JP2002160620A - Physical value estimating device and road surface friction condition estimating device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a physical value estimating device capable of estimating a physical value showing the easiness of slippage between a tire and a road surface when a slip ratio is almost zero, without requiring the maximum road surface friction coefficient, and a road surface friction condition estimating device capable of highly precisely estimating a road surface friction condition during braking or driving. SOLUTION: In accordance with a vehicle speed, a road surface μgradient is estimated by road surface μ gradient estimating means 14, and the braking force of a vehicle is estimated by a high pass filter 16 and braking force estimating means 18. In accordance with a model showing a relationship between the road surface μ gradient and the braking force of the vehicle, the estimated road surface μgradient and the braking force, the estimated road surface μgradient is corrected by road surface μgradient correcting means 20 to establish a road surface μ gradient when the slip ratio is almost zero. In accordance with the vehicle speed, the vibration levels of wheels are estimated by a band pass filter 22 and vibration level operating means 24, and in accordance with the corrected road surface μ gradient and the vibration levels, the road surface friction condition is estimated by road surface friction condition estimating means 26.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a physical quantity estimating apparatus and a road friction state estimating apparatus.
A physical quantity estimating apparatus for estimating a physical quantity indicating the ease of slip between the tire and the road surface such as a gradient, and a road surface friction state estimating apparatus for estimating a friction state of the tire on the road surface based on the physical quantity estimated by the physical quantity estimating apparatus Related to the device.

[0002]

2. Description of the Related Art
ABS (Anti-lock Braking System), TCS (Tracti
on Control System), low air pressure warning system, etc.
Systems that improve vehicle performance and safety have been developed. These systems provide a physical quantity that indicates the ease of slip between the tire and the road surface during driving or braking, for example, the road surface μ.
The gradient is estimated, and various controls are performed based on the estimated value.

[0003] The road surface μ gradient is a gradient of the μ-S characteristic indicating a road surface frictional force with respect to a slip ratio, and varies depending on a running state or a road surface state. Therefore, conventionally, there has been a technique for determining a friction state of a road surface based on a road surface μ gradient.

[0004] As this kind of technology, Japanese Patent Application No. Hei 11-315 is disclosed.
No. 258 focuses on the fact that the vibration level of the wheel speed increases when the road surface layer collapses during running on snowy roads or dirt roads, and based on the road surface μ gradient and the wheel speed vibration level, the friction state of the road surface is determined. A technique has been proposed for estimating.

[0005] However, Japanese Patent Application No. Hei 11-315258.
In the technology proposed in No. 5, the road surface μ gradient is a state quantity that changes in accordance with the braking force or the driving force applied to the tires in addition to the slipperiness of the road surface on which the vehicle is traveling. In the case of applying this technology, the road surface μ gradient is reduced according to the braking force or the driving force, and as a result, the estimated road surface friction state is erroneously determined to be a slippery road surface compared to the actual state. There was a problem that sometimes.

On the other hand, Japanese Patent Application No. 2000-314399 discloses that
Estimating the road μ gradient near the origin (slip ratio is almost zero) based on the braking or driving force and the braking or driving force based on the brush model, with the main purpose of estimating the tire pressure. The technology to do this is described.

[0007] In order to improve the estimation accuracy of the road surface friction state, a road surface μ gradient obtained by the technique described in Japanese Patent Application No. 2000-314399 has been proposed in Japanese Patent Application No. 11-315258. Although a technology for estimating the road surface friction state by applying the technology to the existing technology is also conceivable, this technology has the following problems.

In the technique described in Japanese Patent Application No. 2000-314399, the maximum road surface friction coefficient μ between the tire and the road surface is required as a parameter for estimating the road surface μ gradient near the origin. In order to estimate the frictional state of the vehicle, it is necessary to treat the maximum road friction coefficient μ as an unknown value, and simply apply this technology to the technology proposed in Japanese Patent Application No. 11-315258. I can't.

The present invention has been made in view of the above facts, and estimates a physical quantity indicating the ease of slip between a tire and a road surface when the slip ratio is substantially zero without requiring a maximum road surface friction coefficient. It is an object of the present invention to provide a physical quantity estimating device capable of estimating a road surface friction state at the time of braking or driving and a road surface friction state estimating device capable of estimating the road surface friction state at a high accuracy.

[0010]

First, the principle of physical quantity estimation in the physical quantity estimation apparatus according to the present invention will be described. FIG. 13 shows an example of the relationship between the road surface μ gradient (corresponding to the physical quantity of the present invention) and the braking / driving force in the brush model. As shown in the drawing, the relationship between the road surface μ gradient and the braking / driving force can be approximated by a substantially linear function in other regions except the region near the limit where the road surface μ gradient becomes small.

Therefore, as an example, as shown in FIG.
By assuming a model in which the road surface μ gradient is expressed as a function that monotonically decreases as the braking / driving force increases, the road surface μ gradient during braking or driving and the braking force or driving force at this time are separated. For example, by drawing a straight line (or a curve) in the model passing through this point, it becomes possible to derive a road surface μ gradient near the origin. Note that FIG.
Shows the relationship between the slip speed and the braking / driving force when the model shown in FIG. 14 is used.

