JP3008769B2 - Vehicle lateral speed detection device - Google Patents

Abstract

PURPOSE:To suppress the accuracy reduction of the vehicle lateral speed owing to an integration error of a sensor value, by detecting the actual vehicle lateral speed by correcting the time integration value of the value subtracting the lateral acceleration from the product of a yaw rate and a vehicle longitudinal speed, by the logical laterial speed in case of assuming a two freedom plane model. CONSTITUTION:By detecting Gy, theta, and Vx by sensors 20 to 24, and by integrating (gamma.Vx-Gy) by a computer 30, the actual lateral speed Vy is operated. On the other hand, a sensor 26 detects the actual steering angle delta, and the computer 30 operates a logical lateral speed VYM in case of assuming that the vehicle in a normal circular turning condition under a two freedom plane model, depending on the delta and Vx. And by correcting the Vy by the VyM, the final Vy is found, and by dividing the Vy by the Vx, the slippage angle beta at the center of gravity of the vehicle is found. Furthermore, the friction coefficient mu of the road surface is inferred by the relation of the slippage angle beta and the tire lateral force F.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for detecting a lateral speed of a vehicle, and more particularly to a technique for improving the accuracy of the detection.

[0002]

2. Description of the Related Art It has been already performed to detect a slip angle at a center of gravity of a vehicle body for estimating a behavior of a vehicle or a friction coefficient of a road surface. As a method of detecting the slip angle, for example, a method of detecting the slip speed β (= V _y / V _x ) by detecting the longitudinal speed and the lateral speed of the vehicle, respectively, and dividing the lateral speed V _y by the longitudinal speed V _x. Already exists. In this system, the detection of the front-rear speed is relatively easy, but the detection of the lateral speed is relatively difficult. This is because the lateral speed cannot be detected by the vehicle speed sensor using the rotation of the wheels.

As a method of detecting a lateral velocity, a method of detecting a lateral velocity sensor using the principle of a wave Doppler effect or a spatial filter is already known, but there is a problem that the sensor is relatively expensive.

As another detection method, Japanese Patent Laid-Open No.
As described in JP-A-83247, vehicle yaw
Rate γ and front-rear speed V_xAnd the lateral acceleration G at the center of gravity of the vehicle
_yAnd the yaw rate γ and the front-rear speed
Degree V_xLateral acceleration G from the product of _yIs subtracted from the time t
And integrate to obtain the lateral velocity V_yThe method of detecting
Already exists. This means that V_y= ∫ (γ · V_X-G_y) Using the equation dt, the lateral velocity V_yIs a method of detecting

[0005]

However, the detection method described in this publication has the following problems. That is, in this detection method, since each detection value of the sensor is integrated as it is, if there is an error in each detection value, the error is also integrated. Therefore, the error is, for example,
If the error is a stationary error such as a mounting angle error of the sensor, the error is accumulated as time elapses, and there is a problem that the detection accuracy of the lateral speed is reduced.

[0006] To solve this problem, the present applicant has previously made the following proposal. That is, although the above-described detection method is common in that each detection value is integrated, different from the above-described detection method, each sensor (lateral acceleration sensor, yaw rate sensor , and the like) is different.
Based on raw detection values obtained by the
The transverse velocity derivative is obtained by simply integrating over time.
When the virtual lateral velocity , which is the lateral velocity, is decomposed into a steady component and a variable component, the raw detection values are such that the influence of the variable component appears relatively larger in the detected value of the lateral speed than the effect of the steady component. Proposed a method of integrating. This specific example will be described in detail in Examples. According to the proposed detection method, even if a steady error exists in each raw detection value, the lateral speed can be detected without causing the influence of the steady error. Divergence of the detected speed value is avoided.

However, there is a problem in the proposed detection method. That is, in this detection method, the steady-state component is uniformly suppressed regardless of the cause of the occurrence of the steady-state component in the virtual lateral speed. Even when there is no error, the steady-state component of the virtual lateral velocity is suppressed, and there is a problem that it is difficult to always accurately detect the lateral velocity.

In short, it is difficult to always accurately detect the lateral speed only by using the lateral speed detected by integrating the detected values of the vehicle state quantity.
The invention of the present invention not only estimates the lateral speed by integrating the detected value of the vehicle state quantity, but also estimates the lateral velocity without integration based on the detected value of the vehicle state quantity. It is an object of the present invention to make it possible to always accurately estimate the lateral speed regardless of the steady-state error of each sensor by comprehensively using the sensors.

[0009]

In order to solve the problem, the invention according to claim 1 comprises a vehicle lateral speed detecting device comprising: (a) a lateral acceleration sensor for detecting a lateral acceleration at a center of gravity of the vehicle; ) A yaw rate sensor for detecting the yaw rate of the vehicle, and (c) a longitudinal speed sensor for detecting the longitudinal speed of the vehicle,
(d) A steering angle sensor that detects the steering angle of the vehicle, and (e) A value obtained by subtracting the detection value of the lateral acceleration from the product of the detection value of the front-rear speed and the detection value of the yaw rate with respect to time. estimating the basic value of the actual lateral speed, based on the detected value of the detection value and the steering angle of the longitudinal velocity, the vehicle is to estimate the theoretical lateral velocity of the vehicle when it is assumed to be in steady circular turning state, and the estimated
The basic value of the actual lateral speed is corrected by the theoretical lateral speed
Processing means for estimating the final value of the actual lateral speed of the vehicle.

[0010]

When the vehicle is in a steady circular turning state, a constant relationship is established between the longitudinal speed, the steering angle and the lateral speed. In this relationship, the lateral speed is integrated without integrating the longitudinal speed and the steering angle. Can be induced. Even if a steady-state error exists in the detection value of the sensor, a lateral speed in which the steady-state error does not accumulate can be induced.

