JP2002145038A - Road surface friction coefficient estimating device of vehicle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To ensure reliability and precision stableness of coefficient of friction on a road μ by stopping surely estimation of the coefficient μ, when the coefficient μcan not be estimated with high precision due to particularly large influence of cant running. SOLUTION: A high μ road reference value estimating unit 11 calculates a yaw rate and a transverse acceleration of a high μroad reference with a high μ road vehicle kinetic model, a low μroad reference value estimating unit 12 calculates a yaw rate and a transverse acceleration of a low μ road reference with a low μroad vehicle kinetic model, and an actual value estimating unit 13 calculates an actual yaw rate and an actual acceleration by an observer. A yaw rate comparison coefficient μestimating unit 14 compares the actual yaw rate with the high μ road reference yaw rate and the low μ road reference yaw rate to estimate a coefficient μ. A first execution determining unit 15a stops estimation of the coefficient μ, when a transverse acceleration sensor value is not between the high μ road reference transverse acceleration and the low μ road reference transverse acceleration. A second execution determining unit 15b forms Lisajous' figure with the transverse acceleration sensor value and the actual transverse acceleration, and stops estimation of the coefficient μ, when the area of the figure is small and an inclination is in a set area.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle friction coefficient estimating apparatus for accurately estimating a road friction coefficient based on a lateral acceleration and a yaw rate.

[0002]

2. Description of the Related Art In recent years, various control techniques for traction control, braking force control, torque distribution control and the like have been proposed and put to practical use in vehicles. In many of these techniques, the road surface μ is used for calculation or correction of necessary control parameters, and it is necessary to accurately estimate the road surface μ in order to execute the control reliably.

For estimating the road surface μ, various techniques for estimating the road surface μ based on parameters such as a lateral acceleration and a yaw rate generated by a turning motion of a vehicle have been proposed. For example, the present applicant also discloses in Japanese Patent Application No. 11-217508 a technique for estimating a road surface μ by comparing a yaw rate estimated by an observer with a yaw rate on a high μ road and a low μ road based on a vehicle motion model. is suggesting.

[0004]

However, when the vehicle is traveling on a cant, especially on an actual road surface, the lateral acceleration at the time of cornering has a component which is vertically downward with respect to the local coordinate system of the vehicle. The lateral acceleration is not a value that accurately represents the vehicle behavior. Therefore, not only the lateral acceleration but also the vehicle slip angle, front and rear wheel slip angle, etc. calculated from these values are not values that represent the actual vehicle behavior, and as a result, the estimated road surface μ estimation value is a value with low reliability. When various kinds of vehicle behavior control are performed using the road surface μ, optimal control may not be performed. To solve this problem, it is conceivable to perform road surface μ estimation by combining a method of determining cant from the relationship between the yaw rate, the lateral acceleration, and the vehicle speed. However, this method of cant determination cannot accurately determine when traveling on a low μ road. There's a problem.

The present invention has been made in view of the above circumstances, and particularly when the influence of cant driving is strong and the road surface μ cannot be accurately estimated, the estimation of the road surface μ is reliably stopped, and the reliability is low and the error is large. It is an object of the present invention to provide an apparatus for estimating a road surface friction coefficient of a vehicle, which can reliably prevent a road μ estimated value from being output.

[0006]

According to a first aspect of the present invention, there is provided a vehicle road surface friction coefficient estimating apparatus for estimating a road surface μ based on at least a parameter indicating a turning behavior of the vehicle. In the road surface friction coefficient estimating apparatus, a high μ road reference value related to or the same parameter as the above parameter is calculated based on a high μ road vehicle motion model, and the above parameter is related or the same parameter based on a low μ road vehicle motion model. The low μ road reference value is calculated, and the actually detected value of the parameter related to or the same parameter as the parameter is the high μ road reference value of the parameter related or the same parameter and the low value of the parameter related or the same parameter. first determining means for stopping the estimation of the road surface μ when the vehicle motion model is not in an area between the vehicle motion model and the vehicle motion model Calculating the actual value of the parameter related or the same parameter based on the parameter, and forming a Lissajous figure with the actually detected value of the parameter related or the same parameter and the actual value of the parameter related or the same parameter; And / or a second determining means for stopping the estimation of the road surface μ when the area of the Lissajous figure is small and the inclination is in a preset area.

More specifically, the apparatus for estimating a road surface friction coefficient of a vehicle according to the present invention estimates a road surface μ based on at least a parameter indicating a turning behavior of the vehicle. If at least one of the means stops the estimation of the road surface μ, the estimation of the road surface μ is stopped. For this reason, especially when the influence of cant driving is strong and the road surface μ cannot be accurately estimated, the estimation of the road surface μ is reliably stopped,
It is reliably prevented that a road μ estimated value having low reliability and a large error is output.

According to a second aspect of the present invention, there is provided an apparatus for estimating a road friction coefficient of a vehicle, comprising: an actual value estimating means for calculating at least an actual value of a parameter indicating a turning behavior of the vehicle by an observer; High μ road reference value estimating means for calculating the reference value of the parameter on the road, low μ road reference value estimating means for calculating the reference value of the parameter on the low μ road by a vehicle motion model, and a road friction coefficient estimating device including a μ road reference value and a road μ estimation means for estimating a current road surface μ by comparing with the low μ road reference value, wherein the vehicle for estimating the high μ road reference value A high μ road reference value related to or the same parameter as the above parameter is calculated by the motion model, and the parameter is calculated by the vehicle motion model for estimating the low μ road reference value. The low μ road reference value of the same parameter or the same parameter is calculated, and the actually detected value of the same parameter or the same parameter is related to the above parameter or the high μ road reference value of the same parameter and the same parameter. Or first determining means for stopping the estimation of the road surface μ when it is not in the area between the same parameter and the low μ road reference value, and the observer calculates the actual value of the parameter related or the same parameter by the observer, When a Lissajous figure is formed by the actually detected value of the parameter related to or the same parameter and the actual value of the parameter related to or the same parameter, and the area of the Lissajous figure is small and the inclination is in a preset area. To stop the estimation of road surface μ
And at least one of the determination means.