[0012] Based on the above principle, the physical quantity estimating apparatus according to the first aspect of the present invention provides a wheel speed detecting means for detecting a wheel speed, and a tire and a road surface detected based on the wheel speed detected by the wheel speed detecting means. Physical quantity estimating means for estimating a physical quantity indicating the slipperiness between the vehicle, braking / driving force deriving means for deriving a braking force or a driving force of the vehicle, and a physical quantity indicating the slipperiness between the tire and the road surface and the vehicle braking / driving force. The physical quantity estimating means based on a model indicating a relationship between power or driving force, a physical quantity estimated by the physical quantity estimating means, and a braking force or driving force derived by the braking / driving force deriving means. Physical quantity correction means for correcting the physical quantity estimated by the above so as to be a physical quantity when the slip ratio is substantially zero.

According to the physical quantity estimating apparatus of the first aspect,
The wheel speed is detected by the wheel speed detecting means, and the physical quantity indicating the ease of slip between the tire and the road surface is estimated by the physical quantity estimating means based on the detected wheel speed. Note that any existing estimation method can be applied to the estimation of the physical quantity.

Further, in the present invention, the braking / driving force of the vehicle is derived (estimated or detected) by the braking / driving force deriving means, and the physical quantity indicating the ease of slip between the tire and the road surface is calculated by the physical quantity correcting means. Based on a model indicating the relationship between the braking force or the driving force of the vehicle, the physical quantity estimated by the physical quantity estimating means, and the braking force or the driving force derived by the braking / driving force deriving means, The physical quantity estimated by the physical quantity estimating means is corrected so as to be the physical quantity when the slip ratio is substantially zero.

As described above, according to the physical quantity estimating apparatus of the first aspect, the physical quantity indicating the ease of slip between the tire and the road surface is estimated based on the wheel speed, and the braking force or the driving force of the vehicle is calculated. Derived, a model showing the relationship between the physical quantity indicating the ease of slip between the tire and the road surface and the braking force or driving force of the vehicle, the estimated physical quantity, and the derived braking force or driving force Based on the above, since the estimated physical quantity is corrected to be a physical quantity when the slip ratio is substantially zero, the tire and the road surface when the slip ratio is substantially zero without requiring a maximum road surface friction coefficient. It is possible to estimate a physical quantity indicating the ease of slipping between.

[0016] As in the second aspect of the present invention, the model according to the first aspect of the invention is characterized in that the physical quantity indicating the ease of slipping between the tire and the road surface is increased by an increase in the braking force or the driving force of the vehicle. It is preferable to use a model represented as a function that monotonically decreases. Thus, it is possible to easily and easily estimate a physical quantity indicating the ease of slip between the tire and the road surface when the slip ratio is substantially zero.

According to a third aspect of the present invention, when the braking force is derived by the braking / driving force deriving means, the physical quantity indicating the ease of slipping between the tire and the road surface of the present invention is a braking force gradient. Alternatively, when the driving force is derived from the road surface μ gradient, the driving force gradient or the road surface μ gradient may be used.

On the other hand, to achieve the above object, a road surface friction state estimating device according to a fourth aspect of the present invention includes a physical quantity estimating device according to any one of the first to third aspects, and the wheel speed detecting means. Vibration level estimation means for estimating a vibration level indicating the magnitude of vibration of the wheel based on the detected wheel speed, and a physical quantity corrected by the physical quantity correction means and a vibration level estimated by the vibration level estimation means. Road friction state estimating means for estimating a friction state of the tire with respect to the road surface based on the estimated friction state.

According to a fourth aspect of the present invention, the physical quantity is estimated by the physical quantity estimating apparatus according to any one of the first to third aspects, and the physical quantity estimating apparatus is estimated by the vibration level estimating means. The vibration level indicating the magnitude of the vibration of the wheel is estimated based on the wheel speed detected by the wheel speed detection means provided in the, the road surface friction state estimation means, by the physical quantity correction means provided in the physical quantity estimation device The friction state of the tire on the road surface is estimated based on the corrected physical quantity and the vibration level estimated by the vibration level estimating means.

As described above, according to the road surface friction state estimating apparatus of the fourth aspect, the tire quantity is estimated based on the physical quantity estimated by the physical quantity estimating apparatus according to the present invention, that is, the physical quantity when the slip ratio is substantially zero. Since the friction state on the road surface is estimated, the road surface friction state during braking or driving can be estimated with high accuracy.

[0021]

Embodiments of the present invention will be described below in detail with reference to the drawings.

