JP2001073317A - Road surface state detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a road surface state detector capable of simply detecting the coefficient of friction of a road surface without increasing operating load of a micro-computer or costs. SOLUTION: A wheel speed sensor 2 is provided at each of tires TR of a vehicle 1. The oscillation component of a wheel speed caused by oscillation of each of the tire is extracted from a wheel speed signal of the wheel speed sensor 2 to estimate Q factor of frequency of the tire. In a specific period of time, large or small coefficient of friction of a road can be decided in accordance with frequency when the Q factor is larger than specific value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a road condition detecting device for determining a friction coefficient of a road surface based on a vibration electric signal including a vibration frequency component of a tire of a vehicle.

[0002]

2. Description of the Related Art Heretofore, as a road surface condition detecting device, there has been known a device for detecting a road surface condition by comparing an operation amount and a momentum of a vehicle. In such an apparatus, a road surface friction coefficient (hereinafter, referred to as a “glide rate”) is compared with a vehicle operation (eg, steering, braking, acceleration, etc.) and a vehicle momentum (yaw rate, front / rear G, lateral G, etc.) associated with the operation. , “Road surface μ”).

[0003]

However, in such a road surface state detecting device, since the operation of the vehicle is premised, when the vehicle is running at a constant speed, for example, friction such as asphalt road is generated. Road surface with a large coefficient (hereinafter,
Even if the vehicle enters a low friction coefficient road surface (hereinafter referred to as a “low μ road”) such as a frozen road from a “high μ road”, the road surface μ
Could not be detected.

[0004] To cope with such a problem, for example, one described in Japanese Patent Application Laid-Open No. H11-78843 is known. In the device described in the publication, a transfer function is identified based on a vibration model of a wheel resonance system, and a so-called S-μ characteristic (a change curve of the friction coefficient μ with respect to the slip rate S) indicates a change in the friction coefficient μ with respect to the slip rate S. The road surface condition is detected by estimating the gradient (hereinafter, referred to as “μ gradient”) with high accuracy.

However, in detecting such a road surface condition, it is necessary to solve a complicated equation, and the calculation load on the microcomputer must be increased. Further, in order to deal with similar problems, for example,
No. 94661 is also known. In the device described in the publication, unsprung acceleration that is the vertical acceleration of the wheel and sprung acceleration that is the vertical acceleration of the vehicle body are detected, and the transfer function between these sprung and unsprung accelerations is detected. The road surface μ is detected from the frequency characteristic by obtaining the ratio.

However, the detection of such a road surface condition is premised on the provision of both sprung and unsprung acceleration sensors, which has necessitated an increase in cost. An object of the present invention is to provide a road surface state detecting device capable of easily detecting a friction coefficient of a road surface without increasing a calculation load and cost of a microcomputer.

[0007]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the invention according to claim 1 estimates a Q factor of a tire frequency characteristic based on a vibration electric signal including a vibration frequency component of a vehicle tire. Q factor estimating means to perform
A statistical value calculating unit that calculates a statistical value of the estimated Q factor; and a road surface friction coefficient determining unit that determines whether a road surface friction coefficient is large or small based on the calculated statistical value. And

According to a second aspect of the present invention, there is provided a peak value estimating means for estimating a peak value of a frequency characteristic of a tire based on a vibration electric signal including a vibration frequency component of a tire of a vehicle, and a statistic of the estimated peak value. The gist of the present invention is to include statistical value calculating means for calculating a value and road surface friction coefficient determining means for determining whether the road surface friction coefficient is large or small based on the calculated statistical value.

The invention described in claim 3 is the first or second invention.
Wherein the statistic is at least one of a frequency, a standard deviation, a variance, and an average value.

(Operation) According to the first aspect of the present invention,
Whether the friction coefficient of the road surface is large or small is determined very simply based on the statistical value of the Q factor of the frequency characteristics of the tire without increasing the calculation load and cost of the microcomputer, and the road surface state is detected.

According to the second aspect of the present invention, without increasing the computational load and cost of the microcomputer,
Whether the friction coefficient of the road surface is large or small is determined very easily based on the statistical value of the peak value of the tire frequency characteristic, and the road surface state is detected.