As described above, the theoretical lateral speed of the vehicle can be estimated without integrating each detected value, and this theoretical lateral speed has an advantage that it is hardly affected by the steady-state error of each sensor.
When the vehicle is not actually in a steady circular turning state, it does not match the true lateral velocity with sufficient accuracy, and when the cornering characteristics of the tire are in the nonlinear region, the true lateral speed is sufficiently accurate. The disadvantage is that it does not match the lateral speed. On the other hand, the actual lateral velocity obtained by integrating the lateral velocity derivative based on the detection value of each sensor has the disadvantage that it is easily affected by the steady error of each sensor, but there is no steady error of each sensor. Below, there is the advantage that the vehicle coincides with the true lateral speed with sufficient accuracy irrespective of whether or not the vehicle is actually in a steady circular turning state.

In short, the actual lateral speed and the theoretical lateral speed have a relationship that one of the drawbacks is regarded as an advantage of the other. In the vehicle lateral speed detecting device according to the first aspect of the present invention, the actual lateral speed and the theoretical lateral speed are determined by the theoretical lateral speed. Correct the basic value of the actual lateral speed.
Compensates for the disadvantages between the actual lateral speed and the theoretical lateral speed.
As a result, the final value of the actual lateral speed of the vehicle is estimated. Specifically, for example, the vehicle lateral speed according to claim 2
As in the detection device, the basic actual lateral speed is estimated to gradually decay with time in a state where the virtual lateral speed is stationary, and the estimated basic actual lateral speed is Rather than decay to zero over time, it is corrected to decay toward the theoretical lateral speed, and the resulting value is the final actual lateral speed.
Further, in the vehicle lateral speed detecting device according to the third aspect,
As described above, the basic actual lateral speed is corrected to an average value with the theoretical lateral speed estimated at the same time, and the resulting value is set as the final actual lateral speed.

[0013]

As is apparent from the above description, according to the first aspect of the present invention, both the basic actual lateral speed estimated through the integration process and the theoretical lateral speed estimated without the integration process are obtained. Is used to detect the final actual lateral speed, so that the detection value is less likely to be affected by the steady-state error of the sensor, and the effect of improving the lateral speed detection accuracy is obtained.

[0014]

Preferred embodiments of the present invention will be described below. (1) The invention according to claim 1, wherein the processing means integrates a value obtained by subtracting a detected value of a lateral acceleration from a product of a detected value of a longitudinal speed and a detected value of a yaw rate with respect to time, thereby obtaining a vehicle speed. Based on basic actual lateral speed estimating means for estimating the basic value of the actual lateral speed, and based on the detected value of the longitudinal speed and the detected value of the steering angle,
Theoretical lateral speed estimating means for estimating the theoretical lateral velocity of the vehicle assuming that the vehicle is in a steady circular turning state, and estimating the basic actual lateral speed estimated by the basic actual lateral velocity estimating means by the theoretical lateral velocity estimating means Actual lateral speed correcting means for determining the final value of the actual lateral speed by correcting with the calculated theoretical lateral speed.

(2) The invention according to (1), wherein the basic actual lateral velocity estimating means estimates the basic actual lateral velocity by simply integrating the raw detection values of each sensor.

(3) The invention according to (1), wherein the basic actual lateral velocity estimating means integrates a plurality of raw detection values obtained by each sensor with the passage of time to obtain the basic actual lateral velocity. The speed is estimated and its integral is obtained by each sensor.
The derivative of the lateral velocity based on the raw detection value
When the virtual lateral velocity , which is the lateral velocity obtained by simple integration, is decomposed into a steady component and a variable component, the effect of the variable component appears relatively larger than the effect of the steady component in the basic actual lateral speed. Something to do.

(4) The invention according to (3), wherein the basic actual lateral speed estimating means estimates the basic actual lateral speed for each detection cycle, and calculates the basic actual lateral speed by the previous value and a coefficient smaller than 1. And the product of the current value of the lateral velocity derivative, which is the value obtained by subtracting the current value of the lateral acceleration from the product of the current value of the yaw rate and the current value of the longitudinal speed, and the detection cycle This time the value.

(5) In each of the inventions (1) to (4), the actual lateral speed correcting means calculates an average of the basic actual lateral speed and the theoretical lateral speed estimated at the same time. The final actual lateral speed is determined by correcting to a value (claim 3) .

(6) In each of the inventions (1) to (4), the actual lateral speed correcting means changes the basic actual lateral speed as the period during which the virtual lateral speed is stationary progresses. The estimation is made so as to approach the theoretical lateral speed (claim 2) . Need correction

(7) In the invention of (6), the actual lateral speed correcting means of the invention calculates the present value of the final actual lateral speed as the current value of the theoretical lateral speed, the basic actual lateral speed and the theoretical lateral speed. Calculated as the sum of the product of the previous value of the lateral speed deviation, which is the deviation of the above, and a coefficient smaller than 1, and the product of the current value of the lateral speed differential, which is the time differential value of the basic actual lateral speed, and the lateral speed detection period. What to do.

(8) The invention according to any one of (1) to (7), wherein the actual lateral speed correcting means converts at least the detected values of lateral acceleration, yaw rate, longitudinal speed and steering angle into an input signal, A neural network that has a correction value of the actual lateral speed as an output signal, and has information in which coupling information among a plurality of units constituting the neural network is determined in advance under actual vehicle driving conditions, and the output signal includes Corrects the basic actual lateral speed based on

(9) The invention according to claim 1 or (1) to (8), further comprising a rolling motion state quantity sensor for detecting a rolling motion state quantity of the vehicle body, and the rolling motion state quantity sensor detecting the rolling motion state quantity. Lateral acceleration correction means for correcting the detected lateral acceleration by removing a disturbance based on the rolling motion state amount from the lateral acceleration detected by the lateral acceleration sensor based on the determined rolling motion state amount. .

(10) A vehicle lateral speed detecting device according to the invention of claim 1 or (1) to (9), a vehicle motion state changing device for changing a vehicle motion state, and at least the vehicle lateral speed detecting device. A controller for controlling a vehicle motion state via the vehicle motion state changing device based on the final lateral speed detected by the device.

