodiment in which the forgetting coefficient λ is set to the minimum value of the allowable value or a value near the minimum value, that is, the embodiment of the invention according to claim 2 has been described. In the embodiment, when the vibration level of the wheel speed is equal to or higher than a predetermined level, the forgetting coefficient λ is set to the maximum value of the allowable value or a value near the maximum value, that is, the third embodiment. An embodiment of the present invention will be described.

First, the configuration of the ABS control device according to the second embodiment will be described with reference to FIG. Note that the same reference numerals as in FIG. 1 denote parts performing the same processing as in FIG. 1, and a description thereof will be omitted.

As shown in FIG. 11, in the ABS control device according to the second embodiment, the wheel speed for setting the value of the forgetting coefficient λ according to the vibration level of the wheel speed is used instead of the torque gradient increase / decrease determining means 12A. The only difference from the ABS control device according to the first embodiment is that the vibration level determination means 12B is applied. In addition, about the configuration other than this part,
Since it is the same as the ABS control device shown in the first embodiment, the description is omitted here. The wheel speed vibration level determining means 12B corresponds to the setting means of the present invention (particularly, the setting means according to the third aspect of the present invention).

Next, the operation of the present embodiment will be described. The operation of the parts other than the wheel speed vibration level determination means 12B is the same as that of the ABS control device according to the first embodiment, and the description thereof will be omitted.

The wheel speed vibration level determining means 12B according to the second embodiment sets the forgetting factor λ according to the flow of the flowchart shown in FIG.

That is, first, the vibration level GN of the wheel speed is calculated based on the wheel speed signal after the high-pass filter processing by the processing means 11 (step 400). In addition,
In the present embodiment, the square value of the value indicated by the wheel speed signal input from the processing means 11 is calculated as the vibration level GN.

Next, it is determined whether or not the vibration level GN is larger than a predetermined threshold value TH1 (step 402). If the vibration level GN is larger than the predetermined threshold value TH1 (step 40).
2 affirmative determination), the maximum value or a predetermined value λ _B near the maximum value that is allowed as the forgetting factor for the forgetting factor λ (for example, λ _B =
0.99) is set (step 404).

On the other hand, when the vibration level GN is not larger than the predetermined threshold value TH1, that is, when the vibration level GN is equal to or smaller than the predetermined threshold value TH1 (No at Step 402), the reference value λ ₀ is set to the forgetting coefficient λ (Step 406). ).
The predetermined threshold value TH1 is a value that can be regarded as a value when the vibration level GN exceeds this value when the vehicle is traveling on a rough road or traveling with a chain attached, and is determined in advance by experiments and computer simulations. The obtained values and the like can be applied.

The setting of the forgetting factor λ as described above is as follows.
This is performed at each sample time.

As described above in detail, in the ABS control apparatus according to the present embodiment, when the vibration level of the wheel speed is equal to or higher than the predetermined level, the value of the forgetting coefficient λ is set to the maximum allowable value or Since it is set to a value near the maximum value,
It is possible to stabilize the estimated value of the torque gradient at the time of traveling on a rough road or traveling with a chain attached.

Further, in the ABS control device according to the present embodiment, the wheels are controlled so that the gradient of the braking torque estimated by the torque gradient estimating means 13 according to the present embodiment falls within a predetermined range including the reference value. Since the applied braking force is controlled, the antilock brake control can be accurately performed.

[Third Embodiment] In the third embodiment, when the vibration level of the wheel speed is equal to or higher than a predetermined level, the value of the forgetting coefficient λ is set to the maximum allowable value or a value near the maximum value. If the vibration level is less than the predetermined level, and if the torque gradient estimated by the torque gradient estimating means 13 changes to an increasing side, the value of the forgetting coefficient λ is set to the maximum or maximum allowable value. An embodiment in which the value of the forgetting coefficient λ is set to a value near the minimum value or a value near the minimum value when the torque gradient changes to a value near the value or to a side where the torque gradient decreases, that is, a claim An embodiment of the invention described in Item 4 will be described.