According to the third aspect of the present invention, the statistical value of the Q factor or the peak value of the frequency characteristic of the tire can be calculated very easily based on at least one of the frequency, the standard deviation, the variance, and the average value.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS (First Embodiment) Hereinafter, a road surface condition detecting apparatus according to a first embodiment of the present invention will be described with reference to the drawings. In FIG. 1, each tire TR of the vehicle 1 is provided with a wheel speed sensor (represented by 2) as a wheel speed detector for detecting a rotation speed thereof. The wheel speed sensor 2 is, for example, a well-known electromagnetic induction type sensor including a toothed rotor that rotates with the rotation of each wheel and a pickup provided to face the tooth portion of the rotor. Although a device configured to output a digital signal corresponding to the rotation speed is used, another system may be used. FIG. 2 shows an example of a change in wheel speed. When a vibration component is extracted, the state shown in FIG. 3 is obtained. Each wheel speed sensor 2 is connected to an electronic control unit 3, and an output signal of the wheel speed sensor 2 is supplied to a microcomputer 20 of the electronic control unit 3. It is configured to be detected.

The microcomputer 20 has a general configuration including an input port 21, a CPU 22, a ROM 23, and a RAM.
24, an output port 25, and the like are connected to each other via a common bus, and an output signal of the wheel speed sensor 2 is input from the input port 21 and processed by the CPU 22,
The output port 25 is configured to output to the actuator 4. In addition, a wavelet function (for example, Gabor function) 11 is set for the microcomputer 20. In the microcomputer 20, the ROM 23 stores a program corresponding to the flowchart shown in FIG. 4, the CPU 22 executes the program while an ignition switch (not shown) is closed, and the RAM 24 executes a program necessary for executing the program. Temporarily store variable data.

Here, the definition of the wavelet transform as a premise of the invention of this application and the terms used in this application will be clarified. First, a function serving as a basis of the wavelet transform is called a basic wavelet function h (t), whose norm is normalized by a square integrable function, and is localized at least in the time domain. This basic wavelet function h (t) is represented by the following equation 1 called an admissible condition.
It can be expressed that the equation holds. Equation 1 indicates that the DC component (average value) of the signal is zero.

[0016]

(Equation 1) Then, as shown in the following equation 2, the basic wavelet function is scale-converted by a times, and the origin is shifted by b to set the wavelet function.

[0017]

(Equation 2) When the function to be analyzed is f (t), the wavelet function is defined as shown in the following equation (3).
In this equation, F (a, b) represents a wavelet coefficient, <> represents an inner product, and * represents a complex conjugate.

[0018]

(Equation 3) The wavelet function used for analysis is called an analyzing wavelet (basic wavelet function).
Various functions such as a Gabor function are used. For example,
Morle which is a kind of Gabor function shown in the following equation (4)
The t wavelet is known as an analyzing wavelet suitable for analyzing a signal having a singular point whose derivative is discontinuous.

[0019]

(Equation 4) In the microcomputer 20, the road surface μ (high μ road or low μ road) is determined as described later, and the determination result is output to the actuator 4. This actuator 4
Is configured to set, for example, the control value of the ABS device according to the determined road surface μ.

In the road surface condition detecting device of the present embodiment configured as described above, a series of processes relating to the detection of the road surface condition are performed by the electronic control unit 3. That is, when an ignition switch (not shown) is closed, execution of a program corresponding to the flowchart of FIG.
It is repeated at a predetermined cycle time (for example, 5 milliseconds).

As shown in FIG. 4, when the processing shifts to this routine, first in step 101, the CPU 2
2 sets the period count value j in the RAM 24 to “1”. This period count value j is a predetermined time (for example, 1
This is a count value set corresponding to a certain period Tj divided for each second), and in the present embodiment, “1” to “12”.
It counts up to "0".

CP in which period count value j is set to "1"
U22 proceeds to step 102, and stores various data in RA
After reading from M24, the process proceeds to step 103.
The CPU 22 that has proceeded to step 103 calculates the wheel speed based on the output signal of the wheel speed sensor 2. The wheel speed is used for detecting a road surface state as a vibration electric signal including a vibration frequency component of the tire TR.