(11) A lateral acceleration sensor for detecting the lateral acceleration at the center of gravity of the vehicle, a yaw rate sensor for detecting the yaw rate of the vehicle, a longitudinal speed sensor for detecting the longitudinal speed of the vehicle, and a steering angle for the vehicle. A basic value of the actual lateral speed of the vehicle is estimated by integrating the value obtained by subtracting the detected value of the lateral acceleration from the product of the steering angle sensor and the detected value of the detected longitudinal speed and the detected value of the yaw rate with respect to time. Based on the detected value and the detected value of the steering angle, a theoretical lateral speed of the vehicle when the vehicle is assumed to be in a steady circular turning state is estimated, and based on the basic value and the theoretical lateral speed of the estimated actual lateral speed. Processing means for estimating the final value of the actual lateral speed of the vehicle, and slip angle estimating means for estimating the slip angle at the center of gravity of the vehicle body by dividing the final value of the estimated actual lateral speed by the detected value of the longitudinal speed. And A vehicle body slip angle detection device characterized by including:

(12) The vehicle body slip angle detecting device according to (11),
Tire lateral force estimating means for estimating a tire lateral force generated in the tire in the vehicle lateral direction based on the detected value of the longitudinal speed and the time differential value of the detected value of the yaw rate, and an estimated value of the slip angle and an estimated tire lateral force A road friction coefficient estimating means for estimating a friction coefficient of a road surface on which the vehicle travels based on the value.

(13) In the invention of (12), the road surface friction coefficient estimating means permits updating of the estimated value of the road surface friction coefficient in a state where the estimated value of the slip angle exceeds the set value of the slip angle. , The renewal of which is prohibited without exceeding.

(14) In the invention of (12) or (13), the road surface friction coefficient estimating means sets the slip angle set value in a state where the estimated value of the slip angle exceeds the set value of the slip angle. Update of the estimated value of the road surface friction coefficient is permitted during a period in which the time that continues to exceed does not exceed the first set time, update is prohibited during a period that exceeds the first set time, and estimation of the slip angle is performed. Even if the estimated slip angle shifts to the state where the estimated slip angle exceeds the slip angle set value before the second set time elapses from when the value shifts from the state where the slip angle set value is exceeded to the state where the slip angle does not exceed the slip angle set value, A device that prohibits updating of the estimated value of the coefficient of friction.

(15) A device for detecting a road surface friction coefficient according to any one of (12) to (14), a vehicle motion state changing device for changing a vehicle motion state, and at least a road surface friction detected by the road surface friction coefficient detecting device. A controller for controlling a motion state of the vehicle via the vehicle motion state changing device based on the coefficient.

(16) The vehicle control device according to (15), wherein the vehicle motion state changing device is an anti-lock control device for preventing wheel lock during vehicle braking, and a traction for preventing wheel slip during vehicle driving. A control device, an electric steering device for electrically controlling the steering angle of the vehicle, a driving force distribution control device for controlling the distribution of the driving force to the front and rear wheels, and a control for controlling the distribution of the braking force to the front and rear wheels or the left and right wheels. A vehicle including at least one of a power distribution control device and a roll rigidity distribution control device for controlling distribution of roll rigidity to front and rear wheels.

[0030]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a first embodiment of the present invention;

The illustrated embodiment is a vehicle control device provided with an anti-lock control device as a vehicle motion state changing device. The vehicle control device, the longitudinal velocity V _x of the vehicle and the lateral velocity V _y respectively detected, the vehicle body slip angle by dividing the lateral speed V _y at a longitudinal speed V _x beta (= V _y / V _x)
And a road friction coefficient detecting device for detecting a friction coefficient μ of the road surface based on the detected friction coefficient and the tire lateral force F.

The vehicle control device is mounted on a four-wheel vehicle having left and right front wheels 10 and left and right rear wheels 12, as shown in FIG. Road surface friction coefficient detecting device 14 of the vehicle control device, as shown in the figure, the lateral acceleration sensor 20 for detecting a lateral acceleration G _y of the vehicle center of gravity, a yaw rate sensor 22 for detecting a yaw rate γ of the vehicle, the vehicle longitudinal velocity sensor for detecting a longitudinal velocity V _x 24, left and right front wheels 10
And a computer 30 for detecting the actual steering angle δ of the vehicle. Computer 3
0 is configured to include a CPU, a ROM, and a RAM.

The vehicle control device further comprises an anti-lock control device 32. The antilock control device 32
Includes wheel speed sensors, electro-hydraulic pressure control valves, computers, etc., and brakes each wheel so that each wheel does not lock during vehicle braking based on the rotation status of each wheel (wheel speed, wheel acceleration, locking tendency, etc.). The pressure is electrically controlled. The anti-lock control device 32 can change its brake pressure control characteristics (for example, pressure reduction start timing, pressure reduction time, pressure reduction gradient, pressure increase gradient, etc.) according to the road surface friction coefficient (hereinafter also referred to as road surface μ). ing.

FIG. 2 is a block diagram showing an electrical configuration of the road friction coefficient detecting device 14. As shown in FIG. The road surface friction detecting device 14
As conceptually shown in FIG.
0, theoretical lateral velocity estimating means 52, actual lateral velocity correcting means 54,
It includes slip angle estimating means 56, differentiating means 58, tire lateral force estimating means 60 and road surface friction coefficient estimating means 62.

The basic actual lateral velocity estimating means 50 integrates the lateral velocity differential V _y ′, which is a value obtained by subtracting the detected value of the lateral acceleration G _y from the product of the detected value of the yaw rate γ and the detected value of the longitudinal velocity V _x. By doing so, the actual lateral speed _Vy is estimated.
Therefore, the lateral acceleration sensor 20, the yaw rate sensor 22, and the longitudinal speed sensor 24 are connected to the input side of the basic actual lateral speed estimating means 50. Below is the actual lateral speed V
The principle of estimating _y will be described in detail.

[0036] In this basic actual lateral speed estimating unit 50 is basically a method of detecting actual lateral velocity V _y, method using a _{V y = ∫ (γ · V} X -G y) dt becomes equation is employed ing. _{That, V y (n) = V} y (n-1) + Δt · V y '(n) becomes equation (where, Delta] t represents the detection period of the lateral velocity V _y)
Is used to estimate the actual lateral speed V _y (an example of the basic actual lateral speed). In this equation, the value of the variable with (n) means the current value, and the value of the variable with (n-1) means the previous value. This is the same for other variables.