According to a second aspect of the present invention, there is provided an apparatus for estimating a road friction coefficient of a vehicle, wherein an actual value estimating means calculates at least an actual value of a parameter indicating a turning behavior of the vehicle by an observer to estimate a high μ road reference value. Means for calculating a reference value of the above parameter on a high μ road by a vehicle motion model, and a low μ road reference value estimating means for calculating a reference value of the above parameter on a low μ road by a vehicle motion model.
The estimation unit compares the actual value with the high μ road reference value and the low μ road reference value to estimate the current road surface μ. Here, when at least one of the first judging means and the second judging means stops the estimation of the road surface μ, the estimation of the road surface μ by the road surface μ estimating unit is stopped. For this reason, the influence of cant driving is particularly strong,
When the road surface μ cannot be estimated with high accuracy, the estimation of the road surface μ is reliably stopped, and the output of the road surface μ estimated value having low reliability and a large error is reliably prevented. Further, the three vehicle motion models necessary for estimating the road surface μ can be shared and used in the execution determination of the road surface μ estimation, so that the efficiency is high.

[0010]

Embodiments of the present invention will be described below with reference to the drawings. 1 to 10 show an embodiment of the present invention, FIG. 1 is a functional block diagram showing a configuration of a road surface friction coefficient estimating apparatus, FIG. 2 is a functional block diagram of a road surface μ estimation execution determining unit, and FIG. FIG. 4 is an explanatory diagram showing an equivalent two-wheeled vehicle model of a car, FIG. 4 is an explanatory diagram showing the structure of an observer, and FIG. 5 is an explanatory diagram of a high μ road reference value, a low μ road reference value, and a sensor value of lateral acceleration. 6 is an explanatory diagram of actual values and sensor values of lateral acceleration and yaw rate, FIG. 7 is an explanatory diagram of a Lissajous figure drawn by two waveforms, FIG. 8 is an explanatory diagram of calculation of Lissajous figure area, and FIG. FIG. 10 is an explanatory diagram of a Lissajous figure formed for the lateral acceleration and yaw rate of FIG. 10, and FIG. 10 is a flowchart of road surface μ estimation execution determination control.

In FIG. 1, reference numeral 1 is mounted on a vehicle,
A road surface friction coefficient estimating device for estimating a road surface friction coefficient is shown,
A front wheel steering angle sensor 3, a vehicle speed sensor 4, a lateral acceleration sensor 5, and a yaw rate sensor 6 are connected to the control unit 2 of the road surface friction coefficient estimating device 1, and the front wheel steering angle δfs,
Vehicle speed Vs, lateral acceleration ^{(d 2 y / dt 2)} s, yaw rate so that the (yaw angular velocity) (dψ / dt) s is input. The suffix s of each parameter is used to distinguish that the parameter is a sensor value.

The control unit 2 mainly includes a high μ road reference value estimation unit 11, a low μ road reference value estimation unit 12, an actual value estimation unit 13, a yaw rate comparison road surface μ estimation unit 14, and a road surface μ estimation execution determination unit 15. It estimates and outputs the road surface friction coefficient μ.

The high μ road reference value estimating section 11 serves as high μ road reference value estimating means, and includes a vehicle speed Vs and a front wheel steering angle δfs.
Is input, and the yaw rate corresponding to the detected vehicle speed Vs and the front wheel steering angle δfs is changed to a high μ road reference yaw rate (dψ / dt) H by a vehicle motion model based on a predetermined vehicle motion equation on a high μ road. And outputs it to the yaw rate comparison road surface μ estimation unit 14. Further, the high μ road reference value estimating unit 11 outputs the high μ road reference yaw rate (dψ / d
t) H, the yaw angular acceleration (d ² ψ / dt ² ) H is output to the yaw rate comparison road surface μ estimating section 14 and the high μ
The road-based lateral acceleration (d ² y / dt ² ) H is calculated, and the yaw rate comparison road surface μ estimation unit 14 and the road surface μ estimation execution determination unit 1 are calculated.
5 is output. The subscript H of each parameter output from the high μ road reference value estimating unit 11 indicates that the parameter is a high μ road reference parameter.

The low μ road reference value estimating section 12 serves as low μ road reference value estimating means, and includes a vehicle speed Vs and a front wheel steering angle δfs.
Is input, and the yaw rate corresponding to the detected vehicle speed Vs and the front wheel steering angle δfs is changed to a low μ road reference yaw rate (dψ / dt) L by a vehicle motion model based on a predetermined vehicle motion equation on a low μ road. And outputs it to the yaw rate comparison road surface μ estimation unit 14. Further, the low μ road reference value estimating unit 12 outputs the low μ road reference yaw rate (dψ / d
t) L, the yaw angular acceleration (d ² ψ / dt ² ) L is output to the yaw rate comparison road surface μ estimating unit 14 and the low μ
The road-based lateral acceleration (d ² y / dt ² ) L is calculated and output to the road surface μ estimation execution determination unit 15. The subscript L of each parameter output from the low μ road reference value estimation unit 12 is
Indicates a road-based parameter.

Here, the high μ road reference value estimating unit 11 and the low μ
The vehicle motion model used in the road reference value estimating unit 12 and the calculation of each parameter will be described with reference to FIG. That is, regarding the vehicle motion model of FIG. 3, the equation of motion relating to the translational motion in the vehicle lateral direction is as follows: the cornering forces (one wheel) of the front and rear wheels are Ff and Fr, the body mass is M, and the lateral acceleration is (d ² y / dt ² ) M · (d ² y / dt ² ) = 2 · Ff + 2 · Fr (1)

On the other hand, the equation of motion relating to the rotational motion about the center of gravity is represented by the distances from the center of gravity to the front and rear wheel axes Lf, L
r, the yaw moment of inertia of the vehicle body is Iz, and the yaw angular acceleration is (d ² ψ / dt ² ), and Iz · (d ² ψ / dt ² ) = 2 · Ff · Lf−2 · Fr · Lr (2) Indicated by

If the vehicle slip angle is β and the vehicle slip angular velocity (dβ / dt), the lateral acceleration (d ² y / d)
t ² ) is represented by (d ² y / dt ² ) = V · ((dβ / dt) + (dψ / dt)) (3)

The cornering force has a response close to a first-order delay with respect to the side slip angle of the tire. However, ignoring this response delay, and further linearizing the suspension characteristics using an equivalent cornering power incorporating the tire characteristics. It is as follows. Ff = −Kf · βf (4) Fr = −Kr · βr (5) Here, Kf and Kr are equivalent cornering powers of the front and rear wheels, and βf and βr are side slip angles of the front and rear wheels.