After calculating the wheel speed, the CPU 22 proceeds to step 104 to calculate the filter output. In the calculation of the filter output, the oscillating electric signal output from the wheel speed sensor 2 is converted into, for example, the above-described function f (t) to be analyzed.
Is input to the microcomputer 20, and the microcomputer 20 converts the frequency scale parameter a (scale (a)) into the shift parameter b.
Wavelet transform is performed according to (time position (b)), and wavelet coefficients F (a, b) are calculated. That is, the convolution of the basic wavelet function with respect to the function f (t) is performed.

FIG. 5 shows the form of the wavelet coefficient F (a, b) obtained by the wavelet analysis. The magnitude of the wavelet coefficient F (a, b) is distinguished by a contour pattern as shown in FIG. Note that, when this is three-dimensionally displayed, it is as illustrated in FIG. 6 (in each of the figures, the scale parameter a is displayed in a logarithmic manner, but FIGS. 5 and 6 do not directly correspond to each other). Examples of the wavelet function include a Gabor function, a Mexican hat function, a French function, and a Haar function.

The filter output calculation in step 104 is performed at five predetermined frequencies near the tire resonance frequency, for example, 35 Hz, 37.5 Hz, 40 Hz, 42.5 Hz,
Implemented at 5 Hz, wavelet coefficient F (a, b) ₃₅ at frequency 35 Hz, wavelet coefficient F (a, b) _37.5 at frequency _37.5 Hz, wavelet coefficient F (a, b) ₄₀ at frequency ₄₀ Hz, frequency 42.5
The wavelet coefficient F (a, b) _42.5 at Hz and the wavelet coefficient F (a, b) _{45 at 45} Hz are obtained. The predetermined frequency at which the filter output operation is performed is not limited to five, and may be performed at least three or more predetermined frequencies. Further, the wavelet coefficient corresponding to the selected frequency may be combined in addition to the above.

After calculating the filter output, the CPU 22
The process proceeds to step 105 to perform an addition value calculation. This addition
In the arithmetic operation, the weight calculated in step 104 is calculated.
Bullet coefficient F (a, b)₃₅, F (a, b)_37.5, F
(A, b)₄₀, F (a, b) _42.5, And F (a, b)₄₅
, 200 pieces (1 second / 5 mm) in the period Tj
Second) addition value (wavelet coefficient addition value) SFj
(A, b)₃₅, SFj (a, b)_37.5, SFj (a,
b)₄₀, SFj (a, b)_42.5, And SFj (a, b)
₄₅Is calculated.

After calculating the added value, the CPU 22 proceeds to step 106 to execute a maximum value process. In this maximum value processing, these wavelet coefficient added values SFj
(A, b) ₃₅ , SFj (a, b) _37.5 , SFj (a,
b) ₄₀ , SFj (a, b) _42.5 and SFj (a, b)
Of _45, the frequency fj _i and frequency fj that is adjacent to the back and forth corresponding to the wavelet coefficients addition value becomes the maximum value
_i-1 and fj _{i + 1} and three frequencies fj _i-1 and f
wavelet coefficient addition value SF corresponding to j _i , fj _{i + 1}
j (a, b) _i-1 , SFj (a, b) _i , and SFj
(A, b) Extract _{i + 1} . For example, in the example shown in FIG. 7, since the wavelet coefficient addition value SFj (a, b) _{40 at} the frequency of 40 Hz is the maximum value, the frequencies 37.5H, 40 Hz, and 42.5 Hz are respectively changed to the frequencies fj _{i. -1} , fj _i , fj _{i + 1} and the wavelet coefficient sum SFj (a,
b) _37.5 , SFj (a, b) ₄₀ and SFj (a, b)
_42.5 is calculated by adding the wavelet coefficient sum SFj (a,
b) _i-1 , SFj (a, b) _i and SFj (a, b)
_{i + 1} .