First, the configuration of the ABS control device according to the third embodiment will be described with reference to FIG. Note that the same reference numerals as in FIG. 1 denote parts performing the same processing as in FIG. 1, and a description thereof will be omitted.

As shown in FIG. 13, the ABS control device according to the third embodiment includes a wheel speed vibration level determining means similar to that of the second embodiment between a processing means 11 and a torque gradient increase / decrease determining means 12A. Only the point in which the 12B is arranged is different from the ABS control device according to the first embodiment. Since the configuration other than this portion is the same as that of the ABS control device shown in the first embodiment, the description is omitted here.

The torque gradient increase / decrease determining means 12A and the wheel speed vibration level determining means 12B in this embodiment correspond to the setting means of the present invention (particularly, the setting means in the fourth aspect of the present invention).

Next, the operation of the present embodiment will be described. The operation of the parts other than the wheel speed vibration level determination means 12B is the same as that of the ABS control device according to the first embodiment, and the description thereof will be omitted.

The wheel speed vibration level judging means 12B according to the third embodiment performs processing according to the flow of the flowchart shown in FIG. Steps in FIG. 14 that perform the same processing as in FIG. 12 are assigned the same step numbers as in FIG. 12, and descriptions thereof are omitted.

The vibration level GN of the wheel speed is equal to the predetermined threshold value TH.
If it is not greater than 1, that is, if the vibration level GN is equal to or less than the predetermined threshold value TH1 (No at Step 402), the control unit instructs the torque gradient increase / decrease determining means 12A to execute a torque gradient increase / decrease determination process (see FIG. 7). In response to this, the torque gradient increase / decrease determining means 12A executes the torque gradient increase / decrease determination process in the same manner as in the first embodiment, thereby obtaining the forgetting factor λ according to the increase / decrease state of the torque gradient estimated value.
Set the value of.

The setting of the forgetting factor λ as described above is as follows.
This is performed at each sample time.

As described above in detail, in the ABS control device according to the present embodiment, when the vibration level of the wheel speed is equal to or higher than the predetermined level, the value of the forgetting coefficient λ is set to the maximum value or the maximum allowable value. When the vibration level is less than the predetermined level, when the estimated torque gradient changes to the increasing side, the value of the forgetting coefficient λ is set to the maximum value or the maximum allowable value. When the estimated torque gradient changes to a value close to the value, the forgetting coefficient λ
Is set to the minimum value of the allowable value or a value near the minimum value, so that the forgetting coefficient λ according to the traveling scene is set.
Are appropriately switched, it is possible to realize both an increase in the detection speed of the tendency to lock the wheels and a stabilization of the estimated value of the torque gradient.

Further, in the ABS control device according to the present embodiment, the wheels are controlled so that the gradient of the braking torque estimated by the torque gradient estimating means 13 according to the present embodiment falls within a predetermined range including the reference value. Since the applied braking force is controlled, the antilock brake control can be accurately performed.

In the first, second, and third embodiments, the case has been described in which the braking torque gradient during braking of the vehicle is estimated based on the time-series data of the wheel speed. However, the present invention is not limited to this. Instead, the driving torque gradient at the time of driving the vehicle may be estimated based on the time-series data of the wheel speed. In this case, the same effects as in the above embodiments can be obtained.

[Fourth Embodiment] In the fourth embodiment, when it is determined that the wheels have a tendency to lock, the value of the forgetting coefficient λ does not tend to be locked for a predetermined period from the time when the locked state is released. An embodiment in which control is performed so as to be smaller than a value in the case where the judgment is made, that is, claim 6.
An embodiment of the described invention will be described.

First, the configuration of the ABS control device according to the fourth embodiment will be described with reference to FIG. Note that the same reference numerals as in FIG. 1 denote parts performing the same processing as in FIG. 1, and a description thereof will be omitted.