Assuming that the effects of the roll and the suspension are taken into consideration in the equivalent cornering powers Kf and Kr,
Using the equivalent cornering powers Kf and Kr, the front and rear wheelslip angles βf and βr can be simplified as follows by setting the front wheel steering angle to δf. βf = β + Lf · (dψ / dt) / V−δf (6) βr = β−Lr · (dψ / dt) / V (7)

When the above equations of motion are put together, the following equation of state is obtained. (Dx (t) / dt) = A · x (t) + B · u (t) ... (8) x (t) = [β (dψ / dt)] T u (t) = [δf 0] T a11 = −2 · (Kf + Kr) / (M · V) a12 = −1-2 · (Lf · Kf−Lr · Kr) / (M · V ² ) a21 = −2 · (Lf · Kf−Lr · Kr) ) / Iz a22 = −2 · (Lf ² · Kf + Lr ² · Kr) / (Iz · V) b11 = 2 · Kf / (MV) b21 = 2 · Lf · Kf / Iz b12 = b22 = 0

In the high μ road reference value estimating unit 11, for example, equivalent cornering powers Kf and Kr when the road surface μ is 1.0 are set in advance in the equation (8), and the vehicle motion state (vehicle speed Vs, (Dx) at the front wheel steering angle δfs
(t) / dt) = [(dβ / dt) (d ² ψ / d
t ² )] By calculating ^T , the vehicle slip angular velocity (dβ / dt) H and the yaw angular acceleration (d ² ψ / dt) based on a high μ road are calculated.
² ) Calculate H. Then, the calculated vehicle slip angular velocity (dβ / dt) H and yaw angular acceleration (d ² ψ) based on the high μ road are calculated.
By integrating / dt ² ) H, the vehicle slip angle βH and the yaw rate (dｄ / dt) H based on the high μ road are obtained. Further, by substituting the vehicle slip angle βH based on the high μ road and the yaw rate (dｄ / dt) H into the equation (6), the front wheel slip angle βfH based on the high μ road is calculated. Further, by substituting the vehicle slip angular velocity (dβ / dt) H and the yaw rate (dψ / dt) H based on the high μ road into the equation (3), the high μ road reference lateral acceleration (d ² y / dt ² ) H is obtained. Is calculated.

Similarly, the low μ road reference value estimating unit 12
In equation (8), for example, equivalent cornering powers Kf and Kr when the road surface μ is 0.3 are set in advance, and (dx (t) / dt in the vehicle motion state at that time (vehicle speed Vs, front wheel steering angle δfs). ) = [(Dβ / dt) (d ² } /
dt ² )] By calculating ^T , the vehicle slip angular velocity (dβ / dt) L and the yaw angular acceleration (d ² ψ / d) based on a low μ road are calculated.
t ² ) Calculate L. Then, the calculated vehicle slip angular velocity (dβ / dt) L and the yaw angular acceleration (d ²
By integrating ψ / dt ² ) L, the vehicle slip angle βL and the yaw rate (dψ / dt) L based on the low μ road are obtained. By substituting the vehicle slip angle βL based on the low μ road and the yaw rate (dψ / dt) L into the above equation (6), the front wheel slip angle βfL based on the low μ road is calculated. Furthermore,
By substituting the vehicle slip angular velocity (dβ / dt) L and the yaw rate (dψ / dt) L based on the low μ road into equation (3), the low μ road reference lateral acceleration (d ² y / dt ² ) L is calculated. Is done.

The actual value estimating section 13 serves as actual value estimating means, and includes a vehicle speed Vs, a front wheel steering angle δfs, a lateral acceleration (d ² y / dt ² ) s, and a yaw rate (dψ / dt) s.
Is an observer formed by a vehicle motion model for estimating and calculating an actual yaw rate (dψ / dt) O while feeding back the actual behavior of the vehicle.
The actual yaw rate (dψ) estimated and calculated by the actual value estimation unit 13
/ Dt) O is output to the yaw rate comparison road surface μ estimation unit 14. In addition to the actual yaw rate (dψ / dt) O, the actual value estimating unit 13 outputs the actual yaw angular acceleration (d ²
ψ / dt ² ) O is output to the yaw rate comparison road surface μ estimation unit 14. Further, the actual value estimating unit 13 calculates the actual lateral acceleration (d ² y / dt ² ) O and outputs the calculated value to the road surface μ estimation execution determining unit 15. Note that the subscript O of each parameter output from the actual value estimating unit 13 indicates that it is a parameter from the observer.

Here, the configuration of the observer according to the present embodiment will be described with reference to FIG. When the output that can be measured (detected by the sensor) is shown below, y (t) = Cx (t) (9) The configuration of the observer is as follows. (Dx ′ (t) / dt) = (A−K · C) · x ′ (t) + K · y (t) + B · u (t) (10)

Here, when this observer is applied to a vehicle motion model, x (t) becomes a state variable vector (x ′ (t)
“′” Indicates an estimated value), u (t) is an input vector, A and B are coefficient matrices of a state equation, and these correspond to those described above. Further, y (t) is an observable sensor output vector, y (t) = [βs (dψ / dt) s] ^T , and the vehicle body slip angle βs by the sensor is given by the following equation (3). Lateral acceleration (d ² y / dt ² ) s
And the vehicle slip angular velocity (dβ / dt) s obtained by the sensor based on the yaw rate (dψ / dt) s. Further, C is a matrix (in the present embodiment, a unit matrix) indicating the relationship between the sensor output and the state variable, and K is a feedback gain matrix that can be set arbitrarily, and is shown as follows.

From these relationships, the observer estimates and calculates the actual yaw angular acceleration (d ² ψ / dt ² ) O and the actual vehicle slip angular velocity (dβ / dt) O by the following equations (11) and (12). . (D ² ψ / dt ² ) O = a 11 · (dψ / dt) O + a 12 · βO + b 11 · δfs + k 11 · ((dψ / dt) s-(dψ / dt) O) + k 12 · (βs -βO) 11) (dβ / dt) O = a21 · (dψ / dt) O + a22 · βO + k21 · ((dψ / dt) s− (dψ / dt) O) + k22 · (βs−βO) (12)

The actual yaw angular acceleration (d ² ψ / dt ² ) O and the actual vehicle slip angular velocity (d
β / dt) O, the actual yaw rate (d
ψ / dt) O and the actual vehicle slip angle βO are calculated. Furthermore, the actual vehicle slip angle βO and the actual yaw rate (dψ / dt)
By substituting O into the above equation (6), the actual front wheel slip angle βfO is calculated. Also, the actual vehicle slip angular velocity (dβ
/ Dt) O and the actual yaw rate (dψ / dt) O into equation (3), the actual lateral acceleration (d ² y / dt ² )
O is calculated.

