JP2008030616A - Running state estimation device, automobile and running state estimation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To estimate a sideslip angle easily and accurately by a simple structure. <P>SOLUTION: A yaw rate and a lateral wheel speed difference are normally in a one-to-one linear relationship, but the relationship varies when a sideslip angle develops. The sideslip angle β of a vehicle body is then estimated in accordance with the yaw rate γ of the vehicle body and the lateral wheel speed difference (V<SB>RW</SB>-V<SB>LW</SB>). <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a traveling state estimation device, an automobile equipped with the traveling state estimation device, and a traveling state estimation method.

Conventionally, the side slip angle is estimated by integration calculation using the vehicle speed, yaw rate, and lateral acceleration as parameters, and the side slip angle is estimated by a vehicle model using the state quantity of the vehicle as an input variable. There is one that estimates the angle and adjusts the time constant of the integral calculation in accordance with the side slip angle estimated from the vehicle model (see Patent Document 1).
JP-A-8-332934

However, in the conventional example described in Patent Document 1, it is determined whether or not the vehicle behavior is in the nonlinear region based on the side slip angle estimated from the vehicle model. And the side slip angle depend on the driving environment and the driving state, and therefore the final side slip angle estimation accuracy may be lowered due to the modeling error of the vehicle and the accumulated error of the integral calculation.
In addition, a sensor that detects lateral acceleration generally has problems such as noise easily, difficulty in adjusting to a strict zero point, and the influence of road surface inclination, which may lead to a decrease in estimation accuracy.
An object of the present invention is to estimate the side slip angle with high accuracy.

  In order to solve the above-described problem, a running state estimation device according to the present invention is characterized in that a side slip angle of a vehicle body is estimated according to a yaw rate of the vehicle body and a left and right wheel speed difference.

Normally, the yaw rate and the difference between the left and right wheel speeds are in a one-to-one linear relationship, but if a side slip angle is generated, this relationship is shifted.
According to the traveling state estimation device according to the present invention, focusing on the above characteristics, the side slip angle of the vehicle body is estimated according to the yaw rate of the vehicle body and the difference between the left and right wheel speeds. There is no estimation error due to accumulating errors, and the side slip angle can be estimated easily and with high accuracy with a simple configuration. In addition, since a sensor for detecting acceleration is not required, there is no estimation error unique to the acceleration sensor.

Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings.
<< One Embodiment >>
"Constitution"
FIG. 1 is a schematic configuration diagram of a vehicle. A vehicle 2 is equipped with a controller 2 composed of, for example, a microcomputer in order to estimate a side slip angle β formed between a longitudinal direction and a traveling direction of the vehicle body. .
The controller 2 includes a steering angle detected by the steering angle sensor 11, a wheel speed detected by the wheel speed sensor 12, a yaw rate detected by the yaw rate sensor 13, a lateral acceleration detected by the lateral acceleration sensor 14, and a longitudinal acceleration sensor. The longitudinal acceleration detected at 15 is input.

FIG. 2 is a block diagram of a linear two-input observer used for estimating the side slip angle β by the controller 2. A to D are matrices determined by a linear two-wheel model of the vehicle, and K1 is an observer gain. This observer is provided with an estimated compensator 20, and K2 is its compensation gain.
First, the observer will be described.
The two-wheel model of the vehicle on which the observer is based is expressed by the following equation based on the balance of lateral force and moment.

  If this is rewritten into the form of the state equation, the input u is the tire rudder angle, and the output of the output equation is the yaw rate and lateral acceleration, the following is obtained.