As shown in FIG. 15, the ABS control device according to the fourth embodiment includes a state judging means 12C for judging the locked state of the wheels, and a torque gradient estimating means based on the judgment result by the state judging means 12C. An ABS control device according to the first embodiment is different from the ABS control device according to the first embodiment only in that an estimation control unit 12D for controlling the estimation operation of the torque gradient by the control unit 13 is added and the torque gradient increase / decrease determination unit 12A is omitted. .

In the ABS control device according to the fourth embodiment, the output end of the wheel speed detecting means 10 is branched into two,
One is connected to the processing means 11 and the other is connected to the input end of the state determination means 12C. Further, an output terminal of the state determination unit 12C is connected to an input terminal of the estimation control unit 12D.
The output terminal of the estimation control means 12D is the torque gradient estimation means 13
It is connected to the.

Next, the operation of the present embodiment will be described. The operation of the parts other than the state determination unit 12C and the estimation control unit 12D is the same as that of the ABS control device according to the first embodiment, and the description thereof will not be repeated.

The state determining means 12C determines the locked state of the wheels as follows.

[0179] That is, first, as shown in FIG. 16, the vehicle body by detecting the transition of the wheel speed when repeated increase and decrease of the braking force, to connect the valleys of the velocity v _w obtained from the remained constant gradient Estimate the speed v _v . Since the method of estimating the vehicle speed v _v is well known, a detailed description thereof will be omitted.

Next, the slip speed is calculated by subtracting the wheel speed from the estimated vehicle speed v _v , and if the calculated slip speed is equal to or more than a predetermined value, it is determined that the wheels have entered a locking tendency. Thereafter, when the calculated slip speed becomes less than the predetermined value, it is determined that the lock state has been released. As the predetermined value, a value obtained in advance by an experiment or computer simulation or the like can be applied as a value that can be regarded as a tendency to lock when the slip speed exceeds the value.

On the other hand, the estimation control means 12D according to the fourth embodiment performs processing in accordance with the flow of the flowchart shown in FIG.

First, it is determined whether or not the wheels have a tendency to lock based on the result of the determination by the state determination means 12C (step 500). Based on the determination result, the vehicle waits for the wheel to be released from the locked state (step 502), and the minimum value or a predetermined value λ _S near the minimum value that is allowed as the forgetting factor for the forgetting factor λ (for example, λ _S = 0) .95) is set (step 504).
The forgetting coefficient λ set here is output to the torque gradient estimating means 13, and the torque gradient estimating means 13 estimates the torque gradient using the input forgetting coefficient λ.

Next, the estimation control means 12D waits for the elapse of a predetermined time (step 506), and thereafter the forgetting coefficient λ
Is set to a predetermined reference value λ ₀ (for example, λ ₀ = 0.98) (step 508). The forgetting coefficient λ set here is output to the torque gradient estimating means 13, and the torque gradient estimating means 13 estimates the torque gradient using the input forgetting coefficient λ.

On the other hand, when it is determined in step 500 that the wheels are not in a tendency to lock (in the case of a negative determination), the processing of steps 502 to 508 is not performed.

As described above in detail, in the ABS control device according to the present embodiment, when it is determined that the wheels are in a locking tendency, the value of the forgetting coefficient λ is maintained for a predetermined period from the time when the wheels are released from the locked state. Is controlled so as to be smaller than the value when it is determined that there is no tendency to lock, so that the influence of the wheel speed on the estimated value of the torque gradient in the past locked state is reduced. Accordingly, the accuracy of estimating the torque gradient can be improved.

In the ABS control device according to the present embodiment, the wheels are controlled so that the gradient of the braking torque estimated by the torque gradient estimating means 13 according to the present embodiment falls within a predetermined range including the reference value. Since the applied braking force is controlled, the antilock brake control can be accurately performed.