After executing the maximum value processing, the CPU 22 proceeds to step 107 to execute the second approximation processing. This 2
In the next approximation processing, the wavelet coefficient added value SF
j (a, b) _i-1 , SFj (a, b) _i , and SFj
(A, b) Waveform approximation processing for obtaining a quadratic approximation curve passing through _{i + 1} is performed. That is, the wavelet coefficient sum SF
j (a, b) _i-1 , SFj (a, b) _i , and SFj
(A, b) Using _{i + 1} , a constant k1 in the following equation
j, k2j, and k3j are calculated by the least square method. SFj (a, b) _n = k1j × fj _n ＾ 2 + k2j × fj
_n + k3j (n = i-1, i, i + 1) These constants k1j, k2j, and k3j are calculated by the following equation (5).

[0029]

(Equation 5) The value of the inverse matrix in the above equation (5) is a constant calculated based on a preselected frequency.
It is a value stored in advance in M23.

After executing the quadratic approximation processing, the CPU 22 proceeds to step 108, where the calculated constants k1j, k1
The road surface feature amounts K1j, K2j, and K3j are calculated using 2j and k3j.
Ask for. These road surface features K1j, K2j, K3j
Is the constant k1j, k2j, an expression based on k3j, SFj (a, b) n = K1j (fj n -K2j) ^ 2 + K
3j (n = i-1, i, i + 1), which is the value when deformed, and K1j = k1j K2j = -k2j / (2k1j) K3j = k3j-k2j２2 / (4k1j).

Here, the road surface feature K3j represents the maximum value (peak) of the wavelet coefficient addition value SFj (a, b) in the period Tj. The road surface feature K2j represents the frequency at the maximum value of the wavelet coefficient addition value SFj (a, b), and corresponds to the resonance frequency. Further, the road surface feature K1j represents a feature corresponding to a Q factor indicating the sharpness of resonance of the vibration system, and the reciprocal thereof (1 / −K1j) represents a feature of a frequency width.

The CPU 22 sets the calculated road surface feature values K1j, K2j, and K3j and proceeds to step 109 to determine whether or not the period count value j is "120". Here, if it is determined that the period count value j is not “120”, the CPU 22 proceeds to step 110,
After incrementing the count value j for the same period by “1”,
Steps 102 to 109 are repeated. Such processing is performed in each case in which the period Tj becomes the period T1 to T120, respectively, where the road surface feature amounts K1j, K2j, and K3
This is a process for calculating j (j = 1 to 120).

If it is determined in step 109 that the period count value j is "120", the CPU 22 proceeds to step 111. CPU2 shifted to step 111
2 detects the road surface μ. In the detection of the road surface μ, as shown in the flowchart of FIG. 8, in step 201, the frequency counter cK1 is initialized (set to zero).
This frequency counter cK1 indicates the period (T1 to T12).
0), each of the road surface feature values K1j (j = 1 to 12)
When the magnitude (absolute value) of (0) is equal to or larger than the predetermined value K, the value is incremented by “1”. The predetermined value K is a value previously stored in the ROM 23 as a suitable value for classifying whether the road surface state is a high μ road or a low μ road, but is calculated as a learning value, for example. Value.

Now, such a road surface feature K1j (j =
The determination of the road surface μ based on the size of 1 to 120) is performed for the following reason. That is, as shown in FIG. 9A, especially at the operating point near the origin where the vehicle is traveling steadily (constant speed traveling), the S-μ characteristic shows that the friction is almost in proportion to the increase of the slip ratio S. The coefficient is increasing. As shown in FIG. 9B, the μ gradient on the high μ road has a tendency to be significantly larger than the μ gradient on the low μ road, as indicated by the μ gradient D0 in FIG. Note that the μ gradient D0 generally has a relation of D0∽∂μ / ∂S.

On the other hand, the frequency characteristic of the transfer function during steady running (constant speed running) of the vehicle based on the vibration model of the wheel resonance system described in JP-A-11-78843 is analyzed in the form shown in FIG. Have been. Therefore, according to this frequency characteristic, the μ gradient D0 at the tire resonance frequency is obtained.
Becomes larger, the Q factor, that is, the road surface feature K
It can be confirmed that the size of 1j has increased. As described above, when the road surface characteristic amount K1j corresponding to the Q factor is large, the μ gradient D0 is large, so that the road can be divided into high μ roads. On the other hand, when the road surface characteristic amount K1j is small, the μ gradient D0 is used. Is small and can therefore be divided into low μ roads.