As shown in FIG. 1, a road surface friction state estimating apparatus 10 according to the present embodiment includes a wheel speed sensor 12 for detecting a rotation speed of a wheel provided on a vehicle, and a wheel speed sensor 12 for detecting a rotation speed between the wheel and the road surface. Road surface μ gradient estimating means 14 for estimating a road surface μ gradient which is a physical quantity indicating slipperiness, a high-pass filter 16 for pseudo-differentiating a wheel speed signal output from a wheel speed sensor 12, and a high-pass filter 16 The braking force estimating means 18 for estimating the braking force of the vehicle based on the output signal, and the road μ gradient estimated by the road μ gradient estimating means 14 based on the braking force estimated by the braking force estimating means 18 are slipped. Road surface gradient correction means 20 for correcting the road surface μ gradient when the rate is substantially zero.

The wheel speed sensor 12 is used as the wheel speed detecting means of the present invention, the road surface μ gradient estimating means 14 is used as the physical quantity estimating means of the present invention, the braking force estimating means 18 is used as the braking / driving force deriving means of the present invention, and the road surface μ is obtained. The gradient correction means 20 corresponds to the physical quantity correction means of the present invention.

The road surface friction state estimating apparatus 10 determines, based on the wheel speed signal output from the wheel speed sensor 12, that the frequency including at least one resonance point or at least one anti-resonance point is in a predetermined range larger than the low frequency range. A band-pass filter 22 for extracting a wheel speed signal; a physical quantity representing the magnitude of vibration of the wheel based on the wheel speed signal extracted by the band-pass filter 22; in this embodiment, a vibration level calculation for calculating a vibration level Means 24, a road friction state estimating means 26 for estimating a friction state of the wheel with respect to the road surface based on the road μ gradient derived by the road μ gradient correcting means 20 and the vibration level calculated by the vibration level calculating means 24; , Is provided.

The vibration level calculating means 24 corresponds to the vibration level estimating means of the present invention, and the road surface friction state estimating means 26 corresponds to the road surface friction state estimating means of the present invention.

The wheel speed sensor 12 generates a wheel speed signal ω ₁ corresponding to the wheel speed of each wheel to generate the road surface μ gradient estimating means 1.
4, high-pass filter 16, and band-pass filter 2
Feed to 2.

As shown in FIG. 2, the road surface μ gradient estimating means 14 receives a road surface disturbance ΔT _d from a wheel speed signal ω ₁ of each wheel detected by the wheel speed sensor 12 and outputs a response output of a wheel resonance system. a preprocessing filter 14A for detecting a wheel speed vibration [Delta] [omega ₁ of the wheels, and the transfer function identifying means 14B for the transfer function of each wheel so as to satisfy the detected wheel speed vibration [Delta] [omega ₁ identified by the least square method, identification Μ gradient calculating means 14C for calculating the gradient of the friction coefficient μ between the tire and the road surface for each wheel based on the transfer function thus obtained.

The pre-processing filter 14A includes a band-pass filter that passes only a frequency component in a certain band around a frequency expected to be the resonance frequency of the wheel resonance system,
It is composed of a high-pass filter or the like that passes only high-band frequency components including the resonance frequency component. Here, a parameter that defines the frequency characteristics of the band-pass filter or the high-pass filter is fixed to a constant value.

The pre-processing filter 14A extracts and outputs only the signal from which the DC component has been removed, that is, only the wheel speed vibration Δω ₁ around the wheel speed signal ω ₁ .

Here, the transfer function F (s) of the pre-processing filter 14A is expressed by equation (1).

[0031]

(Equation 1)

Where c _i is the coefficient of the filter transfer function,
s is a Laplace operator.

Next, an operation formula on which the transfer function identification means 14B depends is derived. The calculation of the pre-processing filter 14A is included in the calculation of the transfer function identification means 14B and executed.

The transfer function to be identified is a secondary model in which the road surface disturbance ΔT _d is used as an excitation input and the wheel speed vibration Δω ₁ detected by the pre-processing filter 14A at this time is used as a response output. That is, the vibration model of Expression (2) is assumed.

[0035]

(Equation 2)

Here, v is observation noise included when a wheel speed signal is observed. By transforming equation (2), the following equation (3) is obtained.

[0037]

(Equation 3)

First, the equation obtained by multiplying the equation (3) by the preprocessing filter of the equation (1) is discretized. At this time, Δω
₁ , ΔT _d , v are the discrete data Δω ₁ (k), ΔT _d (k), v sampled every sampling period T _s
(K) (k is a sampling number: k = 1, 2, 3,...)
・). The Laplace operator s can be discretized using a predetermined discretization method. In the present embodiment, as an example, it is assumed that discretization is performed by bilinear transformation of the following equation (4). Note that d is a one-sample delay operator.

[0039]

(Equation 4)

Since the order m of the pre-processing filter is desirably 2 or more, m is set to 2 in consideration of the calculation time, thereby satisfying the expression (5) (however, the expressions (6) to (9) are satisfied). obtain.

[0041]

(Equation 5)

Further, in order to identify the transfer function from each data of the wheel speed vibration Δω ₁ based on the least square method,
The equation (5) is modified as the equation (10) (however, the equation (11) is satisfied) so that the parameter to be identified is in the form of a linear function. Note that “ ^T ” is the transpose of the matrix.