In the present embodiment, the case where the forgetting coefficient λ is set to a constant value smaller than a normal value during a predetermined period after the wheel is released from the locked state has been described. It is not limited to, for example,
For a predetermined period after the wheel is released from the locked state, the forgetting coefficient λ may be set to gradually approach a normal value from a predetermined value smaller than the normal value.
In this case, the same effect as in the present embodiment can be obtained.

[Fifth Embodiment] In the fifth embodiment, when it is determined that the wheels are in the locking tendency, the estimation of the torque gradient by the torque gradient estimating means 13 is interrupted, and it is determined that the vehicle is released from the locking tendency. 7. An embodiment in which the estimation operation by the torque gradient estimating means 13 is controlled so that the estimation of the torque gradient by the torque gradient estimating means 13 is resumed at the point of time, that is, another embodiment of the invention according to claim 6. Will be described.

First, the configuration of the ABS control device according to the fifth embodiment will be described with reference to FIG. In FIG. 15, a portion performing the same processing as in FIG.
The same reference numerals are given and the description thereof will be omitted.

As shown in FIG. 18, the ABS control device according to the fifth embodiment is different from the estimation control means 12D in that when the wheels are determined to be locked, the torque gradient estimation means 13 Estimation control means 12D 'for controlling the estimation operation by the torque gradient estimating means 13 so that the estimation of the gradient is interrupted and the estimation of the torque gradient by the torque gradient estimating means 13 is resumed when it is determined that the tendency to lock is released. Is different from the ABS control device according to the fourth embodiment only in the point that is provided.

Next, the operation of the present embodiment will be described. The operation of the parts other than the estimation control means 12D 'is the same as that of the ABS control device according to the fourth embodiment, and the description is omitted here.

Estimation control means 12 according to the fifth embodiment
D 'performs processing according to the flow of the flowchart shown in FIG.

First, it is determined whether or not the wheels have a tendency to lock based on the result of determination by the state determination means 12C (step 600). If the wheels have a tendency to lock (step 600 affirmative determination), the torque gradient estimating means 13 is determined. Is instructed to suspend the estimation of the torque gradient (step 60).
2). As a result, the torque gradient estimating means 13 interrupts the torque gradient estimating operation.

Next, based on the result of the judgment by the state judging means 12C, the wheel is released from the locked state (step 604), and the torque gradient estimating means 13 is instructed to resume the estimation of the torque gradient (step 604). 606). Thus, the torque gradient estimating means 13 restarts the torque gradient estimating operation.

On the other hand, when it is determined in step 600 that the wheels do not tend to lock (in the case of a negative determination), the processing of steps 602 to 606 is not performed.

While the operation of estimating the torque gradient by the torque gradient estimating means 13 is suspended, the ABS control means 1
In 4, control is performed so that the wheel cylinder oil pressure is reduced at the time when it is detected that the wheels tend to lock, and thereafter the pressure is gradually increased.

As described above in detail, in the ABS control device according to the present embodiment, when it is determined that the wheels have a tendency to lock, the estimation of the torque gradient is interrupted, and it is determined that the wheels are released from the lock tendency. Since the estimation operation is controlled so that the estimation of the torque gradient is restarted at the point in time, the influence of the wheel speed on the torque gradient estimated value in the past locked state can be reduced. ,
As a result, the accuracy of estimating the torque gradient can be improved.

Further, in the ABS control device according to the present embodiment, the wheels are controlled such that the gradient of the braking torque estimated by the torque gradient estimating means 13 according to the present embodiment falls within a predetermined range including the reference value. Since the applied braking force is controlled, the antilock brake control can be accurately performed.