However, if the detection values obtained by the sensors are integrated as they are using the above equations as they are, if there is an error in each detection value, the error is also integrated.
Therefore, if the error is a stationary error such as a mounting angle error of the sensor, the stationary error of the sensor is accumulated in the detected value of the lateral velocity as time passes. Therefore, when the slip angle is estimated using the lateral speed detected in this manner, the estimated slip angle is, for example, as shown in the graph as the first estimated slip angle on the upper side of FIG. The error with respect to (slip angle) changes so as to increase with time.

Therefore, in this embodiment, the integral of the lateral velocity derivative V _y ′ (= γ · V _X −G _y ) is obtained by each sensor.
The lateral velocity derivative based on the raw detection value
When the virtual lateral velocity, which is the lateral velocity obtained by simple integration and integration, is decomposed into a steady component and a variable component, the effect of the variable component is relatively larger than the effect of the steady component. Lateral speed V
Performed to appear in _y .

One of the steady-state component suppression type lateral speed detection methods is described below.
As an example, in the basic actual lateral speed estimation means 50,
Actual lateral speed V_yAnd the product of the coefficient C and the lateral speed
Minute V _y′ As the sum of the product of the current value of
Lateral speed V_yThe method of obtaining the current estimated value of
You. That is, V_{y (n)}= CV_{y (n-1)}+ Δt · V_y’_(n)

In this equation, V_{y (n-1)}In itself
Δt · V_y’_(n)Instead of adding_{y (n-1)}
Multiplied by C and Δt · V_y’_(n)Is added. This
Here, C is represented by the following equation: C = 1−Δt / T, and takes a value of 0 or more and 1 or less. Accordingly
And CV_{y (n-1)}Is V _{y (n-1)}Right after its acquisition
Reached when the detection period Δt elapses when proportionally attenuated
Means C · V_{y (n-1)}And Δt · V_y’
_(n)To add up the past information as it is
Rather than accumulating while attenuating.

Therefore, according to this detection method, for example,
If the true lateral velocity is 0,
If an error occurs, the virtual lateral speed becomes steady.
As a result, the detected value of the actual lateral velocity gradually attenuates and eventually becomes zero.
This causes the detection errors of each sensor to accumulate and diverge.
Evaded. That is, the lateral velocity derivative V_y’₍₀₎~ V_y’
_(n-1)Are all 0, V_{y (n)}Is Cⁿ・
V₀Since C is smaller than 1, V_{y (n)}Is 0
Because

However, there is a problem with the proposed detection method. Even when the true lateral speed is kept almost steadily at a certain value other than 0, the detected value of the lateral speed is attenuated. That is, in the proposed detection method, the steady component in the virtual lateral speed does not distinguish whether it is caused by the fact that the true lateral speed is kept at a steady value or a steady error of the sensor, This is because it is uniformly attenuated. When the slip angle is estimated using the lateral speed detected in this manner, the estimated slip angle is, for example, a steady value in which the actually measured slip angle is not 0 as shown in a graph as a second estimated slip angle in the upper part of FIG. Although it is within the period in which the value is maintained, it attenuates and eventually converges to zero. To solve this problem, the theoretical lateral velocity estimating means 52 is used.
And the actual lateral speed correcting means 54 is provided.

The theoretical lateral velocity estimating means 52, assuming a planar 2 DOF 2-wheel model for four-wheel vehicles, assuming the vehicle is always in steady circular turning state, the front and rear vehicle speed V _x and the actual steering angle δ estimating the lateral velocity V _y of the vehicle body under the current value as a theoretical lateral velocity V _yM. In particular,

[0044]

(Equation 1)

M: body mass a: distance between front wheel and center of gravity (fixed value) b: distance between rear wheel and center of gravity (fixed value) L: wheel base (= a + b) I: yaw moment of inertia of vehicle ( _Kf : Cornering power for two front wheels (fixed value) _Kr : Cornering power for two rear wheels (fixed value)

As described above, the theoretical lateral velocity V_yMIs no integration operation
It can be estimated by
There is an advantage that there is no worry. However, the theoretical lateral velocity V
_yMIs whether the vehicle is actually in a steady circular turning state, that is, the vehicle speed is constant.
Accuracy to true lateral speed only when the steering angle is constant
Good match, vehicle not in steady circular turning
In other words, when the vehicle is in a transient state,
The disadvantage is that the minutes do not match exactly. further,
In this embodiment, the cornering power K_f, K_rBut
Since it is a fixed value, the above equation is
Relationship with tire lateral force F, that is, tire cornering characteristics
Only if the characteristic is in the linear region,
After the transition, i.e. the tire has exceeded the grip limit
After that, the theoretical lateral velocity V_yMIs accurate enough for true lateral speed
There is also the disadvantage that they do not match. Note that the above equation is
Cornering power K_f,
K _rSo-called equivalent cornering power ／ F /
Although a method using ∂α is already known,
Dynamic characteristics of vehicle movements vary with lateral acceleration, etc.
For this reason, even in that method, the true horizontal
The speed cannot be detected with sufficient accuracy.

[0047] On the contrary, the actual lateral speed V _y of the foregoing, the vehicle unless detection errors of the sensors regardless of whether the steady circular turning state, also, the cornering characteristics of the tire in the linear region Irrespective of whether or not there is, there is an advantage that the value coincides with the true lateral speed with sufficient accuracy. However, if there is a detection error of each sensor, it accumulates and the detection accuracy decreases.

As described above, the actual lateral velocity V _y and the theoretical lateral velocity V
_In view of this fact, in this embodiment, the accuracy of estimating the final value of the actual lateral velocity V _y is compensated for by mutual compensation of one defect as the other advantage. In order to improve the lateral speed, the actual lateral speed correcting means 54 is provided.

The actual lateral speed correcting means 54 is connected to its input side.
The basic actual lateral velocity estimating means 50 and the theoretical lateral velocity estimating means 52
Are connected and input from the theoretical lateral speed estimation means 52.
Theoretical lateral velocity V_yMMeans 5 for estimating basic actual lateral speed based on
Actual lateral speed V input from 0 _yIs corrected.