The calculation by the high μ road reference value estimating unit 11, the low μ road reference value estimating unit 12, and the actual value estimating unit 13 is performed by the vehicle speed Vs
In the case of = 0, the operation is divided by 0 and the operation cannot be performed. For this reason, at extremely low speeds (for example, speeds that do not reach 10 km / h),
The yaw rate and the lateral acceleration are sensor values. That is, the (dψ / dt) H = ( dψ / dt) L = (dψ / dt) O = (dψ / dt) s (d 2 y / dt 2) O = (d 2 y / dt 2) s . Further, the slip angle of the vehicle body is set to βH = βL = βO = δfs · Lr / (Lf + Lr) from the geometric relationship of the steady circular turning. At this time, since no cornering force is generated, the front wheel slip angles are all zero. βfH = βfL = βfO = 0

The yaw rate comparison road surface μ estimating section 14 serves as road surface μ estimating means, and includes a vehicle speed Vs, a front wheel steering angle δ.
fs sensor value and yaw rate (d / d
t) H, yaw angular acceleration (d ² ψ / dt ² ) H, lateral acceleration (d ² y / dt ² ) H, and yaw rate (d
ψ / dt) L, yaw angular acceleration ^{(d 2} ψ / dt ²⁾ and L,
Actual yaw rate (dψ / dt) O and actual yaw angular acceleration (d
^{2 ψ} / dt ²⁾ O is entered. Further, from the road surface μ estimation execution determining unit 15, a signal for stopping or executing the road surface μ estimation is input. Then, when the execution condition described later is satisfied and there is no signal to stop the estimation from the road surface μ estimation execution determination unit 15, the high μ road reference yaw rate (dψ / dt)
Based on H, the low μ road reference yaw rate (d （/ dt) L, and the actual yaw rate (dψ / dt) O, a new road surface friction coefficient μγn is calculated based on the following equation (13).

Here, μH is the high μ road reference value estimating unit 11
Road friction coefficient assumed in advance (for example, 1.0)
And μL respectively indicate a road surface friction coefficient (for example, 0.3) assumed in advance by the low μ road reference value estimating unit 12. μγn = ((μH-μL) ・ (dψ / dt) O + μL ・ (dψ / dt) H-μH ・ (dψ / dt) L) / ((dψ / dt) H- (dψ / dt) L) (13)

That is, in equation (13), a linear function is formed from the high μ road reference yaw rate (dψ / dt) H and the low μ road reference yaw rate (dψ / dt) L, and the actual yaw rate (dψ) / Dt) O is substituted to obtain a road surface friction coefficient, which is set as a new road surface friction coefficient μγn. The new road friction coefficient μγn is limited between a predetermined upper limit value (for example, 1.0) and a lower limit value (for example, another road friction coefficient value accurately obtained particularly on a low μ road). And

Then, the previously estimated road friction coefficient μn-1
The current road friction coefficient μγn and the previous road friction coefficient μ
Add a given increment obtained with n-1 (weighted average)
Thus, the road surface friction coefficient μγ based on the yaw rate comparison is calculated and output.

Μγ = μn−1 + κ1 · (μγn−μn−1) (14) Here, κ1 is a high μ road reference yaw rate (dψ / dt).
The larger the difference between H and the low μ road reference yaw rate (d （/ dt) L, the higher the reliability of the estimated value, which is determined in advance as follows. κ1 = 0.3 · | (dψ / dt) H− (dψ / dt) L | / | (dψ / dt) H | (15)

The following conditions are set in advance as conditions for calculating the road surface friction coefficient μγ by the yaw rate comparison in the yaw rate comparison road surface μ estimating section 14.

(1-1) A vehicle that is originally a multi-degree-of-freedom system is approximated by lateral movement + two degrees of freedom around a vertical axis and a two-wheel model is used. No calculation is performed during low-speed running and large turning. For example,
If the vehicle speed Vs does not reach 10 km / h, or if the absolute value of the front wheel steering angle δfs is greater than 500 deg, no calculation is performed.

(1-2) Electric noise of sensor input value,
In consideration of the influence of disturbance or the like not taken into account at the modeling stage, the calculation is not performed when the absolute value of the yaw rate at which the ratio of the influence of noise or disturbance is relatively large is small.
For example, the absolute value of the actual yaw rate (dψ / dt) O is 1.
If it does not reach 5 deg / s, no calculation is performed.

(1-3) Since the road surface friction coefficient is estimated using the difference in the cornering force due to the road surface friction coefficient, the ratio of the influence of noise and disturbance to the influence of the road surface friction coefficient is relatively large. If the increasing cornering force is small, that is, if the absolute value of the lateral acceleration proportional to the cornering force is small, the calculation is not performed. For example, a high μ road reference lateral acceleration (d ² y / dt)
² ) If the absolute value of H does not reach 0.15G, no calculation is performed.

(1-4) The yaw rate response to the steering angle input may change depending on the road surface friction coefficient and cause a delay. If the road surface friction coefficient is estimated when this delay occurs, the error increases. Therefore, when it can be determined that an error due to a delay other than when the yaw rate rises becomes large, no calculation is performed. The rise of the yaw rate is determined by, for example, (yaw rate) · (yaw angular acceleration).

(1-5) If the absolute value of the difference between the high .mu. Road reference yaw rate and the low .mu. Road reference yaw rate does not have sufficient magnitude against the influence of noise, disturbance, etc., no calculation is performed. For example, if the absolute value of the difference between the high μ road reference yaw rate (dψ / dt) H and the low μ road reference yaw rate (dψ / dt) L does not reach 1 deg / s, no calculation is performed.

(1-6) In order to avoid division by zero in the above equation (15), for example, a high μ road reference yaw rate (dψ / dt)
If the absolute value of H does not reach 1 deg / s, no calculation is performed.