In the fourth and fifth embodiments, the wheels tend to lock based on the slip speed obtained by subtracting the wheel speed from the estimated vehicle speed in the state determining means 12C. Although the description has been given of the case where it is determined that the vehicle is released from the locked state, the present invention is not limited to this. For example, the following determination method can be applied. (Example of Lock Tendency Determination) When the estimated torque gradient obtained by the torque gradient estimating means is equal to or less than a predetermined value. Note that the predetermined value here is a value that can be considered as having a tendency to lock when the torque gradient estimated value becomes equal to or less than the value, and a value obtained in advance by an experiment or a computer simulation or the like is applied. Can be. -When the wheel deceleration (the deceleration side is positive) is equal to or greater than a predetermined value. The predetermined value here may be a value obtained in advance by an experiment or a computer simulation or the like as a value that can be considered as having a tendency to lock when the wheel deceleration exceeds the value. it can. When the wheel speed signal after the high-pass filter used for estimating the torque gradient becomes equal to or less than a predetermined value (absolute value is large). The predetermined value here may be a value obtained in advance by an experiment or a computer simulation or the like as a value that can be regarded as having a tendency to lock when the wheel speed signal becomes equal to or less than the value. it can. (Example of unlocking determination)-When the wheel speed signal after the high-pass filter processing used for estimating the torque gradient becomes a predetermined value or more, and the signal decreases. Here, the predetermined value is a value obtained in advance by an experiment or computer simulation as a value that can be regarded as a tendency to lock when the wheel speed signal becomes greater than or equal to the value and the signal decreases. Etc. can be applied.

When these methods are applied, the same effects as those of the fourth and fifth embodiments can be obtained.

In the fourth embodiment, when it is determined that the wheels have a tendency to lock, when it is determined that the value of the forgetting coefficient λ does not tend to be locked for a predetermined period from the time when the locked state is released. In the fifth embodiment, an example in which the control is performed so that the value is smaller than the value of the torque gradient estimating means 1 is performed when it is determined that the wheels have a tendency to lock.
3 and the torque gradient estimating means 13 is stopped when it is determined that the locking tendency has been released.
Each of the embodiments in which the estimation operation by the torque gradient estimating means 13 is controlled so as to restart the estimation of the torque gradient by the above is described. However, the present invention is not limited to this. That is,
When detecting that the wheels have a tendency to lock, the operation of estimating the torque gradient by the torque gradient estimating means 13 is interrupted,
The torque gradient estimating operation may be restarted when the lock state is released, and the torque gradient may be estimated by setting the value of the forgetting coefficient λ to a value smaller than a normal value for a predetermined period from this time. . In this case, an improvement in the estimation speed after the estimation restart can be expected.

FIG. 21 shows the transition of the estimated torque gradient value in this embodiment. As shown in the figure, according to this embodiment, even in the grip state where the wheel speed immediately after the pressure reduction of the wheel cylinder oil pressure is restored, the wheel speed in the past slip state is reflected, and the estimated value of the torque gradient is reduced. It is possible to solve the problem of the prior art that is smaller than the actual one, and to estimate the torque gradient with higher accuracy.

[0203]

According to the torque gradient estimating device of the present invention, a high-pass filter is applied to the time series data of the wheel speed detected by the detecting means for detecting the wheel speed at every predetermined sampling time. While processing, the value of the parameter indicating the degree of removing the past data is set according to the running state of the vehicle, the time series data of the wheel speed subjected to the high-pass filter processing, the wheel motion model, and the braking torque or the driving torque are Using a gradient model that changes linearly with the gradient of the torque against the slip speed as a gradient, and applying the online system identification method using the above parameters, the gradient of the torque is estimated. Scenes in which a large vibration component is superimposed on the wheel speed signal when the vehicle is running with the chain or chain attached, etc. It is possible to appropriately switch the values of the above parameters depending on the driving scene in which the detection of the tendency to slip and the tendency to slip is desired to be speeded up. The effect is obtained that both stabilization of the gradient estimated value can be realized.

According to the torque gradient estimating device of the fifth and sixth aspects, the torque gradient estimating means controls the torque gradient estimating operation based on the determination result of the wheel lock state. When the vehicle is released from the state in which the vehicle tends to lock, the torque gradient estimation operation is controlled so that the influence of the wheel speed on the torque gradient estimation value in the past state of the lock state is reduced. The effect is obtained that the estimation accuracy can be improved.