However, the meaning of each symbol is as follows.
m: vehicle mass I: yaw moment of inertia Lf: distance between the vehicle center of gravity and a front axle Lr: distance between the vehicle center of gravity and a rear axle Cp _f: front wheel cornering power (left total)
Cp _r: rear wheel cornering power (left and right total)
V: Vehicle speed β: Side slip angle γ: Yaw rate Gy: Lateral acceleration a ₁₁ , a ₁₂ , b ₁ : Elements of matrices A and B Based on this equation of state, a linear two-input observer that estimates the vehicle side slip angle β is designed. To do. The input to the observer is the lateral acceleration and the yaw rate, and the observer gain K is set so that it is not easily affected by modeling errors and can be stably estimated. The observer design method is not limited to this, and may be replaced with a completely different estimation method.

Next, the estimation compensator 20 will be described.
FIG. 3 shows a model in which the vehicle is a material point turning at a yaw rate γ and a radius R, and has wheels at left and right L _L and L _R distances. Assuming that the left and right wheel speeds are V _L and V _R , the yaw rate γ and the left and right wheel speed difference (V _R −V _L ) usually have a proportional relationship as shown below. Here, the difference between the left and right wheel speeds is for the driven wheel, but the front wheel and the rear wheel having the higher grip state may be selected.
V _R = (R + L _R ) γ
V _L = (R−L _L ) γ
→ V _R −V _L = (L _R + L _L ) γ

When the side slip angle β is generated in the vehicle body, it rotates while revolving, but the balance of movement is not affected by the side slip angular velocity β ′ as shown below.
V _R = (R + L _R ) × (γ + β ′) − β ′ × L _R
V _L = (R−L _L ) × (γ + β ′) + β ′ × L _L
→ V _R -V _L = (L _R + L _L ) × (γ + β ′) − (L _R + L _L ) × β ′
→ V _R −V _L = (L _R + L _L ) × γ

Further, after the side slip angle β is generated, as shown in FIG. 4, L _L and L _R are shortened by cos β with respect to the turning radius direction, and the balance of motion is expressed by the following equation.
V _R −V _L = {(L _R + L _L ) × cos β} γ
However, since get sideslip angle β is in the tire, unlike the circumferential velocity V _LW and V _RW of V _L and V _R and the wheel, as shown below, it is necessary to correct.
V _RW = V _R × cosβ
V _LW = V _L × cosβ
Therefore, the relationship of the following equation is established among the side slip angle, the yaw rate, and the difference between the left and right wheel speeds.
cos ² β = (V _RW −V _LW ) / {γ × (L _R + L _L )}

The deviation between the side slip angle β estimated according to the above equation and the side slip angle β calculated by the observer is fed back before the integrator in the observer. The feedback gain K2 for applying this correction is calculated according to the ratio ΔV / γ between the left and right wheel speed difference ΔV (= V _RW −V _LW ) and the yaw rate γ with reference to the control map of FIG. This control map is set such that ΔV / γ on the horizontal axis and gain K2 on the vertical axis, and the gain K2 increases as ΔV / γ increases. However, dead zones and limits are provided.

Next, a description will be given of the traveling control that uses the estimated side slip angle β to suppress the unstable behavior of the vehicle when it is detected.
First, as shown below, tire slip angles β _f and β _r of the front and rear wheels are calculated according to the side slip angle β. Each state quantity is an absolute value and is based on the state quantity during straight traveling.
βf = −β−L _f × γ / V + δ
βr = −β + L _r × γ

On the other hand, as shown in FIG. 8, a Kalman filter is constructed from the equation of motion of the vehicle, and the lateral forces F _f and F _r of the front and rear wheels are estimated according to the yaw rate γ and the lateral acceleration Gy. Here, assuming that the disturbance dispersion matrix is Q and the observation noise dispersion matrix is R, the Kalman filter gain K can be calculated by solving the Riccati equation:
^{PA T + AP-PC T R} -1 CP + Q = 0
K = PC ^{TR- 1}
Then, as shown below, the tire cornering powers K _f and K _r of the front and rear wheels are calculated from the tire slip angle and the lateral force.

  However, since it cannot be calculated as it is, it is converted into the following equation. Here, the time differentiation may be a simple difference, but it is desirable to take a sampling time as long as possible to increase the resolution of the change amount.