Next, the road surface μ estimation execution determination unit 15 determines the lateral acceleration (d ² y / dt ² ) s from the lateral acceleration sensor 5 and the high μ road reference lateral acceleration (d ² y) from the high μ road reference value estimation unit 11.
/ Dt ² ) H, low μ road reference lateral acceleration (d ² y / dt ² ) L from the low μ road reference value estimation unit 12, and actual lateral acceleration (d ² y / dt ² ) O from the actual value estimation unit 13. Based on these input values, when it is determined that the road surface μ cannot be estimated with high accuracy, particularly in cant driving, the yaw rate comparison road surface μ estimating unit 1
4, a signal to stop the road surface μ estimation is output.

More specifically, as shown in FIG. 2, the road surface μ estimation execution determining section 15 is composed of a first execution determining section 15a, a second execution determining section 15b, and a determination result output section 15c. I have.

The first execution judging section 15a calculates the lateral acceleration (d ²
y / dt ² ) s, high μ road reference lateral acceleration (d ² y / d
t ² ) H and the low μ road reference lateral acceleration (d ² y / dt ² ) L are input, and the high μ road reference lateral acceleration (d ² y / dt ² ) H>
At the time of the low μ road reference lateral acceleration (d ² y / dt ² ) L, the detected lateral acceleration (d ² y / dt ² ) s is equal to the high μ road reference lateral acceleration (d ² y / dt ² ) H. Low μ road reference lateral acceleration (d ²
y / dt ² ), a signal for stopping the estimation of the road surface μ is output to the judgment result output unit 15c when the difference is not within the range.

That is, as shown in FIG. 5, when the vehicle is traveling on a flat road and the cornering force is determined only by the change in the road surface μ, the turning behavior of the vehicle such as the lateral acceleration and the yaw rate related to the cornering force is shown. In the parameters, the value of the actual vehicle behavior (sensor value) falls between the behavior calculated by the high μ road vehicle motion model and the low μ road vehicle motion model in a range where the phase delay is small, such as when the vehicle behavior parameter rises. . Therefore, when the deviation from the two model behaviors is obtained as in the case of the lateral acceleration (d ² y / dt ² ) s in FIG. 5, the sensor value is correctly obtained due to factors other than the road surface μ, such as the road surface cant. It can be considered that there is a high possibility that the external force has changed due to factors other than the road surface μ, such as no influence of the road surface irregularities. For this reason, the lateral acceleration (d ² y / dt ² ) s, which is the sensor value,
Is a high μ road reference lateral acceleration (d ² y / dt ² ) H based on the high μ road vehicle motion model and a low μ road reference lateral acceleration (d ² y / dt ² ) L based on the low μ road vehicle motion model. If not, the estimation of the road surface μ is stopped. That is, the first execution determination unit 15a is provided as a first determination unit.

The second execution judging section 15b calculates the lateral acceleration (d ²
y / dt ² ) s and the actual lateral acceleration (d ² y / dt ² ) O are inputted, and the lateral acceleration (d ² y / dt ² ) s and the actual lateral acceleration (d ² y / dt ² ) O are used for Lissajous. A figure is formed, and the area Sy of the Lissajous figure is set to a preset value Sc (0
If the inclination θy is less than or equal to a predetermined value (for example, an area of 43 ° or less, 47 ° or more), a signal for stopping the estimation of the road surface μ is output to the determination result output unit 15c.

That is, in a general bank, the cant is gentle at the entrance and the exit and strong at the center of the curve. Therefore, the actual lateral acceleration (d ² y / dt ² ) estimated by a vehicle model (for example, a vehicle model using an observer) assuming traveling on a flat road estimated from the actual steering angle and the vehicle speed.
O and the actual yaw rate (dψ / dt) O, these sensor values (d ² y / dt ² ) s and (dψ / dt) s are
The values are as shown in FIGS. 6A and 6B. That is, when the bank running, for the lateral acceleration sensor value ^(d 2 y / dt ²⁾ s is the actual lateral acceleration obtained in the vehicle model ^(d 2 y / dt ²⁾ smaller than O, the yaw rate , The sensor value (dψ / dt) s is larger than the actual yaw rate (dψ / dt) O obtained from the vehicle model.

As shown in FIGS. 6 (a) and 6 (b), the sensor values (d ² y / dt ² ) s and (d よ う な / dt) when traveling on a bank are the same as those when traveling on a low μ road. s actual lateral acceleration ^{(d 2 y / dt 2)} O, actual yaw rate (dψ / dt) O
Does not occur, the behavior is proportional to the actual lateral acceleration (d ² y / dt ² ) O and the actual yaw rate (d） / dt) O estimated by the vehicle model.

Accordingly, the bank running of the vehicle is determined by analyzing the change and the delay of the sensor value with respect to the value from the vehicle model, and in particular, the situation where the influence of the cant is so strong that the road surface μ cannot be estimated accurately can be determined. You do it.
For this analysis, a Lissajous figure is used as follows.

Generally, as shown in FIGS. 7 (a), 7 (b) and 7 (c), the Lissajous figure formed by a sine wave signal A and a signal B has an area which is reduced when a delay occurs between the two signals. When the signal size increases and the difference in signal magnitude occurs, FIG.
As shown in (b), the inclination changes. Therefore, a Lissajous figure is formed from the values from the vehicle model and the sensor values, and the size of the area Sy and the inclination θy are used for the determination.

The area Sy of the Lissajous figure is obtained by integrating the area of a small triangle, for example, as shown in FIG. In FIG. 8, assuming that the point (xOn-1, ySn-1) is a value one cycle before (Δt = 10 ms before) the point (xOn, ySn), the small triangular area ΔSy is ΔSy = (1/2) ) · | XOn−1 · (dySn−1 / dt) −ySn−1 · (dxOn−1 / dt) | · Δt (16)

The inclination θy of the Lissajous figure is, for example, the sensor value (d ² y / dt ² ) s for the lateral acceleration.
And the actual lateral acceleration (d ² y / dt ² ) O, ryi = (d ² y / dt ² ) O / (d ² y / dt ² ) s θy = tan ⁻¹ ((1 / n) · Σryi) (17) where i is an integer from 1 to n.

As described above, when the Lissajous figure is formed by the sensor value and the value from the vehicle model for each of the lateral acceleration and the yaw rate during canting, the result is as shown in FIG. Are both very small values. In addition, the inclination of the lateral acceleration becomes small, and the inclination of the yaw rate becomes large.