[0043]

(Equation 6)

In the above equation, θ is a parameter of the transfer function to be identified.

The transfer function identification means 14B estimates the unknown parameter θ by applying the least squares method to each data obtained by sequentially applying the discretized data of the detected wheel speed vibration Δω _{1 to} the equation (10). Then, the transfer function is identified.

Specifically, the detected wheel speed vibration Δω ₁
To the discrete data Δω (k) (k = 1, 2, 3,...)
, The data is sampled at N points, and the parameter θ of the transfer function is estimated using the least squares method of the equation (12).

[0047]

(Equation 7)

Here, the crowned amount of the symbol “＾” is defined as the estimated value. The least square method may be calculated as a recursive least square method for obtaining the parameter θ by the following recurrence formulas (13) to (15).

[0049]

(Equation 8)

Here, ρ is a so-called forgetting coefficient, which is usually set to a value of 0.95 to 0.99. At this time, the initial value may be set as in the following equation (16).

[0051]

(Equation 9)

As a method of reducing the estimation error of the above least square method, various modified least square methods may be used.
Here, an example will be described in which an auxiliary variable method, which is a least squares method in which auxiliary variables are introduced, is used. According to this method, (1)
At the stage when the relationship of the expression (0) is obtained, the parameters of the transfer function are estimated using the following expression (17), using m (k) as an auxiliary variable.

[0053]

(Equation 10)

The sequential operation is as shown in the following equations (18) to (20).

[0055]

[Equation 11]

The principle of the auxiliary variable method is as follows.
That is, when equation (17) is substituted into equation (18), (2)
Equation 1) is obtained.

[0057]

(Equation 12)

If the auxiliary variable is selected such that the second term on the right side of the equation (21) becomes zero, the estimated value of θ matches the true value of θ.
Thus, as an auxiliary variable, ζ (k) = [− ξ _y1 (k) −
ξ _y2 (k)] ^T that is delayed so as not to have a correlation with the equation error r (k) is used. That is, m (k) = [− ξ _y1 (k−L) −ξ _y2 (k−L)] ^T. Here, L is a delay time.

After identifying the transfer function as described above,
The μ gradient calculating means 14C calculates a physical quantity related to the road μ gradient D ₀ by using the equation (22).

[0060]

(Equation 13)

As described above, the road surface μ gradient estimating means 14
The μ gradient D ₀ can be calculated by the equation (22).

On the other hand, the braking force estimating means 18 calculates the wheel speed signal ω ₁ , which is pseudo-differentiated by the high-pass filter 16.
, And calculates a wheel deceleration signal as an indicator of the braking state, and supplies the signal to the road surface μ gradient correction means 20.

The road μ gradient correcting means 20 stores in advance a slope SL of a straight line indicating the relationship between the road μ gradient and the braking / driving force approximated by a linear function as shown in FIG. Note that the slope SL is, for example, the road surface μ in the brush model.
Data indicating the relationship between the gradient and the braking / driving force is obtained by an experiment, computer simulation, or the like, and a graph as shown in FIG. 13 is created. It can be obtained by, for example, obtaining a slope of 0 to 600 (N). It is needless to say that the model assumed here is not limited to the brush model, and other various models can be applied.

Then, the road surface μ gradient correcting means 20 calculates the gradient (road surface μ gradient) K _S of the braking force with respect to the slip ratio at the origin by the following equation (23).

K _S = D ₀ + (SS × SL) (23) where SS is the wheel deceleration indicated by the wheel deceleration signal supplied by the braking force estimating means 18.

That is, in the road surface μ gradient correcting means 20 according to the present embodiment, the value obtained by multiplying the wheel deceleration SS by the gradient SL is assumed to be a decrease in the road surface μ gradient due to braking. By adding the value to the road μ gradient D ₀ estimated by the road μ gradient estimating means 14, the road μ gradient K _S with respect to the slip ratio at the origin is derived.

As described above, adding the value obtained by multiplying the wheel deceleration SS by the gradient SL to the road μ gradient estimated by the road μ gradient estimating means 14 means that the road μ gradient and the braking state amount ( Assuming that the relationship with the wheel deceleration (SS) is described by a linear function as shown in FIG. 14, the road surface μ gradient when the braking force is zero is estimated.

By the way, when the wheel speed signal ω ₁ is decomposed in accordance with the frequency, as shown in FIG.
It has two anti-resonance points. Resonance point of frequency is smaller sides of the two resonance points is a longitudinal resonance point based on such a tire inertia, frequency is f ₁ (15~20) Hz. Further, the resonance point of the high frequency side of the two resonance points is a resonance point twisting based on the air pressure and tire rubber elasticity of the tire, the frequency is f ₃ (35~40) Hz. And
A wheel speed signal has a anti-resonance point of the dead zone with respect to various signals, the frequency f ₂ (20~25) Hz. The band-pass filter 22 according to the present embodiment has a wheel speed signal ω
₁ , a predetermined range Δf including the torsional resonance point (frequency f ₃ )
The wheel speed signal is extracted. The band-pass filter 22 may extract a wheel speed signal in a predetermined range Δf including a front and rear resonance point and an anti-resonance point other than the torsional resonance point.