Then, ξ = L _r K _r −L _f K _f is calculated from K _f and K _r , and this is used as an index of the degree of instability, and the direction is determined from the yaw rate γ and the side slip angle β.
The yaw rate γ is set to 0 when stopped. The lateral acceleration Gy is 0 when γ = 0, but because it is affected by road surface cant, etc., the steering angular velocity is below a predetermined value and the value when the yaw rate = 0 is statistically histogrammed. The average value, median value, or mode value may be used.

When an unstable state is detected, the vehicle behavior is controlled in a stable direction using traction / brake control or steering reaction force control. Here, the brake control will be described as an example.
First, a predetermined threshold is set in advance for ξ (= L _r K _r −L _f K _f ). This is a threshold value that serves as a trigger for starting the unstable behavior suppression control, and is set for each of the OS direction and the US direction for each vehicle.

In general, the initial value of the vehicle characteristic often has a weak US tendency. In this case, the OS-side threshold value may be a value in the US region as a result. This is because although it depends on the responsiveness of the control, it is desired to suppress this before the OS tendency is actually reached.
In any case, the brake control is started when ξ exceeds the threshold, and the brake control is continued until the region returns to the region (stable running region) between the OS side threshold and the US side threshold. That is, when the OS-side threshold is exceeded, braking force is applied to the wheels on the outer front turning side, and when the US side threshold is exceeded, braking force is applied to the inner wheels on the rear turning side.

At this time, in order to quickly suppress unstable behavior, the magnitude of the braking force is set to the maximum value within a range where the wheel does not lock, or the pressure is increased to the maximum braking force that can be generated by the braking mechanism. Each time a lock tendency is detected, pressure reduction and pressure holding are repeated.
By performing the above control, the unstable behavior of the vehicle such as spin and drift out is quickly suppressed. FIG. 9 is a flowchart thereof.

<Action>
Next, the operation of the embodiment will be described.
A general linear observer can accurately estimate the side slip angle β in a linear region where the tire side slip angle does not become a nonlinear characteristic under the road surface condition assumed at the time of vehicle model design. However, when the road surface friction coefficient μ changes or approaches the limit of turning performance, the estimated value by the observer gradually deviates from the measured value, and the estimation accuracy decreases.
In addition, a sensor that detects lateral acceleration generally has problems such as noise easily, difficulty in adjusting to a strict zero point, and the influence of road surface inclination, which may lead to a decrease in estimation accuracy.

Therefore, in this embodiment, the yaw rate and the difference between the left and right wheel speeds are usually in a one-to-one linear relationship. However, when a side slip angle is generated, paying attention to the characteristic that a deviation occurs in this relationship, The side slip angle of the vehicle body is estimated according to the wheel speed difference. That is, the side slip angle β is estimated by solving cos ² β according to the above-described formula. As a result, there is no estimation error due to vehicle modeling errors or accumulation of errors due to integration calculations, and the side slip angle can be estimated easily and with high accuracy with a simple configuration. Further, if only the estimated compensator 20 is considered, a sensor for detecting acceleration is not required, and the side slip angle β estimated by the estimated compensator 20 has no estimation error peculiar to the acceleration sensor.

FIG. 6 shows actual vehicle data at the time of acceleration circle turning.
FIG. 7 shows the relationship between the wheel speed difference of the driven wheel and the yaw rate converted value at the same time as in FIG. 6, and it can be seen that the deviation in linearity of both corresponds to the occurrence of the side slip angle β.
In this way, by performing feedback compensation for the observer according to the deviation between the side slip angle β calculated by the observer and the side slip angle β estimated by the estimation compensator 20, estimation by the observer and feedback compensation by the estimation compensator 20 are performed. And make use of the advantages of both. As a result, estimation errors caused by accumulation of errors due to integration calculations are corrected as needed, so the estimated value approximates the measured value, and the system as a whole can estimate the highly accurate side slip angle β that is resistant to environmental changes. it can.