CPU 2 that has initialized frequency counter cK1
In step 2, the process proceeds to step 202, where the count value m is set to "1".
Set to. This count value m is determined by the road surface characteristic amount K1.
This is a count value for performing the above classification for all j (j = 1 to 120).

The CPU 22 having set the count value m proceeds to step 203, and determines whether or not the road surface characteristic amount K1m is equal to or larger than a predetermined value K. Here, if the road surface characteristic amount K1m is determined to be equal to or larger than the predetermined value K, the CPU 22 proceeds to step 204.
Then, the frequency counter cK1 is incremented by "1", and the routine goes to Step 205. On the other hand, if it is determined that the road surface feature K1m is less than the predetermined value K, the CPU 22
The process directly proceeds to step 205 without incrementing the frequency counter cK1.

The CPU 22, which has proceeded to step 205,
It is determined whether or not the count value m is “120”. here,
If it is determined that the count value m is not “120”, the CP
U22 proceeds to step 206, increments the count value m by "1" and increments the count value m in steps 203 to 20 described above.
Step 5 is repeatedly executed.

On the other hand, if it is determined in step 205 that the count value m is “120”, the CPU 22 determines that the counting according to the road surface state has been completed for all the road surface characteristic amounts K1m, and proceeds to step 207.

The CPU 22, which has proceeded to step 207,
It is determined whether or not the counted frequency counter cK1 is equal to or greater than a predetermined value C. Note that the predetermined value C is a value (frequency) stored in advance in the ROM 23 as a value (frequency) having sufficient reliability for confirming that the road surface state is a high μ road. It may be a calculated value.

If the frequency counter cK1 is determined to be equal to or greater than the predetermined value C, the road surface state is determined to be a high μ road,
The process proceeds to step 208, where the road surface state setting flag XK1 is set to “1”, and then proceeds to step 112 in FIG.

On the other hand, the frequency counter cK1 has a predetermined value C
If it is determined that the road surface state is less than the road surface, the road surface state is determined to be a low μ road, and the routine proceeds to step 209, where the road surface state setting flag XK1 is set to “0”. Then, the process proceeds to step 112 of FIG.

In step 112, the CPU 22 determines whether or not the road surface state setting flag XK1 is "1". Here, when the road surface state setting flag XK1 is “1”, the road is a high μ road, and thus the CPU 22
Then, the flow goes to 13 to execute the corresponding high μ control. That is, the CPU 22 drives the actuator 4 to set, for example, the control value of the ABS device to the high μ side, and then temporarily ends the subsequent processing.

On the other hand, the road surface state setting flag XK1 is "0".
If so, the road is a low μ road, and the CPU 22 proceeds to step 114 to execute the corresponding low μ control. That is, the CPU 22 drives the actuator 4 to set, for example, the control value of the ABS device to the low μ side, and temporarily ends the subsequent processing.

As described in detail above, according to this embodiment, the following effects can be obtained. (1) In the present embodiment, the magnitude of the road surface feature K1j (j = 1 to 120) corresponding to the Q factor is compared with the predetermined value K, and the magnitude of the road surface feature K1j (j = 1 to 120) is determined. By a very simple method of calculating the frequency when the value becomes equal to or more than the predetermined value K, the road surface state (high μ road or low μ road) can be detected without increasing the calculation load and cost of the microcomputer. .

(2) In this embodiment, the statistical value of the Q factor of the tire frequency characteristic can be calculated extremely easily by frequency calculation. (3) In the present embodiment, the constant k1 in the second approximation processing
j, k2j, k3j, that is, road surface feature amounts K1j, k2
j and K3j can be easily calculated by multiplying the values of the inverse matrix stored in advance in the ROM 23 and calculated based on the frequency selected in advance.