Therefore, in the present embodiment, a Lissajous figure is formed from the sensor value (d ² y / dt ² ) s and the actual lateral acceleration (d ² y / dt ² ) O for the lateral acceleration.
The area Sy of the Lissajous figure is a predetermined value Sc.
The estimation of the road surface μ is stopped when the inclination θy is equal to or less than (a value close to 0) and the inclination θy is in a preset area (for example, an area of 43 ° or less). Also, when the inclination of the Lissajous figure is, for example, 47 ° or more, a sensor value is not correctly obtained due to a factor other than the road surface μ, or an external force changes due to a factor other than the road surface μ such as an influence of road surface unevenness. The estimation of the road surface μ is stopped because it is considered that there is a high possibility that the road surface μ is present. That is, the second execution determination unit 15b is provided as a second determination unit.

The judgment result output unit 15c receives the judgment result from the first execution judgment unit 15a and the second execution judgment unit 15b, and if any one of the judgment results is the stop of the execution, the yaw rate comparison road surface μ estimation. A signal to stop the estimation of the road surface μ is output to the unit 14.

By providing the first execution judging unit 15a and the second execution judging unit 15b in this way, the first execution judging unit 15a can detect the sensor value (d ² y /
dt ² ) s is the high μ road reference lateral acceleration (d ² y / dt ² ) H
And the low μ road reference lateral acceleration (d ² y / dt ² ) L, the second execution determination unit 15b performs the determination based on the Lissajous figure to reliably execute the determination of the road μ estimation. I can do it.

Next, the road surface μ estimation execution determination control by the road surface μ estimation execution determination unit 15 will be described with reference to the flowchart of FIG. This program is executed every predetermined time (for example, every 10 ms).

First, step (hereinafter abbreviated as “S”) 1
01, the high μ road reference lateral acceleration (d ² y / dt ² ) H is
It is determined whether it is greater than the low μ road reference lateral acceleration (d ² y / dt ² ) L. As a result of this determination, the high μ road reference lateral acceleration (d ² y / dt ² ) H is changed to the low μ road reference lateral acceleration (d ² y).
If it is greater than / dt ² ) L, the process proceeds to S102, and if the high μ road reference lateral acceleration (d ² y / dt ² ) H is equal to or less than the low μ road reference lateral acceleration (d ² y / dt ² ) L, S103. Proceed to.

High μ road reference lateral acceleration (d ² y / dt ² ) H
Is larger than the low μ road reference lateral acceleration (d ² y / dt ² ) L and proceeds to S102, the sensor value (d ² y / dt ² ) s
Is between the high μ road reference lateral acceleration (d ² y / dt ² ) H and the low μ road reference lateral acceleration (d ² y / dt ² ) L. The sensor value ^{(d 2 y / dt 2)} s
Deviates from between the high μ road reference lateral acceleration (d ² y / dt ² ) H and the low μ road reference lateral acceleration (d ² y / dt ² ) L, It can be considered that there is a high possibility that the sensor value is not obtained correctly due to a factor or the external force is changed due to a factor other than the road surface μ such as the influence of the road surface unevenness. Therefore, the process proceeds to S107 and the estimation of the road surface μ is stopped. A signal is output to cause the routine to exit. On the other hand, the sensor value (d ² y / dt ² ) s is between the high μ road reference lateral acceleration (d ² y / dt ² ) H and the low μ road reference lateral acceleration (d ² y / dt ² ) L. If there is, the process proceeds to S103.

[0059] Proceeding S101 or from S102 to S103, the sensor value of the lateral acceleration ^(d 2 y / dt ²⁾ s and the actual lateral acceleration ^(d 2 y / dt ²⁾ in the O to form a Lissajous figure, the area Sy And the inclination θy.

Then, the process proceeds to S104, and it is first determined whether or not the inclination θy of the Lissajous figure is between θy1 (for example, 47 °) and θy2 (for example, 43 °) set in advance. As a result, the inclination θy of the Lissajous figure becomes
If it is between θy1 and θy2, there is a high possibility that the vehicle is traveling on a high μ road, so the process proceeds to S106, and the routine exits while the road surface μ estimation is still executed.

On the other hand, the inclination θy of the Lissajous figure is θy1
If the value of the area deviates from θy2, the process proceeds to S105, and the area Sy of the Lissajous figure is compared with a predetermined value Sc close to 0. When the area Sy is larger than Sc, the lateral acceleration sensor value (d ² y / dt)
² ) s and the actual lateral acceleration (d ² y / dt ² ) O have a delay, and it is highly likely that the vehicle is traveling on a low μ road.
Proceeding to 6, the program exits the routine with the road surface μ estimation still being executed. Conversely, when the area Sy is equal to or smaller than Sc, the sensor value (d ² y / dt ² ) s of the lateral acceleration and the actual lateral acceleration (d ² y)
/ Dt ² ) O, there is no delay, and the sensor value is not correctly obtained due to factors other than the road surface μ, such as cant on the road surface, or the external force changes due to factors other than the road surface μ, such as the influence of road surface irregularities. Since it is considered that there is a high possibility that the road surface μ is estimated, the process proceeds to S107 and outputs a signal to stop the road surface μ estimation and exits the routine.

As described above, in the present embodiment, especially when the vehicle travels on a bank or the like, the difference between the sensor value and the value obtained by the vehicle model is assumed in advance regardless of the road surface μ, and based on this, the difference is obtained. Since the cancellation of μ estimation is determined, especially when the influence of cant driving is strong and the road surface μ cannot be estimated with high accuracy, the estimation of the road surface μ is stopped without fail, and the estimated road surface μ with low reliability and large error is output. Can be reliably prevented.

In particular, by combining the observer, the road friction coefficient estimating apparatus for estimating the road surface μ based on the yaw rate from the high μ road vehicle motion model and the low μ road vehicle motion model, the vehicle necessary for estimating the road surface μ is obtained. Since it is possible to judge the execution of the road surface μ estimation based on the parameters obtained by sharing the motion model, the configuration can be simplified and the efficiency can be improved by adding only a small number of components.

In this embodiment, the first execution determining unit 1
The comparison between the sensor value and the high μ road reference value and the low μ road reference value in 5a is based on the lateral acceleration, but alternatively, a yaw rate or the like may be used. Similarly, the comparison between the sensor value and the actual value in the second execution determination unit 15b may use a yaw rate or the like in addition to the lateral acceleration.

In the present embodiment, both the first execution determination unit 15a and the second execution determination unit 15b are provided to determine the execution of the road surface μ estimation. In consideration of this, only one of them may be used.