[0205] Further, according to the antilock brake control device of the present invention, the anti-lock brake control device includes the torque gradient estimating device according to the present invention, and the gradient of the braking torque estimated by the torque gradient estimating device includes a reference value. Since the braking force acting on the wheels is controlled so as to be within the range, an effect is obtained that antilock brake control can be accurately performed.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of an antilock brake control device according to a first embodiment.

FIG. 2 is a diagram showing a relationship between a slip speed, a braking torque, and a braking torque gradient.

FIGS. 3A and 3B show changes in braking torque F _i and braking torque gradient G _i as a function of slip speed, wherein FIG.
FIG. 7 is a diagram illustrating upper and lower limits of the fluctuation of the braking torque F _i , and FIG. 7B is a diagram illustrating the upper and lower limits of the fluctuation of the braking torque gradient G _i .

4A and 4B are diagrams for explaining a configuration of a wheel speed detection unit according to the embodiment, wherein FIG. 4A is a configuration diagram of the wheel speed detection unit, and FIG. 4B is a diagram illustrating a time of an AC voltage generated in a pickup coil; It is a figure showing a target change.

FIG. 5 is a diagram showing a configuration of an ABS control valve according to the embodiment.

FIG. 6 is a diagram showing a configuration of a system hydraulic circuit including an ABS control valve according to the embodiment.

FIG. 7 is a flowchart showing a flow of a torque gradient increase / decrease determination process according to the embodiment;

FIG. 8 is a flowchart showing a flow of ABS control according to the embodiment.

FIG. 9 is a graph showing a result of estimating a braking torque gradient in the first embodiment.

FIG. 10 is a graph showing changes in wheel speed, brake pressure, and estimated value of braking torque gradient when the torque gradient estimation method according to the first embodiment is applied to an antilock brake control device.

FIG. 11 is a diagram showing a configuration of an antilock brake control device according to a second embodiment.

FIG. 12 is a flowchart illustrating a flow of a wheel speed vibration level determination process according to the second embodiment.

FIG. 13 is a diagram illustrating a configuration of an antilock brake control device according to a third embodiment.

FIG. 14 is a flowchart illustrating a flow of a wheel speed vibration level determination process according to a third embodiment.

FIG. 15 is a diagram showing a configuration of an antilock brake control device according to a fourth embodiment.

FIG. 16 is a graph for explaining a method of estimating a vehicle body speed.

FIG. 17 is a flowchart illustrating a flow of an estimation control process according to the fourth embodiment.

FIG. 18 is a diagram showing a configuration of an antilock brake control device according to a fifth embodiment.

FIG. 19 is a flowchart illustrating a flow of an estimation control process according to the fifth embodiment.

FIG. 20 is a diagram showing a dynamic model of a vehicle to which the ABS control according to the embodiment is applied.

FIG. 21 is a graph showing a transition of a torque gradient estimation value when the torque gradient estimation method described in the fourth embodiment is combined with the torque gradient estimation method described in the fifth embodiment.

FIG. 22 is a graph used to explain a problem of the related art.

[Explanation of symbols]

 Reference Signs List 10 wheel speed detecting means 11 processing means 12A torque gradient increase / decrease determining means (setting means) 12B wheel speed vibration level determining means (setting means) 12C state determining means 12D, 12D 'estimation control means 13 torque gradient estimating means 14 ABS control means 16 ABS control valve

Continuing from the front page (72) Inventor Yuji Murakishi 41-cho, Yokomichi, Nagakute-cho, Aichi-gun, Aichi Prefecture Inside of Toyota Central R & D Laboratories Co., Ltd. No. 1 Inside Toyota Central Research Laboratories Co., Ltd. (72) Inventor Koji Umeno 41-cho, Yokomichi Omachi, Nagakute-cho, Aichi-gun, Aichi Prefecture No. 1 Inside Toyota Central Research Laboratories Co., Ltd. 2-1-1 Asahi-cho, Aisin Seiki Co., Ltd. (72) Inventor Kenji Asano 2-1-1, Asahi-cho, Kariya-shi, Aichi F-term in Aisin Seiki Co., Ltd. 3D046 BB28 HH36 HH52 JJ06 KK06