In this correction, specifically, the current value of the corrected actual lateral velocity V _y (an example of the final actual lateral velocity) is:
It is calculated as the sum of the current value of the theoretical lateral velocity V _yM , the product of the previous value of the lateral velocity deviation ΔV _y and the coefficient C, and the product of the current value of the lateral velocity differential V _y ′ and the detection period Δt. That is, V _{y (n)} = V _{yM (n)} + C · ΔV _{y (n−1)} + Δt · V _y ′ _(n) where ΔV _{y (n−1) is} the previous value of the lateral velocity deviation (= V _{y (n-1)} -V
_{yM (n-1)} )

[0051] By estimating a slip angle using the actual lateral speed V _y of the corrected, estimated slip angle as represented in the graph as a third estimated slip angle on the lower side of FIG. 3, for example, found slip angle substantially Within the period of steady _maintenance , the attenuation decreases with time, but does not converge to 0, but converges to the theoretical slip angle estimated using the theoretical lateral velocity V _yM . That is, in the present embodiment, during the period in which the measured lateral speed (true lateral speed) is kept almost steady, the corrected actual lateral speed attenuates with time but converges to zero. However, since it converges to the theoretical lateral velocity, the lateral velocity detection accuracy is improved as compared to the case where it converges to 0,
As a result, the slip angle estimation accuracy is improved.

The vehicle body slip angle estimating means 56 is connected to the actual lateral speed correcting means 54 and the front / rear speed sensor 24.
It estimates the vehicle body slip angle β by dividing the longitudinal speed V _x input _y from longitudinal velocity sensor 24.

The differentiating means 58 is connected to the yaw rate sensor 22. The yaw rate γ is sequentially input from the yaw rate sensor 22 for each detection cycle Δt, and a value obtained by subtracting the previous value from the current value thereof is divided by the detection cycle Δt to calculate a yaw rate derivative γ ′ which is a time differential value of the yaw rate γ. I do. That is, γ ′ _(n) = (γ _(n) −γ _(n−1) ) / Δt

The tire lateral force estimating means 60 is connected to the differentiating means 58 and the lateral acceleration sensor 20, and based on the yaw rate differential γ 'and the lateral acceleration _Gy , the tire lateral force F generated for all four wheels. Is calculated. Specifically, F = (b · M · G _y + I · γ ′) / L, where b: distance between the center of gravity and rear wheel (fixed value) M: body mass (fixed value) I: yaw inertia of the vehicle Moment (fixed value) L: Wheel base (fixed value)

The road friction coefficient estimating means 62 is connected to the tire lateral force estimating means 60 and the vehicle body slip angle estimating means 56. A relationship is established between the slip angle β, the tire lateral force F, and the road surface μ, for example, as shown by a solid line graph in FIG. 4, where the tire lateral force increases as the road surface friction coefficient increases even at the same slip angle. . Road surface friction coefficient estimating means 6
2 estimates the current road surface μ based on the slip angle β and the tire lateral force F according to this relationship.

In a region where the slip angle β is sufficiently close to 0,
As is clear from the solid line graph in FIG. 7, the tire lateral force F does not differ even if the road surface μ differs, so that it is difficult to estimate the road surface μ with sufficient accuracy in this region. Therefore, the road surface friction coefficient estimating means 62 permits the estimation of the road surface μ in the good estimation region where the estimated slip angle β exceeds the threshold value β _th (an example of the slip angle set value). _The estimation of the road surface μ is prohibited in the estimation difficult region not exceeding _th , and the road surface μ is fixed to the last estimated value. The update of the estimated value of the road surface μ is prohibited. In the present embodiment, the threshold value β _th is set so as to increase with the estimated tire lateral force F, as shown by a broken line graph in the figure, but may be a fixed value.

However, the road surface friction coefficient estimating means 62 does not always permit the update of the road surface μ if it is in the good estimation region. After the estimated slip angle β has been maintained for a long time in the region equal to or larger than the threshold value β _th , although the actual lateral velocity _Vy has been corrected, as is clear from the lower graph of FIG. The error from the measured slip angle tends to be slightly larger. Therefore, in the present embodiment, as conceptually shown in the graph of FIG. 5, after the estimated slip angle β increases to reach the threshold value β _th and the estimation of the road surface μ is started, the estimated slip angle β _{Until the} time T ₁ that continues to exceed the threshold β _th exceeds the threshold T _1th (an example of the first set time), it is determined that the reliability R of the estimated value of the road surface μ is high.
_Is permitted (indicated by the prohibition flag OFF in the figure), but after exceeding the threshold value T _1th , it is determined that the reliability R of the estimated value of the road surface μ is low, and the update of the road surface μ is prohibited ( (Indicated by a prohibition flag ON in the figure).

The road friction coefficient estimating means 62
After the update of the road surface μ is prohibited as described above, the estimated slip angle β falls below the threshold value β _th , and thereafter, the update of the road surface μ is immediately permitted because the estimated slip angle β exceeds the threshold value β _th again. Do not. Threshold beta time T ₂ has elapsed from the time of transition _th from the state beyond the state below the threshold value beta _th threshold T _2th (an example of the second set time)
If the estimated slip angle β does not exceed the threshold value β _th , the update of the road surface μ continues to be inhibited even if the estimated slip angle β exceeds the threshold value β _th , and only after the time T ₂ exceeds the threshold value T _2th
The update of the road surface μ is permitted. The estimated slip angle β used for estimating the road surface μ in this period has a low reliability R that is strongly influenced by past detection values acquired in a period in which the reliability R is low, and such If the estimated slip angle β is used, the accuracy of estimating the road surface μ also decreases, so that the updating of the road surface μ during this period is also prohibited.

In order to execute the above contents, various routines, tables, functions, etc., including a road friction coefficient detection routine shown in the flowchart of FIG. 6, are stored in the ROM of the computer 30 in advance.

The routine shown in FIG. 6 is executed every detection cycle Δt (for example, about 10 ms). Prior to the first execution, initialization is performed, and a prohibition flag indicating that the update of the road surface μ is permitted in the OFF state and prohibiting the update in the ON state is set to OFF. Further, the time T ₁ during which the estimated slip angle β _{remains in} the region equal to or larger than the threshold value β _th.
And the time T ₂ that _{remains in} the region below the threshold β _th
Are both set to 0.