  However, in order for the estimated compensator 20 to operate normally, the wheels need to rotate while gripping the road surface. For example, if the rotation is disturbed by braking force or driving force, the estimated accuracy Will fall. Therefore, as the slip ratio increases during braking or driving, the difference between the left and right wheel speeds approaches 0 and the ΔV / γ ratio decreases. Therefore, the feedback gain K2 decreases as the ΔV / γ ratio decreases, and the estimation accuracy increases. To prevent the loss of

《Application example》
In the above embodiment, the value of the feedback gain K2 is changed according to the ΔV / γ ratio, but the slip ratio may be calculated and the feedback gain K2 may be changed accordingly. According to this, the weight of feedback compensation can be changed more accurately.
In the above-described embodiment, the side slip angle β is estimated by the observer, the side slip angle β is calculated by the estimation compensator 20, and feedback is provided for the integration operation by the integrator 21 of the observer according to the deviation between the two. Although compensation is performed, the present invention is not limited to this. That is, since the side slip angle β can be estimated only by the estimation compensator 20, this β may be directly estimated as the side slip angle β.

  In the above-described embodiment, the side slip angle β is merely estimated, but the acceleration Ay in the turning radius direction may be estimated using the side slip angle β, the yaw rate γ, and the vehicle speed V. . According to this, since an estimation error peculiar to the acceleration sensor does not occur, it is possible to estimate the acceleration Ay with high accuracy. Note that the acceleration Ay in the turning radius direction is equivalent to the detection value of the lateral acceleration sensor during steady circle turning on a flat road. Here, the above processing corresponds to “Ay estimating means”.

Furthermore, the road surface gradient θ may be estimated according to the side slip angle β, the acceleration Ay in the turning radius direction, the acceleration in the longitudinal direction of the vehicle body, and the vehicle speed change amount. First, the acceleration in each direction is calculated by subtracting the change in the vehicle speed from the acceleration (sensor value) in the front-rear direction and the left-right direction. Then, as shown below, the acceleration Ay in the turning radius direction is decomposed into a front-rear direction component and a left-right direction component, and a change in vehicle speed is subtracted from each. In this way, the amount of acceleration generated by gravity is extracted from the difference between the sensor value considering the change in vehicle speed and the decomposition component of Ay, and the ratio between this and the gravitational acceleration (≈9.8 m / s ² ) is taken. The road surface gradient θ in the road longitudinal direction and the road crossing direction can be estimated. According to this, the road surface gradient θ in the direction detected by the acceleration sensor can be estimated with high accuracy. Here, the above processing corresponds to “θ estimation means”.
Longitudinal component ← Ay × sinβ
Horizontal component ← Ay × cosβ

"effect"
As described above, the estimation compensator 20 corresponds to the “β estimation unit”, the observer in FIG. 2 corresponds to the “β calculation unit”, and feedback according to the deviation between the observer calculation result and the estimation compensator 20 and the estimation result. The compensation process corresponds to “compensation means”.
(1) β estimation means for estimating the side slip angle of the vehicle body according to the yaw rate of the vehicle body and the difference between the left and right wheel speeds is provided.
As a result, there is no estimation error due to vehicle modeling errors or accumulation of errors due to integration calculations, and the side slip angle can be estimated easily and with high accuracy with a simple configuration. In addition, since a sensor for detecting acceleration is not required, there is no estimation error unique to the acceleration sensor.