The vibration level calculating means 24 calculates the vibration level G (N) of the wheel speed signal determined by the following equation. Note that the output of the bandpass filter 22 is ω (k).

[0070]

[Equation 14]

Note that the vibration level calculating means 24 actually calculates the

[0072]

(Equation 15)

Are sequentially calculated. The reason why the vibration level is calculated by the vibration level calculating means 24 is as follows. That is, as shown in FIG.
If the vehicle gets over a protrusion while traveling on an asphalt road, the wheel speed signal becomes smaller, and the estimated value of the road surface μ gradient estimated at this time becomes smaller, which is almost the same as the estimated value of the road surface μ gradient on a low μ road. Thus, there is a case where it is determined that the vehicle is traveling on a low μ road when the vehicle goes over a protrusion. As described above, the road surface μ gradient may not be properly estimated due to, for example, riding over a protrusion.

On the other hand, when the vibration level and the estimated value of the road μ gradient obtained in this way are plotted on an asphalt road surface and a low μ road surface, as shown in FIG. In addition to being clearly distinguished and recognized, even if a protrusion is passed over the asphalt road surface, it is clearly distinguished and recognized from a region on a low μ road. In view of the above fact, when the relationship between the vibration level and the estimated value of the road μ gradient was determined for each of various road surface conditions, as shown in FIG. For example, low μ road, asphalt road, cobblestone road,
The results of the experiment showed that it was possible to clearly distinguish and recognize the compacted snow road, the gravel and irregular roads for each area. That is, for example, a snowy road has a slightly smaller estimated value of the road μ gradient than a high μ road (asphalt road, cobblestone road), but a difference appears in the area due to a difference in vibration level.
Gravel roads are located in areas where the vibration level is even higher.
The estimated value of the gradient is lower than the high μ road.

Therefore, the road surface friction state estimating means 26 calculates the vibration level calculated by the vibration level calculating means 24 and the estimated value of the road μ gradient corrected by the road μ gradient correcting means 20 and the estimated value of both of the road surface conditions. Relationship (see Figure 5)
, The friction state of the wheel on the road surface is estimated.

In the present embodiment, the band-pass filter 22 converts the wheel speed signal at _one of the two resonance points or the wheel speed signal at the anti-resonance point out of the wheel speed signal ω _1. Extract. That is, as shown in FIG.
A wheel speed signal whose frequency is larger than the low frequency region is detected. The reason why the wheel speed signal whose frequency is higher than the low frequency region is detected as described above is as shown in FIG.
This is because the influence of the wheel speed signal whose frequency is smaller than the low frequency region is large because the error is large, and the road surface friction state cannot be accurately estimated as shown in FIG.

That is, FIG. 6 shows the frequency characteristics of the wheel speed signal when traveling on an asphalt road or a low μ road. Note that the low μ road has very small irregularities as compared with the asphalt road.

Looking at the resonance characteristics at around 40 Hz, it can be understood that the vibration level of the low μ road is smaller than that of the asphalt road, and that the low μ road reflects the unevenness of the road surface well. Also, the strength of resonance is small, and it can be estimated that the road is a low μ road by this method.

Looking at the vibration components in the low frequency region (5 Hz or less), the vibration level in this region increases in proportion to the reciprocal of the road surface μ gradient, so that the vibration of the wheel speed signal in the low frequency region on the low μ road is reduced. The level is higher than the vibration level of the wheel speed signal in the low frequency region of the asphalt road.

Accordingly, when the road surface state is estimated by calculating the vibration level and the road μ gradient based on the wheel speed signal in the low frequency range, the following problem occurs.

That is, the vibration component in the low frequency range depends on the road surface μ gradient and does not reflect the unevenness of the actual road surface. For this reason, many adaptation man-hours are needed to create a map for road surface determination.

Further, since the wheel speed signal in the low frequency range includes a vibration component due to the rolling and pitching motion of the vehicle, it may not be possible to accurately determine the road surface condition depending on running conditions.

Further, for the above reason, the judgment map must be prepared according to the physical parameters of the tires and the vehicle. However, it is difficult to cope with the case where the specifications of the vehicle are changed by changing tires or the like. May not be able to determine the road surface condition.

On the other hand, in the present embodiment, since the wheel speed signal in a region larger than the low frequency region is targeted, the above-described problem does not occur, and the road surface state can be accurately determined.

FIG. 8 shows changes in the wheel speed with time on the high μ road and the low μ road during braking (FIG. 8 (A)) and the road surface μ gradient on the high μ road and the low μ road at this time. Correction means 20
(I.e., the road μ gradient estimated by the road μ gradient estimating means 14) before the correction (FIG. 8).
(B)) and a change in the road surface μ gradient after the correction by the road surface μ gradient correcting means 20 (FIG. 8C).