(Second Embodiment) Hereinafter, a road surface condition detecting apparatus according to a second embodiment of the present invention will be described with reference to the drawings. In the first embodiment, the magnitude of the road surface feature amount K1j (j = 1 to 120) is equal to the predetermined value K.
Although the road surface state is detected based on the frequency of the above, in the second embodiment, the road surface characteristic amount K1j (j =
1 to 120) to detect the road surface condition based on the standard deviation. Therefore, detailed description of the same parts as those in the first embodiment will be omitted.

In the present embodiment, the CPU 22 which has proceeded to step 111, as shown in the flowchart of FIG.
Based on j (j = 1 to 120), the standard deviation σ (K1) is calculated based on the following equation (6).

[0049]

(Equation 6) Then, the CPU 22 that has calculated the standard deviation σ (K1)
The process proceeds to step 302, where it is determined whether the standard deviation σ (K1) is equal to or greater than a predetermined value σ0. Note that this predetermined value σ0
Is a value stored in advance in the ROM 23 as a suitable value for classifying whether the road surface state is a high μ road or a low μ road, but may be a value calculated as a learning value, for example. Good.

Now, such a road surface feature K1j (j =
The determination of the road surface μ based on the standard deviation σ (K1) of 1 to 120) is performed for the following reason. That is, as shown in FIG. 12, the vehicle is, for example, 30 km / h.
(Km / h), 60 km / h, and 80 km / h, in a steady running state (constant speed running), each wheel (four wheels) on a high μ road It has been experimentally confirmed that the standard deviation σ (K1) is larger than the standard deviation σ (K1) when traveling on a low μ road. Therefore, the standard deviation σ (K1) and FIG.
By comparing with the predetermined value σ0 shown in (1), the high μ road and the low μ road can be appropriately classified.

If it is determined that the standard deviation σ (K1) is equal to or larger than the predetermined value σ0, the road surface state is determined to be a high μ road, and the routine proceeds to step 303, where the road surface state setting flag XK
1 is set to “1”, and the process proceeds to step 112 in FIG.

On the other hand, the standard deviation σ (K1) is equal to a predetermined value σ
If it is determined to be less than 0, the road surface state is a low μ road,
In step 304, the road surface state setting flag XK1 is set to "0". Then, the process proceeds to step 112 of FIG.

In the step 112, the CPU 22 selectively executes the high μ control or the low μ control (steps 113 and 114) according to the value of the road surface state setting flag XK1, as in the first embodiment. It is.

As described in detail above, according to the present embodiment, effects similar to the effects (1) to (3) of the first embodiment can be obtained. Note that the embodiment of the present invention is not limited to the above-described embodiment, and may be modified as follows.

In the second embodiment, the road surface state is detected based on the standard deviation σ (K1) of the road surface characteristic amount K1j (j = 1 to 120). The road surface state is calculated by the following equation (7). Variance V of feature K1j (j = 1 to 120)
The road surface condition may be detected based on (K1).

[0056]

(Equation 7) In the second embodiment, the road surface feature K1j (j
= 1 to 120), the road condition is detected based on the standard deviation σ (K1), which is calculated based on the average value M (K1) of the road feature values K1j (j = 1 to 120) calculated by the following equation (8). The road condition may be detected based on the condition.

[0057]

(Equation 8) In addition, such a road surface characteristic amount K1j (j = 1 to 120)
The determination of the road surface μ based on the average value M (K1) is performed for the following reason. That is, as shown in FIG. 13, the vehicle is, for example, 30 km / h, 60 km / h,
In a state where the vehicle is traveling at a constant speed (constant speed traveling) at each speed of 80 km / h, the average value M (K1) of each wheel (four wheels) when traveling on a high μ road is low. It has been experimentally confirmed that the average value M (K1) when traveling on the μ road is smaller than the average value M (K1). Therefore, the average value M (K1)
However, when the value is less than the predetermined value M0 shown in FIG. 13, the road is determined to be a high μ road, and when the value is equal to or more than the predetermined value M0, the road is determined to be a low μ road. Can be.

In the second embodiment, the road surface state is detected based on the standard deviation σ (K1) of the road surface characteristic amount K1j (j = 1 to 120), which is calculated by the following equation (9). The road surface condition may be detected based on the standard deviation σ (K3) of the characteristic amount K3j (j = 1 to 120).