(2) The β estimating means estimates the side slip angle according to the following mathematical formula.
cos ² β = (V _RW −V _LW ) / {γ (L _R + L _L )}
β: Side slip angle V _RW −V _LW : Left and right wheel speed difference γ: Yaw rate L _R + L _L : Distance between left and right wheels Thereby, the above-described effect can be obtained with certainty.
(3) Through an integration calculation with the running state as an auxiliary variable, according to the β calculation means for calculating the side slip angle of the vehicle body, and the deviation between the side slip angle calculated by the β calculation means and the side slip angle estimated by the β estimation means And compensation means for performing feedback compensation for the β calculation means.
As a result, it is possible to estimate the highly accurate side slip angle β that is resistant to environmental changes as a whole system.

(4) The compensation means changes the weight of feedback compensation according to the slip tendency of the left and right wheels.
Thereby, the fall of estimation accuracy can be prevented.
(5) Ay estimating means for estimating the acceleration in the turning radius direction according to the side slip angle estimated by the β estimating means, the yaw rate of the vehicle body, and the vehicle speed change amount is provided.
According to this, since an estimation error peculiar to the acceleration sensor does not occur, it is possible to estimate the acceleration Ay with high accuracy.
(6) θ estimating means for estimating a road surface gradient in a predetermined direction of the vehicle according to the side slip angle estimated by the β estimating means, the acceleration in the turning radius direction estimated by the Ay estimating means, the acceleration in the predetermined direction of the vehicle body, and the amount of change in the vehicle speed. Is provided.
According to this, the road surface gradient θ in the direction detected by the acceleration sensor can be estimated with high accuracy.

1 is a schematic configuration diagram of a vehicle. This is an observer for estimating the side slip angle β. It is a geometric model of yaw rate and wheel speed difference. This is a geometric model when a side slip angle β occurs. It is a control map for gain calculation. It is actual vehicle data at the time of acceleration circle turning. This is a comparison between wheel speed difference and yaw rate. It is a Kalman filter that estimates lateral force. It is a flowchart which shows an unstable behavior suppression process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Car 2 Controller 11 Rudder angle sensor 12 Wheel speed sensor 13 Yaw rate sensor 14 Lateral acceleration sensor 15 Longitudinal acceleration sensor 20 Estimation compensator

Claims (8)

  A travel state estimation device comprising β estimation means for estimating a side slip angle of a vehicle body according to a yaw rate of the vehicle body and a difference between left and right wheel speeds. The travel state estimation apparatus according to claim 1, wherein the β estimation means estimates a side slip angle according to the following mathematical formula.
cos ² β = (V _RW −V _LW ) / {γ (L _R + L _L )}
β: Side slip angle V _RW −V _LW : Left and right wheel speed difference γ: Yaw rate L _R + L _L : Distance between left and right wheels   Through an integral calculation with the running state as an auxiliary variable, β calculation means for calculating the side slip angle of the vehicle body, and according to the deviation between the side slip angle calculated by the β calculation means and the side slip angle estimated by the β estimation means, The traveling state estimation apparatus according to claim 1, further comprising: a compensation unit that performs feedback compensation on the β calculation unit.   The travel state estimation apparatus according to claim 3, wherein the compensation means changes a weight of the feedback compensation according to a slip tendency of the left and right wheels.   The Ay estimating means for estimating the acceleration in the turning radius direction according to the side slip angle estimated by the β estimating means, the yaw rate of the vehicle body, and the amount of change in the vehicle speed is provided. The travel state estimation device according to 1.   Θ estimating means for estimating a road surface gradient in the predetermined direction of the vehicle body according to the side slip angle estimated by the β estimating means, the acceleration in the turning radius direction estimated by the Ay estimating means, the acceleration in the predetermined direction of the vehicle body, and the amount of change in vehicle speed. The travel state estimation device according to claim 5, comprising: In a car equipped with a running state estimation device,
The vehicle is characterized in that the running state estimation device estimates a side slip angle of a vehicle body according to a yaw rate of the vehicle body and a difference between left and right wheel speeds.   A running state estimation method characterized by estimating a side slip angle of a vehicle body according to a yaw rate of the vehicle body and a difference between left and right wheel speeds.

Images