As shown in FIG. 8B, when the correction according to the braking state is not performed, it can be seen that the road μ gradient becomes small at the time of braking, and the difference due to the road surface state becomes small. In this case, the road μ gradient at the time of braking on the high μ road falls to almost the same level as that of the low μ road, and as a result, the road surface friction state estimating means 26 incorrectly determines that the road is a low μ road.

On the other hand, when the road surface μ gradient is corrected by the road surface μ gradient correcting means 20, FIG.
As shown in (1), the decrease in the road surface μ gradient due to braking is compensated, and a substantially constant value (road surface μ gradient in a steady running state) is obtained. Therefore, in this case, the road surface μ gradient during braking on the high μ road is at a sufficiently large level as compared with the low μ road, and the road surface friction state estimating means 26 appropriately sets the high μ road.
It is determined to be a road.

The wheel speed sensor 12, the road μ gradient estimating means 14, the high-pass filter 16, and the braking force estimating means 1
8. The portion constituted by the road surface μ gradient correction means 20 corresponds to the physical quantity estimation device of the present invention.

As described above in detail, the road surface friction state estimating apparatus 10 according to the present embodiment estimates the road surface μ gradient based on the wheel speed, derives the braking force of the vehicle, and obtains the road surface μ.
Based on the model indicating the relationship between the gradient and the braking force or the driving force of the vehicle, the estimated road surface μ gradient, and the derived braking force, the estimated road surface μ gradient is calculated based on the slip ratio. Is corrected so as to be a physical quantity when the road surface friction coefficient is substantially zero. Therefore, the road surface μ gradient when the slip ratio is substantially zero can be estimated without requiring a maximum road surface friction coefficient.

In the road surface friction state estimating apparatus 10 according to the present embodiment, a model in which the road surface μ gradient is expressed as a function that monotonously decreases with an increase in the braking force or the driving force of the vehicle is applied as the model. Therefore, the road μ gradient when the slip ratio is substantially zero can be easily and quickly estimated.

In the road surface friction state estimating apparatus 10 according to the present embodiment, the friction state of the tire against the road surface is estimated based on the road μ gradient when the slip ratio is substantially zero. The road surface friction state at the time can be estimated with high accuracy.

In this embodiment, the case where the road surface μ gradient is estimated by the road surface μ gradient estimating means 14 as the physical quantity of the present invention has been described. However, the present invention is not limited to this. As a physical quantity that can be equivalently treated as, for example, a braking force gradient which is a slope of a tangent of a curve representing a relationship between a slip ratio or a slip speed and a braking force, and a relationship between a slip ratio or a slip speed and a driving force. A driving force gradient or the like, which is a gradient of a tangent to a curve, can also be applied.

Hereinafter, as an example, a procedure for estimating the braking force gradient when the braking force gradient is applied as a physical quantity according to the present invention will be described. Note that the estimation procedure described here is
The transfer characteristic from the road surface disturbance to the wheel speed is approximated to a first-order lag model, the band frequency is estimated from the frequency response of the first-order lag model based on the time-series data of the wheel speed, and the braking force gradient is estimated from the estimated band frequency. Is estimated.

[0094] As shown in FIG. 9, an example of embodiment in this case is based on the time-series data of wheel speed indicated by the wheel speed signals omega _1, the transfer characteristics from road surface disturbance to the wheel speed in first-order lag model In a gain diagram representing the frequency response of the model when approximated, a band frequency estimating means 14A 'for estimating a frequency at which the gain changes from a constant value as a band frequency (wheel speed frequency characteristic amount) is stored in advance. Based on the map representing the relationship between the band frequency and the braking force gradient,
And a braking force gradient estimating means 14B 'for estimating a braking force gradient with respect to the estimated band frequency.

Although FIG. 9 shows the configuration of one wheel, in the case of a vehicle having a plurality of wheels, such as a four-wheel vehicle, the configuration shown for each wheel is provided.

In the band frequency estimating means 14A ', it is assumed that a white disturbance, which is a disturbance including all frequencies, is inputted to the tire from the road surface, and the band frequency of the first-order lag model is identified using the least square method.

FIG. 10 is an algorithm for identifying the band frequency, and FIG. 11 is a gain of the first-order lag model corresponding to the band frequency identified by the algorithm of FIG. 10 when a white disturbance is applied to the wheel full model. FIG.

First, an algorithm for identifying a band frequency will be described with reference to FIG. Step 3
At 00, data obtained by adding white disturbance to the time-series data of the wheel speed detected by the wheel speed sensor 12 is fetched. At step 302, a high-pass filter of, for example, 2 Hz and a 20-Hz
Pre-processing by a filter composed of a low-pass filter. By inputting the wheel speed signal to the high-pass filter and performing high-pass filtering, a steady component of the acceleration of the wheel is removed, and smoothing of the wheel speed signal can be performed by performing low-pass filtering.