[0066]

As described above, according to the present invention, when the influence of cant driving is particularly strong and the road surface μ cannot be accurately estimated, the estimation of the road surface μ is stopped without fail, and the reliability is low and the error is large. It is possible to reliably prevent the road μ estimation value from being output.

[Brief description of the drawings]

FIG. 1 is a functional block diagram showing a configuration of a road surface friction coefficient estimating apparatus.

FIG. 2 is a functional block diagram of a road surface μ estimation execution determination unit;

FIG. 3 is an explanatory diagram showing an equivalent two-wheeled vehicle model of a four-wheeled vehicle;

FIG. 4 is an explanatory diagram showing a configuration of an observer.

FIG. 5 is an explanatory diagram of a high μ road reference value, a low μ road reference value, and a sensor value of lateral acceleration.

FIG. 6 is an explanatory diagram of actual values and sensor values of lateral acceleration and yaw rate.

FIG. 7 is an explanatory diagram of a Lissajous figure drawn by two waveforms.

FIG. 8 is an explanatory diagram of calculating a Lissajous figure area;

FIG. 9 is an explanatory diagram of a Lissajous figure formed with respect to a lateral acceleration and a yaw rate during cant driving.

FIG. 10 is a flowchart of road surface μ estimation execution determination control;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Road surface friction coefficient estimation device 3 Front wheel steering angle sensor 4 Vehicle speed sensor 5 Lateral acceleration sensor 6 Yaw rate sensor 11 High μ road reference value estimation unit (high μ road reference value estimation means) 12 Low μ road reference value estimation unit (low μ road) (Reference value estimating means) 13 Actual value estimating section (Actual value estimating means) 14 Yaw rate comparison road surface μ estimating section (Road surface μ estimating means) 15 Road surface μ estimating execution judging section 15a First execution judging section (First judging means) 15b Second execution determination unit (second determination unit) 15c Determination result output unit

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] October 30, 2001 (2001.10.
30)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0006

[Correction method] Change

[Correction contents]

[0006]

According to a first aspect of the present invention, there is provided a vehicle road surface friction coefficient estimating apparatus for estimating a road surface μ based on at least a parameter indicating a turning behavior of the vehicle. In the road surface friction coefficient estimating apparatus, a high μ road reference value related to or the same parameter as the above parameter is calculated based on a high μ road vehicle motion model, and the above parameter is related or the same parameter based on a low μ road vehicle motion model. low μ road and calculates the reference value, the actually detected values of the above parameters with related or identical parameters, road surface when not in the region between the upper Symbol high μ road reference value and the upper Symbol low μ road reference value of Stop estimation of μ
A first determination means for the determination, based on the vehicle motion model calculates the actual value of the associated or the same parameters as the parameters to form a Lissajous figure in the detected value and the upper Symbol actual value when upper you a second determination means for the determination for stopping the estimation of the road surface μ when in the area of the Lissajous figure is set small and the gradient is pre region, the first
The road surface is determined by either one of the judgment means and the second judgment means.
If it is decided to stop estimating μ,
And a determination result output means for canceling . That is, the road surface of the vehicle according to the present invention described in claim 1
The friction coefficient estimating device indicates at least the turning behavior of the vehicle
The road surface μ is estimated based on the parameters.
The stage is provided for either one of the first determining means and the second determining means.
If it is determined to stop estimating the road surface μ,
Cancel the setting. Therefore, the impact of cant driving is particularly strong.
If the road surface μ cannot be estimated accurately, the road surface μ is estimated.
Is reliably stopped and the road surface μ
The output of the constant value is reliably prevented.

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name] 0007

[Correction method] Change

[Correction contents]

[0007] The vehicle according to the second aspect of the present invention.
The road friction coefficient estimating device is
The actual value of the parameter that indicates the turning behavior of the vehicle is calculated.
Threshold value estimation means and vehicle motion model on high μ road
High μ road reference value estimator that calculates the reference values of the above parameters
Step and vehicle motion model
A low μ road reference value estimating means for calculating a reference value of
Compare the actual value with the high μ road reference value and the low μ road reference value.
Road μ estimation means for estimating the current road μ by comparing
The road friction coefficient estimating device for
The above parameters are obtained from the vehicle motion model for estimating the quasi-value.
Calculate high μ road reference value related to or same parameter
And the vehicle motion estimating the low μ road reference value
Depending on the model, it is related to the same parameter or the same parameter.
Calculates the low μ road reference value of the
Or the actually detected value of the same parameter is
If it is not in the area between the road reference value and the low μ road reference value
First judging means for judging to stop estimation of road surface μ
And the observer is related to the above parameters.
Calculates the actual value of the same parameter,
A Lissajous figure is formed from the calculated value and the actual value, and
Area where the area of the jewel figure is small and the inclination is set in advance
To stop the estimation of the road surface μ when there is
Determining means, the first determining means, and the second determining means
One of the steps determines that the estimation of the road surface μ is stopped.
Judgment result output means for stopping the estimation of the road surface μ in the case of
It is characterized by having. That is, according to claim 2
The apparatus for estimating the road friction coefficient of a vehicle according to the present invention estimates the actual value.
At least the turning behavior of the vehicle by the observer
Calculates the actual value of the indicated parameter and estimates the reference value for the high μ road.
The above parameters on high μ road by vehicle motion model at step
The reference value of the vehicle is calculated, and the vehicle
The reference value of the above parameters on low μ road is
Calculate and use the road surface μ estimation means to calculate the actual value as
The current road surface μ is estimated by comparing the
Set. Here, the judgment result output means is the first judgment means.
Estimation of the road surface μ is stopped in either of the second judgment means.
When it is determined that the road surface μ is to be estimated, the estimation of the road surface μ is stopped. this
Therefore, the influence of cant driving is particularly strong, and the road surface μ is accurately determined.
If the estimation cannot be performed, the estimation of the road
The road surface μ estimated value with low reliability and large error is output
Is reliably prevented. Necessary for estimating road surface μ
Three vehicle motion models in the road surface μ estimation
Efficiency is high because it can be used by having.