At each execution of this routine, first, in step S10 (hereinafter simply referred to as S10, and the same applies to other steps), the lateral acceleration G is obtained from various sensors.
_y, yaw rate gamma, each current value of the longitudinal speed V _x and the actual steering angle δ is input and stored in RAM. Subsequently, in S20, the lateral acceleration G _y of them, based on the yaw rate γ and longitudinal velocity V _x, the current value of the basic actual lateral speed V _y as described above is estimated.

Thereafter, in S30, the current value of the yaw rate differential γ 'is calculated by dividing the value obtained by subtracting the previous value from the current value of the yaw rate γ by the detection cycle Δt. Then, in S40, the yaw rate differential γ' is calculated. And the lateral acceleration G _y based on the tire lateral force F
Is estimated this time. Thereafter, in S50, based on the front and rear speed V _x and the actual steering angle [delta], the current value of the theoretical lateral velocity V _yM as the is estimated, then, in S60, based on the theoretical lateral velocity V _yM this value of the final actual lateral speed V _y is estimated by the basic actual lateral speed V _y is corrected Te.

Thereafter, in S70, the final actual lateral speed V
This value of the slip angle β are estimated by dividing _y by longitudinal speed V _x. Subsequently, in S80, T ₁ is:
In S90, both T ₂ are increased by the same value as the detection period Δt.

Thereafter, in S100, it is determined whether or not the current value of the estimated slip angle β has exceeded a threshold value β _th . Assuming that it has not exceeded this time, the determination is NO
Next, in S110, T ₁ is set to 0. T ₁ is incremented by independent detection period Δt is the relationship between the estimated slip angle beta and the threshold beta _th in S80 but returns to 0 as long as the estimated slip angle beta does not exceed the threshold beta _th in this S110 since the, eventually, T ₁ would represent the duration of the state where the estimated slip angle beta is greater than the threshold value beta _th. Then, in S120, the data indicating that the reliability R is low in this estimation value of the road surface μ is RA
M.

[0065] Subsequently, in S130, the current value of T ₁ is whether exceeds the threshold T _1th is determined. Since the current value of T ₁ is 0, the determination moves to NO next, S140. In this S140, the current value of T ₂ whether exceeds the threshold T _2th is determined. Assuming that it has not exceeded this time, the determination is NO, and one execution of this routine is immediately terminated.

[0066] Thereafter, if the current value of T ₂ within the execution of S10~140 is repeated many times exceeds the threshold T _2th, determination of S140 is OFF the prohibition flag YES, in S200 the command Is displayed, but since the prohibition flag is OFF this time, the state of the prohibition flag does not change by the current execution of S200.

Thereafter, if the current value of the estimated slip angle β exceeds the threshold value β _th while the execution of S10 to S140 is repeated many times, the determination in S100 becomes YES, and in S150, T ₂ Is set to 0. T ₂ is S
At 90, the detected slip angle β is increased by the detection period Δt irrespective of the relationship between the estimated slip angle β and the threshold value β _th. However, since the estimated slip angle β is returned to 0 as long as the estimated slip angle β exceeds the threshold value β _th at S150. Eventually, T ₂ represents the duration of the state where the estimated slip angle β does not exceed the threshold value β _th . Thereafter, in S160, it is determined whether or not the current prohibition flag is OFF. Since it is assumed to be OFF this time, the determination is YES, and in S170, the current value of the road surface μ is estimated based on the relationship between the estimated slip angle β and the tire lateral force F. Then, in S180, the estimate is the reliability R
Is stored in the RAM.

[0068] Then, in S130, the current value of T ₁ is whether exceeds the threshold T _1th is determined. Assuming the current value of T ₁ is not the still exceeds the threshold T _1th, determination of S130 shifts to NO next, S140. In S140, the current value of T ₂ is equal to the threshold value T _2th
If it is determined whether or not exceeded, this value of T ₂ are S15
Since 0 is set to 0, the determination is NO, and one cycle of this routine is immediately terminated.

Thereafter, S10-100, 150-18
If the current value of T ₁ within the execution of 0,130 and 140 is repeated many times exceeds the threshold T _1th, S
The determination at 130 is YES, and an instruction to turn on the prohibition flag is issued in S190. As a result,
Will be prohibited from being updated. This completes one execution of this routine.

Subsequently, if S10 to S100 and S150 are executed, the prohibition flag is ON this time, so that S16
If the determination of 0 is NO, S170 and S180 are skipped, and updating of the estimated value of the road surface μ is prohibited. As a result, the current value of the road surface μ is maintained at the previous value. Thereafter, in S120, data indicating that the reliability R of the current estimated value of the road surface μ is low is stored in the RAM. Thereafter, in S130, when it is determined whether the present value of T ₁ is greater than the threshold T _1th is, since the current value of T ₁ is assumed to have exceeded the threshold value T _1th, determination Is YES, and in S190, a command to turn on the prohibition flag is issued. Prohibition flag is already ON
Therefore, the state of the prohibition flag does not change by this execution of S190. This completes one execution of this routine.

Thereafter, S10-100, 150-18
Assuming 0,130 and estimated slip angle of which is also repeated execution times of 190 beta is equal to or less than the threshold value beta _th, the determination becomes NO S100, returned to T ₁ is 0 in S110, In S120, data indicating that the reliability R of the current value of the road surface μ is low is stored in the RAM. Subsequently, in S130, when it is determined whether the present value of T ₁ is greater than the threshold T _1th is, since the current value of T ₁ is zero at S110, the determination is NO, the S140 Transition. In this S140, T ₂
Of but this value whether exceeds the threshold T _2th is determined, assuming that the current value of T ₂ are not in yet exceed the threshold T _2th, a negative decision (NO) is obtained immediately the routine Is completed.

[0072] Then, assuming that the current value of T ₂ within the execution of S10~140 is repeated exceeds the threshold T _2th, determination is YES in S140, the instruction to turn OFF the prohibition flag in S200 Will be issued. The prohibition on updating the road surface μ is released. This completes one execution of this routine.