In the next step 304, the time series data of the band frequency is estimated from the time series data of the wheel speed preprocessed by using the online least squares method. First, the time series data of the wheel speed detected by the wheel speed sensor 12 discretized for each sample time τ is converted into the time series data of the wheel speed after the pre-processing by the filter of step 302.
[k] (k is a sample time in units of a sample time τ, k = 1, 2,...), and the following steps 1 and 2 are repeated to obtain a time series of the detected wheel speeds. The time series data of the braking force gradient is estimated from the data. Step 1:

[0100]

(Equation 16)

Note that φ [k] in the equation (24) is a value obtained by multiplying the change amount of the wheel speed in one sample time by the sample time τ (physical quantity relating to the change in the wheel speed), and (25)
Y [k] in the equation is 1 of the change amount of the wheel speed (ω [k−1] −ω [k−2], ω [k] −ω [k−1]) during one sample time.
Change in sample time (ω [k-1] -ω [k-2]-
(Ω [k] −ω [k−1])) (physical quantity relating to a change in wheel speed change). Step 2:

[0102]

[Equation 17]

From the recurrence formula, the estimated value θ, that is, the braking force gradient is estimated. Here, λ in the expressions (27) and (28) is a forgetting factor (for example, λ = 0.98) indicating the degree of removing past data, and T indicates transposition of a matrix.

Note that θ [k] in the equation (26) is a physical quantity representing the history of physical quantities relating to changes in wheel speed and the history of physical quantities relating to changes in wheel speed.

In the above description, an example in which the band frequency is estimated using the online least squares method has been described. However, the band frequency can be estimated using another online identification method such as the auxiliary variable method.

FIG. 11 shows an example of band frequency estimation results in the first-order lag model estimated as described above.
Further, as understood from the gain diagram of FIG. 11, the gain of the approximated first-order lag model has a braking force gradient of 300.
At other than Ns / m, it has been identified as a characteristic that passes the steady gain and the gain at the anti-resonance point (around 40 Hz) in the gain diagram of the wheel full model.
The resonance before and after the suspension around 40 Hz and the resonance characteristic of the tire rotational vibration around 40 Hz are neglected. When the braking force gradient is as small as 300 Ns / m, resonance does not appear because the first-order lag model does not pass through the anti-resonance point, and the vibration characteristics of the first-order lag model and the characteristics of the wheel full model are good. I can understand that we are doing. This is because in the braking region near the limit where the braking force gradient is 300 Ns / m or less, the influence of the front-rear suspension resonance and the resonance due to the tire rotational vibration is small, and the wheel deceleration motion model is dominant. Therefore, near such a limit, it is considered that the wheel motion can be approximated by the following wheel deceleration motion model.

[0107]

(Equation 18)

Where ν _w is the wheel speed (m / s), w is the road surface disturbance, k is the braking force gradient (Ns / m), R _C is the tire effective radius (m), and J is the vehicle inertia moment. ν _w
Represents the band frequency.

By the way, the above equation (29) indicates that, in the limit region, the range between the band frequency ω ₀ and the braking force gradient is

[0110]

[Equation 19]

It is shown that there is a relationship as follows.

In the low slip region, the relationship shown in FIG. 12 can be derived by applying the least squares method. This diagram shows the relationship between the braking force gradient in the wheel full model and the band frequency identified from the wheel speed data when a white disturbance is applied. The unit of the band frequency in FIG. 12 is [rad / s]. The braking force gradient monotonically increases as the band frequency increases. The relationship between the band frequency and the braking force gradient in FIG. 12 is stored as a map in the memory of the braking force gradient estimating means 14B ',
By calculating the braking force gradient corresponding to the band frequency estimated by the band frequency estimating means 14A 'based on the wheel speed signal using the map, the braking force gradient is estimated from the band frequency estimation (identification) result. Becomes possible.

In the present embodiment, the case where the relationship between the road surface μ gradient and the braking / driving force is approximated by a linear function as shown in FIG. 14 has been described. However, the present invention is not limited to this. Instead, if the road surface μ gradient monotonically decreases as the braking / driving force increases, the form may be described by a polynomial or a map of a quadratic function or higher. In this case, the same effect as in the present embodiment can be obtained.

[0114]

According to the physical quantity estimating device of the present invention, a physical quantity indicating the ease of slipping between the tire and the road surface is estimated based on the wheel speed, and the braking force or the driving force of the vehicle is estimated. Derived, a model showing the relationship between the physical quantity indicating the ease of slipping between the tire and the road surface and the braking force or driving force of the vehicle, the estimated physical quantity, and the derived braking force or driving Based on the force, the estimated physical quantity is corrected to be a physical quantity when the slip ratio is substantially zero, so that the tire when the slip ratio is substantially zero without requiring a maximum road surface friction coefficient. The effect of being able to estimate a physical quantity indicating the ease of slipping with the road surface is obtained.

Further, according to the road surface friction state estimating device of the fourth aspect, the tire friction on the road surface is determined based on the physical amount estimated by the physical amount estimating device according to the present invention, that is, the physical amount when the slip ratio is substantially zero. Since the state is estimated, an effect is obtained that the road surface friction state during braking or driving can be estimated with high accuracy.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a road friction state estimating apparatus 10 according to an embodiment.

FIG. 2 is a block diagram illustrating a configuration of a road μ gradient estimating unit according to the embodiment;

FIG. 3 is a graph showing a relationship between a frequency and an amplitude of a wheel speed signal.

FIG. 4 is a graph showing a relationship between an asphalt road and a low μ road between a vibration level of a wheel speed signal and a road μ gradient.

FIG. 5 is a graph showing a relationship between a vibration level of a wheel speed signal and a road μ gradient for each road surface state.

FIG. 6 is a diagram showing a frequency characteristic of a wheel speed signal.

FIG. 7 is a graph showing a change in a road surface μ gradient estimated value when the vehicle gets over a protrusion before traveling on a low μ road.

FIG. 8A shows a change in wheel speed over time on a high μ road and a low μ road during braking, and FIG. 8B shows a road surface μ gradient correcting means 20 on the high μ road and the low μ road at this time. (C) is a graph showing a change in the road surface μ gradient before correction by the road surface gradient correction, and (C) is a graph showing a change in the road surface μ gradient after correction by the road surface μ gradient correction means 20 at this time.

FIG. 9 is a block diagram illustrating a configuration example when a braking force gradient is estimated.

FIG. 10 is a flowchart showing an algorithm for estimating a band frequency in the configuration of FIG. 9;

FIG. 11 is a gain diagram illustrating a frequency response from a road surface disturbance to a wheel speed of the first-order lag model.

FIG. 12 is a diagram showing a relationship between a band frequency and a braking force gradient.

FIG. 13 is a graph showing an example of the relationship between the road surface μ gradient and the braking / driving force in the brush model.

FIG. 14 is a diagram for explaining the principle of the present invention, and is a graph showing an example of a model in which the road surface μ gradient monotonously decreases in accordance with the braking / driving force.

FIG. 15 is a graph showing the relationship between the slip speed and the braking / driving force when the model of FIG. 14 is used.

[Explanation of symbols]

 Reference Signs List 10 road surface friction state estimating device 12 wheel speed sensor (wheel speed detecting means) 14 road surface μ gradient estimating means (physical quantity estimating means) 16 high-pass fister 18 braking force estimating means (braking / driving force deriving means) 20 road surface μ gradient correcting means ( Physical quantity correction means) 22 Bandpass filter 24 Vibration level calculation means (Vibration level estimation means) 26 Road surface friction state estimation means

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Eiichi Ono 41-Cho, Yokomichi, Nagakute-cho, Aichi-gun, Aichi Prefecture Inside of Toyota Central Research Laboratory Co., Ltd. (72) Inventor Koji Umeno Nagakute-machi, Aichi-gun (1) Inside Toyota Toyota Central Research Institute Co., Ltd. (72) Inventor Katsuhiro Asano 41-cho Nagakute-cho, Aichi-gun, Aichi Prefecture Toyota Motor Co., Ltd. 1 Toyota Town, Toyota Prefecture (72) Inventor Satoshi Onozawa 2-1-1 Asahimachi, Kariya City, Aichi Prefecture F-term in Aisin Seiki Co., Ltd. 3D046 BB28 BB29 HH36 HH46

Claims (4)

[Claims] 1. A wheel speed detecting means for detecting a wheel speed, and a physical quantity estimating means for estimating a physical quantity indicating the ease of slip between a tire and a road surface based on the wheel speed detected by the wheel speed detecting means. A braking / driving force deriving means for deriving a braking force or a driving force of the vehicle, and a model showing a relationship between a physical quantity indicating the ease of slipping between the tire and the road surface and the braking force or the driving force of the vehicle, Based on the physical quantity estimated by the physical quantity estimating means and the braking force or the driving force derived by the braking / driving force deriving means, the physical quantity estimated by the physical quantity estimating means is calculated when the slip rate is substantially zero. A physical quantity estimating device comprising: physical quantity correcting means for correcting the physical quantity to be a physical quantity. 2. The model according to claim 1, wherein the model is a model in which a physical quantity indicating slipperiness between a tire and a road surface is monotonously decreased with an increase in a braking force or a driving force of the vehicle. Physical quantity estimation device. 3. The physical quantity is a braking force gradient or a road μ gradient when the braking force is derived by the braking / driving force deriving means, and is a driving force gradient or a road μ gradient when the driving force is derived. 3. The physical quantity estimating apparatus according to claim 1 or 2. 4. A device for estimating a physical quantity according to claim 1, further comprising: estimating a vibration level indicating a magnitude of vibration of a wheel based on a wheel speed detected by said wheel speed detecting means. A vibration level estimating means, and a road surface friction state estimating means for estimating a friction state of the tire against a road surface based on the physical quantity corrected by the physical quantity correcting means and the vibration level estimated by the vibration level estimating means. Road friction state estimation device provided.