Claims (7)

[Claims] 1. A detecting means for detecting a wheel speed at every predetermined sample time; a processing means for performing high-pass filtering on time-series data of the wheel speed detected by the detecting means; and a degree of removing past data. Setting means for setting the value of the parameter indicating according to the running state of the vehicle, time series data of the wheel speed subjected to high-pass filtering by the processing means, the wheel motion model, and the braking torque or the driving torque to the slip speed A torque gradient estimating means for estimating the torque gradient by applying an on-line system identification method using the parameter, using a gradient model that changes linearly with the gradient of the torque as a gradient; Estimation device. 2. The method according to claim 1, wherein when the torque gradient estimated by the torque gradient estimating means changes to an increasing side, the setting unit sets the value of the parameter to a maximum value or a maximum value of a value permitted as the value of the parameter. When the torque gradient estimated by the torque gradient estimating means changes to a value near the minimum value, the value of the parameter is changed to a minimum value or a value near the minimum value allowed as the value of the parameter. The torque gradient estimating device according to claim 1, wherein each is set. 3. The method according to claim 1, wherein the setting unit sets the parameter value to a maximum value or a maximum value that is allowed as the parameter value when the vibration level of the wheel speed detected by the detection unit is equal to or higher than a predetermined level. The torque gradient estimating device according to claim 1, wherein the torque gradient estimating device is set to a value near the value. 4. When the vibration level of the wheel speed detected by the detecting means is equal to or higher than a predetermined level, the setting means sets the value of the parameter to a maximum value or a maximum value of a value permitted as the value of the parameter. When the vibration level is set to a value in the vicinity and the vibration level is less than the predetermined level, when the torque gradient estimated by the torque gradient estimating means changes to the increasing side, the value of the parameter is set as the value of the parameter. When the torque gradient estimated by the torque gradient estimating means changes to the maximum value or a value near the maximum value of the allowable value, the torque gradient estimated by the torque gradient estimating means decreases the value of the parameter to a value allowable as the value of the parameter. 2. The torque gradient estimating device according to claim 1, wherein the torque gradient estimating device is set to a minimum value or a value near the minimum value. 5. A detecting means for detecting a wheel speed at every predetermined sampling time; a processing means for performing high-pass filtering on time-series data of the wheel speed detected by the detecting means; and a high-pass filter by the processing means. The time series data of the processed wheel speed, the wheel motion model, and the gradient model in which the braking torque or the driving torque changes linearly as a gradient of the torque with respect to the slip speed as a gradient are used to determine the degree of removing past data. A torque gradient estimating means for estimating a torque gradient by applying an online system identification method using the indicated parameters; a state determining means for determining a lock state of the wheel; and Estimating control means for controlling the torque gradient estimating operation by the torque gradient estimating means. Torque gradient estimating apparatus. 6. The estimation control means interrupts the estimation of the torque gradient by the torque gradient estimation means when the state determination means determines that the wheels are in a locking tendency, and releases the vehicle from the locking tendency. At the time when the determination is made, the estimation operation by the torque gradient estimating means is controlled so that the estimation of the torque gradient by the torque gradient estimating means is restarted, or the wheel is determined to be locked by the state determining means. 6. The torque gradient estimating device according to claim 5, wherein in the case, the value of the parameter is controlled to be smaller than the value when it is determined that the parameter does not have the locking tendency for a predetermined period from the time when the lock state is released. 7. The torque gradient estimating device according to claim 1, wherein the gradient of the braking torque estimated by the torque gradient estimating device is a value in a predetermined range including a reference value. Control means for controlling a braking force acting on the wheels.