Therefore, if the estimated slip angle β exceeds the threshold value β _th thereafter, the prohibition flag is OFF, so that the determination in S160 becomes YES, and the update of the road surface μ is restarted in S170. Become.

The road surface μ estimated as described above is sequentially supplied to the antilock control device 32. The anti-lock control device 32 can change the brake pressure control characteristics such as the control start reference and the pressure reduction characteristics.
When it is determined that the road is a road, the brake pressure control characteristics are changed so that the antilock control is started earlier than when the road is determined to be a high μ road, or the pressure is continuously reduced for a longer time. It is designed to change the brake pressure control characteristics so that Anti-lock control is performed with brake pressure control characteristics that match the actual road
The design is such that the tire is not locked as much as possible on the road, and that the braking force is not wasted on a high μ road and the braking distance is not extended, thereby achieving both.

As is apparent from the above description, the part of the computer 30 that executes S10 and S20 in FIG. 6 is transmitted to the basic actual lateral speed estimating means 50 in FIG.
0 is executed by the theoretical lateral speed estimating means 52 in S60.
Corresponds to the actual lateral velocity correcting means 54, and the part for executing S30 corresponds to the differentiating means 58. And, among them, the basic actual lateral speed estimation means 5
0, theoretical lateral speed estimating means 52 and actual lateral speed correcting means 5
4 constitutes an example of the "processing means" in the first aspect of the present invention. 6 of the computer 30.
The step of executing S70 is performed by the slip angle estimating means,
The part executing S30 and S40 corresponds to the tire lateral force estimating means, and the part executing S80 to S200 corresponds to the road surface friction coefficient estimating means. In addition, the above basic actual lateral speed
Degree estimating means 50, theoretical lateral velocity estimating means 52 and actual lateral velocity
Each of the degree correction means 54 is
The actual lateral speed estimating means, theoretical lateral speed estimating means and actual lateral speed
This constitutes an example of the degree correction means.

[0076] In the embodiment described Shoki above, the final actual lateral speed V _y based only on both the theoretical lateral velocity V _yM not the actual lateral speed V _y in need of integration operations require integration operation value, but is adapted to be estimated, it is possible to estimate the final value of the actual lateral speed V _y also based on other variables. An example will be described below.

[0077] In previous embodiments, the present value of the actual lateral speed V _y of the corrected, and the current value V _yM theoretical lateral speed _(n), the previous value of the lateral speed deviation ΔV _{y (n-1) (} = _{Vy (n-1)} -V
_{yM (n-1)} ) and the coefficient C, and the sum of the product of the current value V _y ′ _{(n) of} the lateral velocity derivative and the detection period Δt. That is, V _{y (n)} = V _{yM (n)} + C · ΔV _{y (n−1)} + Δt · V _y ′ _(n)

On the other hand, in the present embodiment, the current value V _y ′ _(n) of the lateral velocity derivative is corrected using the value N calculated using a neural network (hereinafter simply referred to as a network). That is, the present value V _y ′ _{(n) of} the lateral velocity derivative is calculated using the equation V _y ′ _(n) = V _y ′ _(n) + N _(n) , and the actual lateral velocity V _y The correction is made.

The configuration of the network can be, for example, as follows. That is, as shown in the model in FIG. 7, each of the actual steering angle δ, the longitudinal velocity V _x , the lateral acceleration G _y , the yaw rate γ, the product of the yaw rate γ and the longitudinal velocity V _x , and the longitudinal acceleration G _x are input signals. It has an input layer input to each unit, an intermediate layer having a plurality of units, and one unit, and has an actual lateral velocity derivative V _y ′ and a theoretical lateral velocity derivative V
_This can be a hierarchical type including an output layer that outputs N representing an error from _yM ′ as an output signal. It should be noted that the number of input signals can be increased / decreased, the type can be changed, the number of intermediate layers can be plural, or a mutual coupling type can be used instead of the hierarchical type.

The learning of the connection weight (an example of the connection information) ω between the units, that is, the learning of the network, is performed by assuming the actual running state of the vehicle and presenting each input signal many times. Further, for example, as a teacher signal, the measured lateral speed differential value of (the true horizontal velocity) V _yT and the theoretical lateral velocity V _yM
Lateral speed differential V _y from the sum of the differential value of the 'minus, _{sand, V yT' (n) -V} y (n) = V yT '(n) - (γ (n) · V X _(n) −G _{y (n)} ) can be selected. In the learning stage, the connection weight ω of each unit is adjusted so that the output matches the teacher signal.

[0081] the lateral acceleration G _y which is used to estimate the vehicle lateral speed in theory a component in the direction perpendicular to the horizontal and the vehicle traveling direction with respect to the road surface of the acceleration acting on the vehicle. Therefore, the lateral acceleration sensor 20
The vehicle is fixedly mounted on the vehicle with a direction horizontal to the road surface and perpendicular to the traveling direction of the vehicle as a detection direction. But,
During the turning of the vehicle, the detection direction of the lateral acceleration sensor 20 is strictly not horizontal with respect to the road surface due to the presence of the roll angle φ of the vehicle body. The effect of gravity acting on the vehicle as well as _GyT will be affected. Further, during the turning of the vehicle body, the lateral acceleration sensor 2
The detected value of 0 is affected not only by the true lateral acceleration G _yT but also by the roll angular acceleration φ ″.

That is, during turning of the vehicle, a relationship represented by the following equation is established between the true lateral acceleration G _yT and the lateral acceleration G _y detected by the lateral acceleration sensor 20 (see FIG. 8). G _y = G _yT · cos φ + g · sin φ−h · φ ”where g: gravitational acceleration h: Roll arm that is the height of the vehicle center of gravity point from the roll axis

Here, considering the fact that φ is sufficiently close to 0, cos φ can be approximated to 1 and sin φ can be approximated to φ. Therefore, the above equation can be expressed as G _y = G _yT + g · φ−h · φ ”, and the true lateral acceleration G _yT is eventually described by the following equation: G _yT = G _y −g · φ + h · φ”.

[0084] Therefore, in order to improve the estimation accuracy of the lateral velocity V _y of the vehicle, it detects the lateral acceleration G _y instead of estimating the vehicle lateral velocity directly used, by adding the correction based on the above formula It is desirable to obtain the true lateral acceleration G _yT and use it to estimate the lateral velocity V _y .

In the above equation, it is necessary to obtain the roll angle φ and the roll angular acceleration φ ″. For this purpose, for example, the following method can be adopted.

That is, a sensor (for example, a vehicle height sensor) for directly or indirectly detecting the height from the road surface on each of the right side and the left side of the vehicle is provided, and the relative height between the right side and the left side is provided. It is possible to adopt a method of directly detecting the roll angle φ and the roll angular acceleration φ ″ from the relationship. In addition, a method of indirectly detecting the roll angle φ and the like using the equation of motion relating to the rolling motion of the vehicle body is also available. For example, the following equation can be adopted as the equation of motion: I · φ ″ + C · φ ′ + G · φ = _ms · _Gy · h where I: the sprung member of the vehicle Moment of inertia (fixed value) C: Roll damping (fixed value) G: Roll stiffness (fixed value) φ ': Roll angular velocity _ms : Mass of sprung member (fixed value) _Gy : Detected lateral acceleration h: Roll arm ( Fixed value)

This equation can be transformed into the following equation. φ ″ = 4π ² / _Tr ² · (G _a · G _y -φ) -φ '/ T
_g However, T _r: roll period (fixed value) G _a: gain (fixed value) T _g: roll damping time (fixed value)

[0088] Thus, by using the detected lateral acceleration G _y with discretized with respect to this expression time, can be detected sequentially indirectly roll angle phi and the roll angle acceleration phi ". That is, in this embodiment, The lateral acceleration sensor 20 is an example of the rolling motion state quantity sensor.

Although the inventions of claims 1 and 2 have been specifically described based on the illustrated embodiment, various modifications may be made based on the knowledge of those skilled in the art without departing from the scope of the claims. The inventions of claims 1 to 3 can be carried out in modified and improved embodiments.

[Brief description of the drawings]

FIG. 1 is a block diagram conceptually showing a configuration of a vehicle control device according to an embodiment of the present invention.

FIG. 2 is a block diagram conceptually showing an electrical configuration of a road surface friction coefficient detecting device 14 in FIG.

FIG. 3 is a graph for explaining a relationship between a temporal transition of an estimated slip angle by various estimation methods and an actually measured slip angle.

FIG. 4 is a graph for explaining a relationship between a slip angle β and a tire lateral force F used by the road surface friction coefficient detecting device 14 for estimating a road surface friction coefficient μ.

FIG. 5 is a graph for explaining the timing at which the road friction coefficient detection device 14 permits and prohibits the updating of the friction coefficient μ of the road surface in relation to the temporal transition of the estimated slip angle.

FIG. 6 is a flowchart showing a routine used by the road surface friction coefficient detecting device 14 to estimate a road surface friction coefficient μ.

FIG. 7 is a block diagram conceptually showing a flow of various signals in a vehicle lateral speed detecting device according to another embodiment of the first and second aspects of the present invention.

FIG. 8 is a front view for explaining how the detected lateral acceleration is affected by the roll angle of the vehicle body and the roll angular acceleration when the vehicle body rolls.

[Explanation of symbols]

 14 Road surface friction coefficient detecting device 20 Lateral acceleration sensor 22 Yaw rate sensor 24 Front / rear speed sensor 26 Actual steering angle sensor 30 Computer 32 Antilock control device

Claims (3)

(57) [Claims] 1. A lateral acceleration sensor for detecting a lateral acceleration at a center of gravity of a vehicle, a yaw rate sensor for detecting a yaw rate of the vehicle, a longitudinal speed sensor for detecting a longitudinal speed of the vehicle, and a rudder for detecting a steering angle of the vehicle. The basic value of the actual lateral speed of the vehicle is estimated by integrating the angle sensor and the value obtained by subtracting the detected value of the lateral acceleration from the product of the detected value of the longitudinal speed and the detected value of the yaw rate, and detecting the longitudinal speed. based on the detected value of the value and the steering angle, the vehicle is to estimate the theoretical lateral velocity of the vehicle when it is assumed to be in steady circular turning state, its
The basic value of the actual lateral speed is calculated according to the theoretical lateral speed estimated by
Processing means for estimating the final value of the actual lateral speed of the vehicle by correcting the vehicle lateral speed. 2. The processing means according to claim 1, The product of the detected value of the longitudinal speed and the detected value of the yaw rate
Integrate the value obtained by subtracting the detected value of lateral acceleration from time with respect to time
To estimate the basic value of the actual lateral speed of the vehicle
Actual lateral speed estimation means, Based on the detected value of the front-rear speed and the detected value of the steering angle,
Vehicle theory assuming the vehicle is in a steady circular turning state
Theoretical lateral velocity estimating means for estimating theoretical lateral velocity, Basic actual lateral speed estimated by basic actual lateral speed estimation means
The lateral acceleration sensor, the yaw rate sensor and the front and rear
Lateral velocity differentiation based on raw detection values acquired by velocity sensor
Lateral velocity obtained by simply integrating over time
The theoretical lateral velocity is estimated when the virtual lateral velocity
Decay toward the theoretical lateral velocity estimated by the means
The final value of the actual lateral speed.
Actual lateral speed correction means Claim characterized by including
Item 4. The vehicle lateral speed detection device according to Item 1. 3. The processing means according to claim 2, The product of the detected value of the longitudinal speed and the detected value of the yaw rate
Integrate the value obtained by subtracting the detected value of lateral acceleration from time with respect to time
To estimate the basic value of the actual lateral speed of the vehicle
Actual lateral speed estimation means, Based on the detected value of the front-rear speed and the detected value of the steering angle,
Vehicle theory assuming the vehicle is in a steady circular turning state
Theoretical lateral velocity estimating means for estimating theoretical lateral velocity, Basic actual lateral speed estimated by the basic actual lateral speed estimation means
Is estimated at the same time by theoretical lateral speed estimation means.
Corrected to the average value with the calculated theoretical lateral speed and final actual lateral speed
Actual lateral speed correction means Characterized by containing
The vehicle lateral speed detecting device according to claim 1.