[Procedure amendment 4]

[Document name to be amended] Statement

[Correction target item name] 0008

[Correction method] Change

[Correction contents]

Further , the vehicle road surface friction coefficient estimating apparatus according to the third aspect of the present invention provides at least a turning behavior of the vehicle.
Road friction of vehicles estimating road μ based on indicated parameters
In the coefficient estimating device, based on the vehicle motion model,
Performs the actual value of the parameter related to or the same as the parameter.
Of the same or related parameters
Relevant or same as the actually detected value and the above parameter
A Lissajous figure is formed with the above actual values of the parameters,
The area of the Lissajous figure is small and the inclination is set in advance.
Determining means to stop estimating the road surface μ when the vehicle is in the
It is characterized by having. That is, according to claim 3
An apparatus for estimating a road surface friction coefficient of a vehicle according to the present invention has at least
Also estimates the road surface μ based on parameters that indicate the turning behavior of the vehicle
However, the judgment means is that the area of the Lissajous figure is small and
Estimate the road surface μ when the slope is in the preset area.
To stop. Therefore, the influence of cant driving is particularly strong, and
If the surface μ cannot be estimated accurately, the estimation of the road surface μ is reliable.
And the road surface μ estimated value with low reliability and large error
Output is reliably prevented.

[Procedure amendment 5]

[Document name to be amended] Statement

[Correction target item name] 0009

[Correction method] Change

[Correction contents]

Further, the road surface friction coefficient estimating apparatus for a vehicle according to the present invention according to claim 4, at least by the observer
The actual value of the parameter that indicates the turning behavior of the vehicle is calculated.
Threshold value estimation means and vehicle motion model on high μ road
High μ road reference value estimator that calculates the reference values of the above parameters
Step and vehicle motion model
A low μ road reference value estimating means for calculating a reference value of
Compare the actual value with the high μ road reference value and the low μ road reference value.
Road μ estimation means for estimating the current road μ by comparing
In the apparatus for estimating the road surface friction coefficient of a vehicle,
Related to or the same parameter as above
Calculate the actual value of
Associated with the actually detected value of the parameter and the above parameter
Or Lissajous figure with the above actual value of the same parameter
Shape, and the area of the Lissajous figure is small and tilted.
Estimation of road surface μ is stopped when
And determining means for making the determination. That is,
An apparatus for estimating a road friction coefficient of a vehicle according to the present invention according to claim 4.
Means at least the vehicle
Calculate the actual values of the parameters that indicate the turning behavior of
On a high μ road using a vehicle motion model with reference value estimation means
Calculates the reference value of the above parameters and estimates the low μ road reference value.
The above parameters on the low μ road are calculated by the vehicle motion model at the step.
Calculate the reference value of the data and increase the actual value
Compared to the high μ road reference value and the low μ road reference value
Is estimated. At this time, the judgment means is Lissajous figure
If the area of the shape is small and the slope is in a preset area,
In this case, the estimation of the road surface μ is stopped. Because of this, especially cant
When the influence of traveling is strong and the road surface μ cannot be estimated accurately
Indicates that the estimation of the road surface μ is stopped
Output of a large road surface μ estimation value is reliably prevented.
You. Also, the three vehicle motions required to estimate the road surface μ
The model can be shared and used for the execution judgment of road surface μ estimation
It is efficient.

[Procedure amendment 6]

[Document name to be amended] Statement

[Correction target item name] 0054

[Correction method] Change

[Correction contents]

The judgment result output unit 15c receives the judgment result from the first execution judgment unit 15a and the second execution judgment unit 15b, and if any one of the judgment results is the stop of the execution, the yaw rate comparison road surface μ estimation. A signal to stop the estimation of the road surface μ is output to the unit 14. That is, the judgment result
The force unit 15c is provided as a determination result output unit.
You.

[Procedure amendment 7]

[Document name to be amended] Statement

[Correction target item name] Explanation of sign

[Correction method] Change

[Correction contents]

[Description of Signs] 1 Road surface friction coefficient estimation device 3 Front wheel steering angle sensor 4 Vehicle speed sensor 5 Lateral acceleration sensor 6 Yaw rate sensor 11 High μ road reference value estimation unit (high μ road reference value estimation means) 12 Low μ road reference value estimation Section (low μ road reference value estimation means) 13 actual value estimation section (actual value estimation means) 14 yaw rate comparison road surface μ estimation section (road surface μ estimation means) 15 road surface μ estimation execution determination section 15a first execution determination section (first 15b Second execution determination unit (second determination unit) 15c Determination result output unit (determination result output unit)

Claims (2)

[Claims] An apparatus for estimating a road surface friction coefficient of a vehicle based on at least a parameter indicating a turning behavior of the vehicle, comprising: While calculating the reference value, based on the low μ road vehicle motion model, calculate the low μ road reference value related to or the same parameter as the above parameter,
The value actually detected of the parameter related or the same parameter is in an area between the high μ road reference value of the parameter related or the same parameter and the low μ road reference value of the parameter related or the same parameter. First determining means for stopping the estimation of the road surface μ when there is no actual value of the parameter related to or the same as the above parameter based on the vehicle motion model, and actually detected value of the parameter related to or the same as the parameter Second determining means for forming a Lissajous figure with the parameter and the actual value of the same parameter or the same parameter, and stopping the estimation of the road surface μ when the area of the Lissajous figure is small and the slope is in a preset area. And a road friction coefficient estimating device for a vehicle, comprising at least one of the following. 2. An actual value estimating means for calculating at least an actual value of a parameter indicating a turning behavior of the vehicle by an observer, and a high μ road reference value estimating means for calculating a reference value of the parameter on a high μ road by a vehicle motion model. A low μ road reference value estimating means for calculating a reference value of the parameter on the low μ road by a vehicle motion model; comparing the actual value with the high μ road reference value and the low μ road reference value to obtain a current value. A road friction coefficient estimating apparatus for a vehicle, comprising: a road surface μ estimating means for estimating a road surface μ; wherein the vehicle motion model for estimating the high μ road reference value is used to calculate a high μ road reference value related to or the same parameter as the parameter. Calculating and calculating a low μ road reference value related to or the same parameter as the parameter by the vehicle motion model estimating the low μ road reference value, The actually detected value of the parameter related or the same parameter is not in an area between the high μ road reference value of the parameter related or the same parameter and the low μ road reference value of the parameter related or the same parameter. First determining means for stopping the estimation of the road surface μ in the case; calculating an actual value of the parameter related to or the same as the parameter by the observer; A second determining means for forming a Lissajous figure with the above-mentioned actual values of related or same parameters, and stopping the estimation of the road surface μ when the area of the Lissajous figure is small and the slope is in a preset area. An apparatus for estimating a road friction coefficient of a vehicle, comprising: