JP3535076B2 - Road surface friction coefficient determining apparatus and method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a road surface friction coefficient determining apparatus and method. More specifically, by determining the coefficient of friction between the road surface and the tires (road surface friction coefficient) by using the rotation information of the four tire wheels, a road surface friction coefficient determination device capable of improving the performance and safety of the vehicle. And about the method.

[0002]

2. Description of the Related Art When a vehicle is subjected to sudden acceleration or braking on a slippery road surface, there is a risk that the tires may slip and spin. In addition, sudden steering may cause the vehicle to slip or spin.

Therefore, conventionally, before the braking force between the tire and the road surface exceeds the maximum value and the tire is locked, the brake torque acting on the wheel is reduced to prevent the wheel from being locked. An anti-lock brake device and the like for controlling the rotational speed of the wheel that provides the maximum braking force have been proposed (Japanese Patent Laid-Open No. 60-99757, Japanese Patent Laid-Open No. 1-99757).
249559, etc.).

For example, in the control of the anti-lock brake device, a slip ratio is calculated from the vehicle judgment speed and the detected wheel speed (rotational speed), and then the calculated slip ratio matches a preset reference slip ratio. The braking force is controlled so that the maximum braking force is followed.

In the control of such an ABS device, the friction coefficient μ of the road surface is used. That is, depending on the road surface friction coefficient μ (road surface μ), for example, the control content is changed between high μ and low μ to perform optimum control.

[0006]

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention
In the device disclosed in Japanese Patent No. 759, the vehicle acceleration is obtained from the driven wheels when the slip occurs, and the road surface μ is determined using this acceleration.

That is, according to this publication, when the vehicle acceleration is A during slip and the vehicle weight is W, the driving force F required for accelerating the vehicle is obtained by F = W · A / g (g is Gravitational acceleration). On the other hand, the driving force F is determined by the frictional force between the driving wheel and the road surface, and F = μ · is calculated by using the load Wr applied to the driving wheel and the road surface μ.
It can be expressed as Wr. From these two formulas, the road surface μ
Is calculated as μ = W / Wr · g × A.

However, the road surface μ obtained by this equation
Is the road surface μ at the time when the acceleration of the driven wheel, which is simply obtained by differentiating the rotational speed of the driven wheel, is replaced with the vehicle acceleration A, and the vehicle acceleration A is calculated. The road surface between the actual road surface and the tire is There is an overwhelming possibility when it is unknown whether it is μ or not stochastically.

Therefore, if various vehicle motion controls such as ABS are performed based on such a road surface μ, there is a possibility of executing an inappropriate control because the control does not correspond to the actual road surface μ. Further, even when the driver is warned that the road surface is slippery, there is a possibility that a false alarm may be issued on the determined road surface μ. Therefore, normally, ABS control is performed assuming a high μ road, and the control is switched based on the rotation behavior of the tire after that to determine whether the road is really high or low μ.

Therefore, more effective control can be performed by determining the road surface condition in advance and performing ABS control according to the determined road surface μ. However, a technique for accurately determining the road surface μ in the normal running condition is available. Absent.

According to the present invention, a road surface friction coefficient determining apparatus and method capable of enhancing the performance and safety of a vehicle by determining the friction coefficient between a road surface and a tire using rotation information of four tire wheels. The purpose is to provide.

[0012]

A road surface friction coefficient determination device of the present invention comprises a rotational speed detecting means for periodically detecting the rotational speeds of four tires of a vehicle, and a measured value by the rotational speed detecting means. The first calculating means calculates the acceleration / deceleration of the vehicle and the slip ratio of the tire, the moving average processing means performs a moving average of the acceleration / deceleration of the vehicle and the slip ratio in a predetermined time, and the moving average processing means. Weighted moving average processing means for further performing moving average processing on the moving average value, and obtaining a regression line of vehicle acceleration / deceleration and slip ratio of the weighted moving average value obtained by the weighted moving average processing means
Computing means and regression line obtained by the second computing means
Linear regression coefficient indicating the slope of the
It is characterized by comprising friction coefficient determining means for determining a friction coefficient between the road surface and the tire based on comparison with a threshold value .

Further, the road surface friction coefficient determining method of the present invention comprises the steps of periodically detecting the rotational speeds of the tires of the four wheels of the vehicle, and the acceleration / deceleration of the vehicle and the slip ratio of the tires based on the measured rotational speeds. A step of calculating and a step of performing a moving average of the acceleration / deceleration and slip ratio of the vehicle in a predetermined time,
And performing further moving average processing a moving average value obtained by said process, and obtaining a regression line of the acceleration and the slip ratio of the vehicle weighted moving averages obtained by the process, the slope of the regression line And a step of comparing a first-order regression coefficient with a preset threshold value, and a step of determining a friction coefficient between a road surface and a tire based on a result of the comparison.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION A road surface friction coefficient determining apparatus and method according to the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an embodiment of a road surface friction coefficient determining apparatus of the present invention, FIG. 2 is a block diagram showing an electrical configuration of the road surface friction coefficient determining apparatus in FIG. 1, and FIG.
Is a schematic diagram showing the relationship between road surface μ and slip ratio, FIG.
5 (a), (b) and (c) are diagrams showing the frequency ratio of the regression coefficient K1 on the road surfaces of dry asphalt, compressed snow and frozen asphalt in Example 1, respectively.
6 (a), (b) and (c) are diagrams showing the frequency ratio of the regression coefficient K1 on the road surfaces of dry asphalt, snow compaction and frozen asphalt in Example 2, respectively.
7 (a), (b) and (c) are graphs showing the frequency ratio of the regression coefficient K1 on the road surfaces of dry asphalt, compressed snow and frozen asphalt in Comparative Example 1, respectively.
8 (a), (b) and (c) are diagrams respectively showing the frequency ratio of the regression coefficient K1 on the road surfaces of dry asphalt, compressed snow and frozen asphalt in Comparative Example 2;
(A), (b) and (c) are figures which respectively show the frequency ratio of regression coefficient K1 in each road surface of the dry asphalt, the pressure snow, and the frozen asphalt in the comparative example 3, and FIG. In Examples 1-3, it is a figure which shows the relationship between the response time from the start of sampling until the regression coefficient K1 is obtained, and the rate at which the value of the regression coefficient K1 was not updated when the correlation coefficient R was less than 0.6.

As shown in FIG. 1, a road surface friction coefficient determining apparatus according to an embodiment of the present invention regularly determines the rotational speeds of wheel tires provided on tires FLW, FRW, RLW and RRW of a four-wheel vehicle. Equipped with a rotational speed detecting means S for positive detection, the output of the rotational speed detecting means S is transmitted to the control unit 1 such as ABS. Further, as shown in FIG. 2, the control unit 1 is connected with an alarm indicator 2 which is a display means constituted by a liquid crystal display element, a plasma display element, a CRT or the like. Three
Is an initialization switch operated by the driver.

As the rotation speed detecting means S, a rotation speed is generated using an electromagnetic pickup or the like to generate a rotation pulse and the rotation speed is measured from the number of pulses. An angular velocity sensor or the like including one that measures the rotation speed from this voltage can be used.

As shown in FIG. 2, the control unit 1 has an I / O interface 1a necessary for exchanging signals with an external device, and a C functioning as a center of arithmetic processing.
It comprises a PU 1b, a ROM 1c in which a control operation program for the CPU 1b is stored, and a RAM 1d in which data and the like are temporarily written when the CPU 1b performs a control operation and the written data and the like are read. There is.

In the present embodiment, the control unit 1
As a first solution means, the acceleration / deceleration of the vehicle and the first slip ratio of the tire (the ratio of the wheel speed of the front tires to the wheel speed of the rear tires) are determined from the measured value by the rotation speed detecting means S. A moving average processing means for performing a moving average in a time period of, and a weighted moving average processing means for further performing a moving average processing on the moving average value obtained by the moving average processing means,
Second calculating means for obtaining the relational expression between the acceleration / deceleration of the vehicle and the second slip ratio of the weighted moving average value obtained by the weighted moving average processing means, and the slope of the relational expression obtained by the second calculating means. On the basis of this, there is provided a friction coefficient judging means for judging the friction coefficient between the road surface and the tire. The friction coefficient determining means includes comparing means for comparing the slope of the relational expression obtained by the second calculating means with a preset threshold value.

In the present embodiment, the rotational speeds of the tires of the four wheels are detected periodically every 0.2 seconds or less. The acceleration / deceleration of the vehicle can be measured by a G sensor.
It is preferable in terms of cost to calculate the average wheel speed of the wheels or the driven wheels.

Then, the acceleration / deceleration and the slip ratio of the vehicle are obtained as a moving average for each sampling time as an average value of data for a fixed time, for example, data for at least 0.1 seconds. Then, with respect to the acceleration / deceleration and the slip ratio of the moving averaged vehicle, data for a certain period of time is obtained as a moving average value for each sampling time. In this way, based on the values (moving average) of two or three times or more (a fixed number of slip ratios and acceleration / deceleration of the vehicle), the linear regression coefficient and correlation coefficient of the slip ratio with respect to the acceleration / deceleration of the vehicle Ask for. Here, when the slip ratio obtained by the weighted moving average is equal to or greater than a set value, the slip ratio may not be used for calculation and an alarm may be issued as a slip alarm.

The sampling time needs to be 0.2 seconds or less in order to reduce the variation of data and to make a discrimination in a short time. More preferably, it is 0.05 seconds or less. The average time of the first moving average is 0.2 to 2 seconds, and the average time of the second and subsequent moving averages is preferably 0.1 to 1 second.

Then, when the value of the correlation coefficient is equal to or larger than the set value, the regression coefficient is updated and held, and the value of the regression coefficient is compared with a preset threshold value to compare the tire with the road surface. Determine the coefficient of friction.

In the present embodiment, when the road surface friction coefficient determining means determines that the road surface is low μ, the alarm display 2 issues an alarm.

Hereinafter, the operation of the road surface friction coefficient determining apparatus according to the present embodiment will be described in accordance with the procedure.

Vehicle four-wheel tires FLW, FRW, RL
From the respective rotation speeds of W and RRW, the wheel speed (V
1 _n , V2 _n , V3 _n , V4 _n ) are calculated.

For example, each wheel tire FLW, FRW, RLW of the vehicle obtained from a sensor such as an ABS sensor,
The wheel speed data at a certain point of RRW is used as the wheel speeds V1 _n , V
_Let 2 _n , V3 _n , and V4 _n .

Next, the average wheel speeds (Vf _n , Vd _n ) of the driven wheels and the driving wheels are calculated.

In the case of the front wheel drive, the average wheel speeds Vf _n and Vd _n of the driven wheels and the drive wheels at a certain point are _expressed by the following equation (1),
It is obtained by (2). _{Vf n = (V3 n + V4} n) / 2 (1) Vd n = (V1 n + V2 n) / 2 (2)

Next, the average wheel acceleration / deceleration (that is, vehicle acceleration / deceleration) Af _n of the driven wheels is calculated.

If the average wheel speed Af _n-1 is obtained from the wheel speed data immediately before the average wheel speed Vf _{n of} the driven wheels, the average wheel acceleration / deceleration Af _n of the driven wheels is calculated by the following equation (3). Desired. Af _n = a · (Vf _n −Vf _n−1 ) / Δt / g (3)

Here, Δt is a time interval (sampling time) between the wheel speeds Vf _n and Vf _n-1 calculated from the wheel speed data, g is a gravitational acceleration, and a is a wheel speed (k.
It is a constant (1 / 3.6) for matching the unit of m / h) with the unit of acceleration (m / s). The sampling time needs to be 0.2 seconds or less in order to reduce the variation of data and make a determination in a short time. More preferably, it is 0.05 seconds or less.

Then, the slip ratio is calculated according to the value of the acceleration / deceleration Af _n of the vehicle.

First, when the vehicle is slipping with the drive wheels locked in the acceleration state (Vd _n = 0, Vf _n ≠ 0), or in the deceleration state, the vehicle wheels are stopped and the drive wheels cause wheel spin. (Vf _n = 0, Vd _n ≠ 0), the slip ratio S _n is calculated by the following equation (4):
Calculate from (5).

If Af _n ≧ 0 and Vd _n ≠ 0, S _n = (Vd _n −Vf _n ) / Vd _n (4) If Af _n <0 and Vf _n ≠ 0, S _n = (Vd _n− Vf _n ) / Vf _n (5) In other cases, S _n = 1.

Then, the data of the acceleration / deceleration of the vehicle and the slip ratio are subjected to moving averaging processing every sampling time.

Since the road surface μ during actual traveling is not constant and changes every moment, it is necessary to estimate the road surface μ in a short time. Further, when performing linear regression, the reliability of the obtained regression coefficient is poor unless there is a certain number of data. Therefore, the data is sampled at a sampling time of a single time, for example, every several tens of ms, and the data having a large variation obtained in this sampling time is subjected to a moving average to reduce the variation of the data without reducing the number of the data. be able to.

Regarding the acceleration / deceleration of the vehicle, MAf _n = (Af ₁ + Af ₂ + ... + Af _n ) / N (6) MAf _{n + 1} = (Af ₂ + Af ₃ + ... + Af _{n + 1} ) / N (7) MAf _{n + 2} = (Af ₃ + Af ₄ + ... + Af _{n + 2} ) / N (8) Regarding the slip ratio, MS _n = (S ₁ + S ₂ + ... + S _n ) / N (9) MS _{n + 1} = (S ₂ + S ₃ + ... + S _{n + 1} ) / N (10) MS _{n + 2} = (S ₃ + S ₄ + ... + S _{n + 2} ) / N (11 )

In general, when performing regression calculation,
If the variation of the regressed data is large, the correlation coefficient becomes low, and the accuracy of the regression coefficient becomes poor.

Therefore, when the correlation coefficient is lower than a certain level, the calculation result (regression coefficient) is not updated and the previous calculation result is succeeded. As a result, if the correlation coefficient is low and the calculation result is not updated frequently, the road surface μ cannot be estimated so much. Therefore, it is important to suppress the variation of data (to eliminate noise) and reduce the ratio of low correlation coefficient. for that reason,
Moving average processing is being performed.

However, although there are various causes for the variation of the data, the major factor is the variation due to the rotation of the tire. This appears at the same frequency as the tire rotation cycle. For example, the radius of a typical tire of an ordinary passenger car is around 300 mm, so the circumference of the tire is about 1.9 m, and when traveling at 30 km / h, it is about 4.4 H.
A frequency variation of z occurs. That is, since it changes depending on the traveling speed, there is a possibility that the moving average processing may not be able to sufficiently suppress the fluctuation.

For example, when the moving average processing is performed for 0.5 seconds (2 Hz), frequency components that are positive multiples of 2, such as 2, 4, 6, 8 Hz, can be eliminated, but other than that. The frequency component cannot be erased. Time T of moving average
1 second (1 Hz), 2 seconds (0.5 Hz), 3 seconds (0.3
3Hz) makes it possible to eliminate many frequency components, but the longer it becomes, the worse the response becomes, and the necessary frequency fluctuations (accelerator ON, OF
F) can also be erased.

Therefore, in the present embodiment, by performing the weighted moving average, variation is suppressed without increasing the response time (the time required to judge the friction coefficient μ), and the ratio of the low correlation coefficient is reduced. Lower. Next, moving average processing is further performed.

[0044]

[Equation 1]

A moving averaging process is similarly performed for the slip ratio.

[0046]

[Equation 2]

Next, the first-order regression coefficients of the acceleration / deceleration of the vehicle and the slip ratio, that is, the regression coefficient K1 for the acceleration / deceleration of the vehicle of the slip ratio and the regression coefficient K2 for the slip ratio of the acceleration / deceleration of the vehicle, are respectively given by the following equations. (1
8), (19).

[0048]

[Equation 3]

P is the number of data.

[0050]

[Table 1]

The correlation coefficient R is       R = K1 × K2 (20) Becomes

This correlation coefficient R is a set value, for example, 0.
If it is 6 or more, the value of the regression coefficient K1 is updated.

Here, the relationship between the acceleration / deceleration of the vehicle and the slip ratio is the same as the μ-s curve of a general tire and the road surface, and the difference of the road surface μ (dry asphalt road R1
3 of the high μ road, the medium μ road of the compressed snow road R2, and the low μ road of the frozen asphalt road R3). And
The regression coefficients K1 and K2 are obtained by finding the slope of the μ-s curve. This μ-s curve is originally a curve,
It is almost a straight line in the range of the slip ratio that occurs during actual traveling. That is, the μ-s curve is y = aX + b
Can be expressed by the equation Coefficient a at this time
Is a regression coefficient (K1, K2), which means the slope of a straight line. Here, a = K1 or a = K2 depending on whether y is a slip ratio or acceleration. In the present embodiment, the road surface μ
Is being determined. Of course, the road surface μ can also be determined from the regression coefficient K2.

The reason for obtaining the correlation coefficient R is to judge whether or not the value of the obtained regression coefficient is appropriate. That is, when the value of the correlation coefficient R is large, there is a correlation between the slip ratio and the acceleration, and the obtained regression coefficient is appropriate, but when the value of the correlation coefficient R is small, the correlation coefficient R is large. Since there is no correlation and the obtained regression coefficient is inappropriate, the road surface μ should not be judged by its value. Then, the road surface μ is estimated from the value of the regression coefficient K1.

For example, K1 ≦ 0.1 high μ road (μ = 0.7 or more) 0.1 <K1 ≦ 0.16 medium μ road (μ = 0.3 to 0.7) 0.16 <K1 low μ path (μ = 0.3 or less) The threshold value of this regression coefficient K1 can be obtained from, for example, experimental data up to now. Next, the driver is warned of road surface information (such as slipperiness).

Furthermore, the condition of the road surface can be determined by ABS device or TR.
It is used to control the C device.

Next, the present invention will be explained based on examples, but the present invention is not limited to such examples.

[0058]

Example 1 First, a front-wheel drive vehicle was prepared, and tires were studless tires (Glass Pick D manufactured by Sumitomo Rubber Industries, Ltd.).
S-1) was attached. Then, the vehicle was driven on a dry asphalt road R1, a snow-packed road R2 and a frozen asphalt road R3. The running condition at this time is 20 to 8 on each road surface.
It is running around 0 km / h. Regarding the sampling time of the wheel speed of the wheel, in order to eliminate the variation and the measurement error with a large number of data, for example, 1 second is too long, so the sampling time is set to 40 ms.

Then, based on the wheel speed pulse output from the rotation speed detecting means, the wheel speed is fetched and 40 ms
The average wheel acceleration / deceleration (vehicle acceleration / deceleration) of each driven wheel and the slip ratios of the front and rear wheels were calculated. Table 2 shows only the average wheel acceleration / deceleration of the driven wheels.

[0060]

[Table 2]

Next, about the average wheel acceleration / deceleration of the driven wheels and the slip ratio of the front and rear wheels, 25 pieces of data for 1 second were averaged, and a moving average value was obtained for each sampling time (40 ms) (see Table 2). .

Further, the moving-averaged average wheel acceleration / deceleration of the driven wheels and the slip ratio of the front and rear wheels are 0.5.
Data for 2 seconds (13 pieces) was averaged, and a weighted moving average value was obtained for each sampling time (40 ms) (see Table 2).

Next, 50 pieces of data of the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels obtained by weighted moving average for each sampling time are accumulated, and 1 of the average wheel acceleration / deceleration of the driven wheels with respect to the slip ratio is stored. The following regression coefficient K1 was obtained (see Table 2). The regression coefficient K1 is calculated for each sampling time by replacing the newest data of the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels, which are weighted moving averaged at each sampling time, with the oldest data. However, when the weighted moving average slip ratio was 0.5 or more or −0.5 or less, it was not added to the accumulation of 50 and a slip alarm was issued at that time.

Further, the correlation coefficient R at this time was obtained, and if the correlation coefficient R was 0.6 or more, the value of the regression coefficient K1 was updated and held. When the correlation coefficient R is less than 0.6,
The value of the regression coefficient K1 is not updated. In this case, the shortest time (response time) from the start of sampling until the regression coefficient K1 is obtained is, as shown in Table 2, the processing start time of the first moving average (No. 25), 1 second, and the second moving. Add the processing time of 0.52 seconds for the average (No. 37) and the processing time of up to 50 for storing the number of data for regression (No. 37 to 86) for 2 seconds, and overlap the data (No. 25 and No. 25). No. 37) time 0.08 (2 × 0.
04) Subtracting seconds gives 3.44 seconds.

The value of the regression coefficient K1 is compared with a preset threshold value (0.12 and 0.18) to determine the friction coefficient μ between the tire and the road surface. That is, the regression coefficient K1
And the correlation coefficient R is calculated by sampling time (40 m
The road surface μ was determined by the following threshold value.

[0066]       K1 ≦ 0.12 High μ road (No alarm or warning is issued)       0.12 <K1 ≦ 0.18 Medium μ road (Slip warning is issued)       0.18 <K1 Low μ road (slip alarm is issued)

Example 2 In Example 2, weighted moving averaging was performed twice for the same data arithmetic processing as in Example 1. That is, with respect to the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels, 0.76 seconds (19 pieces) of data were averaged, and a moving average value was obtained every sampling time (40 ms).

Then, the average wheel acceleration / deceleration of the driven wheel and the slip ratio of the front and rear wheels, which were moving averaged, were 0.36.
The data for seconds (9 pieces) was averaged, and the weighted moving average value was obtained for each sampling time (40 ms).

Further, with respect to the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels, which are subjected to the weighted moving average,
The data for 0.16 seconds (4 pieces) were averaged, and the weighted moving average value was calculated for each sampling time (40 ms).

Subsequent procedures were the same as those in Example 1 above. The shortest time (response time) from the start of sampling until the regression coefficient K1 is obtained is 3.16 seconds (= 0.76 seconds + 0.36 seconds + 0.16 seconds + 2 seconds-0.
12 seconds).

Then, the coefficient of friction μ between the tire and the road surface was determined from the value of the regression coefficient K1 thus obtained, as in Example 1.

Comparative Example 1 Based on the same data as in Example 1, for the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels, the data for 0.8 seconds (20 pieces) was used without performing the weighted moving average. Were averaged, and a moving average value was calculated only once every sampling time (40 ms).

The subsequent procedure was the same as in Example 1 above. The shortest time (response time) from the start of sampling until the regression coefficient K1 is obtained is 2.76 seconds (0.8 seconds + 2 seconds−0.04 seconds).

Then, according to the value of the obtained regression coefficient,
In the same manner as in Example 1, the friction coefficient μ between the tire and the road surface was determined.

Comparative Example 2 Based on the same data as in Example 1, for the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels, data for 1.2 seconds (30 pieces) was obtained without performing the weighted moving average. Were averaged and determined as a moving average value only once for each sampling time (40 ms).

Subsequent procedures were the same as those in Example 1 above. The shortest time (response time) from the start of sampling until the regression coefficient K1 is obtained is 3.16 seconds (1.2 seconds + 2 seconds−0.04 seconds).

According to the value of the obtained regression coefficient, Example 1
The friction coefficient μ between the tire and the road surface was determined in the same manner as in.

Comparative Example 3 With respect to the same data as in Example 1, with respect to the average wheel acceleration / deceleration of the driven wheels and the slip ratios of the front and rear wheels, the data for 1.6 seconds (40 pieces) was averaged without performing the weighted moving average. And the moving average value was calculated only once every sampling time (40 ms).

Subsequent procedures were the same as those in Example 1 above. The shortest time (response time) from the start of sampling until the regression coefficient K1 is obtained is 3.56 seconds (1.6 seconds + 2 seconds−0.04 seconds).

According to the value of the regression coefficient obtained,
The friction coefficient μ between the tire and the road surface was determined in the same manner as in.

The frequency ratios of the regression coefficient K1 on the road surfaces of dry asphalt, compressed snow and frozen asphalt in Examples 1 and 2 and Comparative Examples 1 to 3 are shown in FIGS.
In (a), (b) and (c). Figures 4-8
As a result of judging the road surface μ by the following threshold values, the high μ road, the medium μ road and the low μ road are A, B and C, respectively.
Is displayed in.

[0082]       K1 ≦ 0.12 High μ road A (No alarm or warning is issued)       0.12 <K1 ≦ 0.18 Medium μ Road B (Slip warning is issued)       0.18 <K1 Low μ road C (Slip alarm is issued)

In all of Examples 1 and 2 and Comparative Examples 1 to 3, as shown in FIGS. 4 (a) to 8 (a),
In the dry asphalt road, 97% or more of the regression coefficient K1 which is the judgment value is 0.12 or less, and it can be judged as the high μ road A, while as shown in FIGS. 4 (b) to 8 (b). In the snowy road, around 50% of the regression coefficient K1 is 0.12.
It was judged to be a medium μ road B and a low μ road C exceeding the above, and a warning or a warning was frequently issued.

As shown in FIGS. 4 (c) to 8 (c), 80% of the regression coefficient K1 is obtained on the frozen asphalt road.
It is judged to be a medium μ road B and a low μ road C that exceed 0.12 in the front and back, and further, a low μ road C that exceeds 0.18 in half of them.
, And a warning or warning was issued almost all the time.

As described above, the friction coefficient μ of the road surface can be sufficiently determined without performing the weighted moving average based on the above results alone.

On the other hand, in Table 2 and FIG. 9, in Examples 1 and 2 and Comparative Examples 1 to 3, the shortest time from the start of sampling until the regression coefficient K1 is obtained, that is, the response time and the correlation coefficient R are 0.6. It shows the relationship of the proportions where the value of the regression coefficient K1 is not updated when the value is less than. The correlation coefficient R <
The ratio of 0.6 is the average of all road surfaces of dry asphalt, snow and frozen asphalt.

[0087]

[Table 3]

Since the friction coefficient μ of the road surface is not uniform,
It is considered that the shorter the response time, the better, and the smaller the rate at which the regression coefficient K1 is not updated, the better. Therefore, in comparison with Comparative Examples 1 to 3, Examples 1 and 2 in which the weighted moving average is performed have a smaller rate of not updating the regression coefficient K1 with respect to the response time, which is more excellent.

In the present embodiment, the number of pieces of data for obtaining the regression line is set to 50, but the present invention is not limited to this. However, if the number of data is too small, the value of the regression coefficient K1 will fluctuate, or the correlation coefficient R will be 0.6.
If the number of data is less than 5, the value of the regression coefficient K1 may not be updated easily, so at least 5 are required. Conversely, if the number of data is too large, not only the response time becomes long, but also When the condition of the road surface changes one after another, the variation of the data becomes large and conversely the low ratio of the correlation coefficient R increases, resulting in more cases where the road surface μ cannot be discriminated. It is desirable to decide in.

As described above, by using this system, the road surface μ can be accurately determined in a short time, and the driver can be informed that the vehicle is in a slippery and dangerous state.

Then, the information of the determined road surface μ is set to A
By using it for BS equipment and TRC equipment, etc.
Optimal control can be performed according to Also, low μ
When the road is determined, the driver can be warned that the road surface is slippery.

[0092]

As described above, according to the present invention,
By performing the weighted moving average of the acceleration / deceleration of the vehicle and the slip ratios of the front and rear wheels, the variation in the data is suppressed, so that the rate at which the determination result is not updated can be reduced without lengthening the response time. As a result, the road surface μ can be accurately determined in a short time, and the performance and safety of the vehicle can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a road surface friction coefficient determination device of the present invention.

FIG. 2 is a block diagram showing an electrical configuration of the road surface friction coefficient determination device in FIG.

FIG. 3 is a schematic diagram showing a relationship between a road surface μ and a slip ratio.

4 (a), (b) and (c) are diagrams showing the frequency ratio of the regression coefficient K1 on each road surface of dry asphalt, compressed snow and frozen asphalt in Example 1, respectively.

5 (a), (b) and (c) are graphs showing the frequency ratio of the regression coefficient K1 on each road surface of dry asphalt, compressed snow and frozen asphalt in Example 2.

6 (a), (b) and (c) are graphs showing the frequency ratio of the regression coefficient K1 on each road surface of dry asphalt, compressed snow and frozen asphalt in Comparative Example 1.

7 (a), (b) and (c) are graphs showing the frequency ratio of the regression coefficient K1 on each road surface of dry asphalt, compressed snow and frozen asphalt in Comparative Example 2.

8 (a), (b) and (c) are diagrams showing a frequency ratio of a regression coefficient K1 on each road surface of dry asphalt, compressed snow and frozen asphalt in Comparative Example 3, respectively.

9A and 9B show that in Examples 1 and 2 and Comparative Examples 1 to 3, the response time from the start of sampling until the regression coefficient K1 is obtained and the correlation coefficient R is less than 0.6, and the value of the regression coefficient K1 is not updated. It is a figure which shows the relationship of the ratio.

[Explanation of symbols]

1 control unit 2 alarm indicators 3 initialization switch FLW, FRW, RLW, RRW tires S rotation speed detection means

─────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-7-112659 (JP, A) JP-A-7-192086 (JP, A) Fukuihei 1-78911 (JP, U) (58) Field (Int.Cl. ⁷ , DB name) B60T 8/58 G01N 19/02 B62D 6/00

Claims (3)

(57) [Claims] 1. A rotational speed detecting means for periodically detecting rotational speeds of tires of four wheels of a vehicle, and a vehicle acceleration / deceleration and a tire slip ratio are calculated from measured values by the rotational speed detecting means. Calculating means, moving average processing means for performing moving average of acceleration / deceleration and slip ratio of the vehicle in a predetermined time, and weighted moving average processing for further performing moving average processing of the moving average value obtained by the moving average processing means. Means and
A second computing means for obtaining a regression line of the acceleration and the slip ratio of the vehicle weighted moving averages obtained by the pressurized heavy moving average processing means, the regression straight obtained by the second computing means
A linear regression coefficient indicating the slope of the line and a preset predetermined
A road surface friction coefficient judging device comprising: a friction coefficient judging means for judging a friction coefficient between a road surface and a tire based on comparison with a threshold value . 2. A step of periodically detecting rotation speeds of tires of four wheels of a vehicle, a step of calculating acceleration / deceleration of the vehicle and a slip ratio of the tires from the measured rotation speeds, and an acceleration / deceleration of the vehicle. A step of performing a moving average of the speed and the slip ratio at a predetermined time; a step of further performing a moving average process on the moving average value obtained by the step; and an acceleration / deceleration of the vehicle of the weighted moving average value obtained by the step. Regression with slip ratio
A step of determining a straight line, primary times indicating the slope of the regression line
A road surface friction coefficient determination method comprising: a step of comparing a natural coefficient with a preset threshold value; and a step of determining a friction coefficient between a road surface and a tire based on a result of the comparison. 3. The data of the acceleration / deceleration of the vehicle and the slip ratio of the weighted moving average value are used to determine the linear regression coefficient of the acceleration / deceleration of the vehicle and the slip ratio, and the slip ratio of the vehicle.
The regression coefficient for both accelerations and decelerations and the acceleration / deceleration threshold of the vehicle
The correlation coefficient , which is the product of the regression coefficient and the up ratio, is obtained, and when the value of the correlation coefficient is equal to or greater than the set value, the regression coefficient is updated and held, and the value of the regression coefficient and the preset value are set. The road surface friction coefficient determination method according to claim 2, wherein the friction coefficient between the tire and the road surface is determined by comparing with a threshold value.