[0059]

(Equation 9) In addition, such a road surface characteristic amount K3j (j = 1 to 120)
The determination of the road surface μ based on the standard deviation σ (K3) is performed for the following reason. That is, as shown in FIG. 14, the vehicle is, for example, 30 km / h and 60 km / h.
h, when the vehicle is traveling at a constant speed (constant speed traveling) at each speed of 80 km / h, the standard deviation σ (K3) of each wheel (four wheels) when traveling on a high μ road is It is experimentally confirmed that the standard deviation becomes larger than the standard deviation σ (K3) when traveling on a low μ road. Therefore, the standard deviation σ
When (K3) is equal to or more than the predetermined value σ1 shown in FIG. 14, it is determined that the road is a high μ road, and when it is less than the predetermined value σ1, it is determined that the road is a low μ road. Can be divided.

Similarly, the road surface feature K3j (j
= 1 to 120), the road surface condition may be detected. In the second embodiment, the road surface feature K1j (j
= 1 to 120), the road surface state is detected. This is the average value M (K3) of the road surface feature K3j (j = 1 to 120) calculated by the following equation (10).
The road surface condition may be detected based on the condition.

[0061]

(Equation 10) In addition, such a road surface characteristic amount K3j (j = 1 to 120)
The determination of the road surface μ based on the average value M (K3) is performed for the following reason. That is, as shown in FIG. 15, the vehicle is, for example, 30 km / h, 60 km / h,
In a state in which the vehicle is traveling at a constant speed (constant speed traveling) at each speed of 80 km / h, the average value M (K3) when traveling on a high μ road is when traveling on a low μ road. It has been experimentally confirmed that the average value M (K3) of each wheel (four wheels) is larger than the average value M (K3). Therefore, the average value M (K3)
However, when the road is equal to or more than the predetermined value M1 shown in FIG. 15, it is determined that the road is a high μ road, and when the road is less than the predetermined value M1, it is determined that the road is a low μ road. Can be.

In each of the above embodiments, the road surface condition is detected based on the statistical value of the road surface feature K1j corresponding to the Q factor. Similarly, the value − (K1j / (2 × K3j)) and (-(K1
j / (2 × K3j))) ＾ 0.5 or K1j / K3j, and the road surface state may be detected based on the statistical value.

In each of the above embodiments, the output signal (wheel speed) of the wheel speed sensor 2 is used as the vibration electric signal including the vibration frequency component of the tire of the vehicle. The speed of change or acceleration of a vehicle height sensor or the like may be used.

In the above embodiments, a wavelet filter composed of an FIR (finite impulse response) filter is generally used for the wavelet operation. However, in order to reduce the load on the microcomputer, the same impulse as in the wavelet transform is used. An IIR (infinite impulse response) type filter having a response or a frequency response can be used in combination. Regarding the combination of IIR filters, for example, when a Gabor function is used as a mother function, it is divided into a real part and an imaginary part, so that ((real part) ＾ 2 + (imaginary part) ＾ 2) ＾ 0 .5, one of | real part | + | imaginary part |, max (| real part |, | imaginary part |), low-pass | real part |, or low-pass | imaginary part | The load on the microcomputer can be further reduced.

In each of the above embodiments, a wavelet filter is employed as a frequency filter. However, this may be a general-purpose band-pass filter having a predetermined pass band.

In the above embodiments, the maximum value processing or the like is performed based on the added value (wavelet coefficient added value) in the period Tj. For example, this is the average value (wavelet coefficient average value) in the period Tj. May be performed based on the maximum value processing.

Next, technical ideas other than the claims which can be grasped from the above embodiments will be described below together with their effects. (A) A vibration electric signal including a vibration frequency component of a tire of a vehicle is converted into a time by a wavelet function enlarged / reduced by a scale parameter with respect to at least three predetermined frequencies based on a basic wavelet function localized in time. A wavelet transform is performed according to the shift parameter indicating the position, a wavelet coefficient is calculated by the wavelet transform, an approximate curve is obtained by the least square method using the wavelet coefficient at each predetermined frequency, and the Q factor of the tire frequency characteristic is obtained. Factor estimating means for estimating the estimated Q factor, a Q factor statistical value calculating means for calculating a statistical value of the estimated Q factor, and a road surface for determining whether the friction coefficient of the road surface is large or small based on the calculated statistical value Road surface condition comprising friction coefficient determining means Detection device.

According to this configuration, it is extremely easy to determine whether the coefficient of friction of the road surface is large or small based on the statistical value of the Q factor of the frequency characteristics of the tire without increasing the computational load and cost of the microcomputer. , A road surface condition is detected.

(B) A vibration electric signal including a vibration frequency component of a tire of a vehicle is converted into a wavelet function enlarged / reduced by a scale parameter with respect to at least three predetermined frequencies based on a time-localized basic wavelet function. A wavelet transform is performed in accordance with a shift parameter indicating a time position, a wavelet coefficient is calculated by the wavelet transform, and an approximate curve is obtained by a least square method using a wavelet coefficient at each predetermined frequency by a least square method, thereby obtaining a tire frequency characteristic. Peak value estimating means for estimating the peak value of, peak statistical value calculating means for calculating a statistical value of the estimated peak value, and determining whether the friction coefficient of the road surface is large or small based on the calculated statistical value Road surface condition detecting device, comprising: .

According to this configuration, it is extremely easy to determine whether the coefficient of friction of the road surface is large or small based on the statistical value of the peak value of the tire frequency characteristics without increasing the computational load and cost of the microcomputer. , A road surface condition is detected.

[0071]

As described in detail above, according to the first and second aspects of the present invention, it is possible to easily detect the friction coefficient of the road surface without increasing the operation load and cost of the microcomputer.

According to the third aspect of the present invention, the statistical value of the Q factor or the peak value of the frequency characteristic of the tire can be calculated extremely easily based on at least one of the frequency, the standard deviation, the variance, and the average value. it can.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing a first embodiment of the present invention.

FIG. 2 is a graph showing an example of a change state of a wheel speed.

FIG. 3 is a graph showing an example of a vibration state of a wheel speed.

FIG. 4 is a flowchart showing the embodiment.

FIG. 5 is a graph showing an example of a mode of a wavelet coefficient.

FIG. 6 is a three-dimensional graph showing an example of a mode of a wavelet coefficient.

FIG. 7 is a graph for explaining how to obtain constants k1, k2, and k3.

FIG. 8 is a flowchart showing the embodiment.

FIG. 9 is a graph showing a relationship between a slip ratio and a friction coefficient.

FIG. 10 is a graph showing frequency characteristics during steady running (constant speed running) of the vehicle.

FIG. 11 is a flowchart showing a second embodiment of the present invention.

FIG. 12 is a graph showing a relationship between a road surface μ and a standard deviation σ (K1).

FIG. 13 is a graph showing a relationship between a road surface μ and an average value M (K1).

FIG. 14 is a graph showing a relationship between a road surface μ and a standard deviation σ (K3).

FIG. 15 is a graph showing a relationship between a road surface μ and an average value M (K3).

[Explanation of symbols]

 Reference Signs List 1 vehicle 2 wheel speed sensor 3 electronic control unit 20 microcomputer TR tire

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification code FI Theme coat ゛ (Reference) G01B 21/30 102 G01B 21/30 102 G01N 19/02 G01N 19/02 B

Claims (3)

[Claims] 1. A Q factor estimating means for estimating a Q factor of a frequency characteristic of a tire based on a vibration electric signal including a vibration frequency component of a tire of a vehicle, and a statistical value calculating for calculating a statistical value of the estimated Q factor. And a road surface friction coefficient determining unit that determines whether the road surface friction coefficient is large or small based on the calculated statistical value. 2. A peak value estimating means for estimating a peak value of a frequency characteristic of a tire based on a vibration electric signal including a vibration frequency component of a tire of a vehicle, and a statistical value calculation for calculating a statistical value of the estimated peak value. And a road surface friction coefficient determining unit that determines whether the road surface friction coefficient is large or small based on the calculated statistical value. 3. The road condition detecting device according to claim 1, wherein the statistical value is at least one of a frequency, a standard deviation, a variance, and an average value.