METHOD OF ADAPTIVE ESTIMATION OF ADHESION COEFFICIENT OF VEHICLE ROAD SURFACE CONSIDERING COMPLEX EXCITATION CONDITIONS

Abstract

A method for adaptive estimation of a road surface adhesion coefficient for a vehicle with complex excitation conditions taken into consideration comprises the following steps: 1) designing an estimator according to a single-wheel dynamics model of a vehicle, and estimating a longitudinal tire force and a road surface peak adhesion coefficient under longitudinal excitation; 2) designing an estimator according to a two-degree-of-freedom kinematic model of the vehicle, and estimating a tire aligning moment and a road surface peak adhesion coefficient under excitation of a lateral force; and 3) determining an excitation condition met by the vehicle according to a vehicle state parameter, performing fuzzy inference to obtain limits achievable by current longitudinal and lateral tire forces, and designing a fusion observer to fuse estimation results. The method achieves favorable robustness, improves real-time capability, and can be performed quickly and accurately.

Claims

1. A method of an adaptive estimation of an adhesion coefficient of a vehicle road surface considering complex excitation conditions, the method comprising the following steps: 1) designing an estimator based on a single-wheel dynamical model of a whole vehicle, and estimating a peak adhesion coefficient of a road surface under a longitudinal tire force and a longitudinal excitation; 2) designing an estimator based on a two-degree-of-freedom kinematic model of the whole vehicle, and estimating the peak adhesion coefficient of the road surface under a tire aligning torque and a lateral force excitation; 3) determining excitation conditions met by a vehicle from vehicle state parameters, obtaining limits that current longitudinal and lateral tire forces can reach by a fuzzy inference, and thereby designing a fusion observer to fuse estimation results.

2. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 1, wherein in the step 1), the single-wheel dynamical model of the whole vehicle is as follows: $\stackrel{.}{\omega }=\frac{1}{I_w}\left[T_m-{\mu }_x\left({\theta }_x,\lambda \right).Math.F_z.Math.R\right]\lambda =\{\begin{array}{c}\frac{\omega .Math.R-v_x}{\omega .Math.R};v_x<\omega .Math.R\\ \frac{v_x-\omega .Math.R}{v_x};v_x\ge \omega .Math.R\end{array}$ wherein ω is an angular velocity of a wheel, {dot over (ω)} is an angular acceleration of the wheel, R is a radius of the wheel, T.sub.m is a driving/braking torque acting on the wheel, F.sub.z is a vertical load acting on the wheel, I.sub.w is a rotational inertia of the wheel, λ is a slip rate of the wheel, v.sub.x is a longitudinal speed at a center of the wheel, and μ.sub.x(θ.sub.x,λ) is a current utilization adhesion coefficient of a tire to the road surface obtained based on a tire model.

3. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 2, wherein an expression of the tire model is as follows: $\mu \left(\theta ,\lambda \right)=\theta -{\theta e}^{-\frac{c_1}{\theta }\left(\lambda +c_2{\lambda }^2\right)}-c_3\lambda \mathrm{sgn}\left(\lambda \right)+c_4{\lambda }^2$ wherein θ is the peak adhesion coefficient of the road surface, i.e., the peak adhesion coefficient of the road surface corresponding to a highest point of a μ−λ curve, c.sub.1 is a longitudinal slip stiffness of the tire, i.e., a slope of the μ−λ curve at an origin, and c.sub.2, c.sub.3, and c.sub.4 are respectively control parameters for a descending section of a curve of the peak adhesion coefficient of the road surface versus the slip rate.

4. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 2, wherein in the step 1), an expression for estimating the peak adhesion coefficient of the road surface under the longitudinal tire force and the longitudinal excitation is as follows: ${\hat{F}}_x=\frac{I_w}{R}\left(y+K\omega \right)-F_z.Math.{\mu }_x\left({\hat{\theta }}_x,\lambda \right)\stackrel{.}{y}=-\frac{K}{I_w}\left(T_m+R{\hat{F}}_x\right)+\frac{R}{I_w}.Math.\frac{\partial {\mu }_x\left({\hat{\theta }}_x,\lambda \right)}{\partial {\theta }_x}.Math.{\stackrel{\stackrel{.}{ˆ}}{\theta }}_x{\stackrel{\stackrel{.}{ˆ}}{\theta }}_x=\gamma \left[{\theta }_x\left(\lambda ,{\hat{F}}_x\right)-{\hat{\theta }}_x\right]$ wherein {circumflex over (F)}.sub.x is an estimated value of a tire longitudinal force, μ.sub.x({circumflex over (θ)}.sub.x,λ) is a utilization adhesion coefficient calculated based on an estimated value of the adhesion coefficient of the road surface and the slip rate, K is a gain of a longitudinal force estimator, θ.sub.x(λ,{circumflex over (F)}.sub.x) is the peak adhesion coefficient of the road surface calculated from a curve described by the tire model based on a current longitudinal force and the slip rate, {circumflex over (θ)}.sub.x is an estimated value of the peak adhesion coefficient of the road surface under the longitudinal excitation, γ is a gain of an adhesion coefficient estimator of the road surface, y is an intermediate variable, {dot over (y)} is a derivative of y with respect to time, and {circumflex over ({dot over (θ)})}, is a derivative of θ.sub.x with respect to time.

5. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 4, wherein in the step 2), the two-degree-of-freedom kinematic model of the whole vehicle is as follows: ${\alpha }_f=\beta +\frac{l_{f}R}{v_0}-\delta {\alpha }_r=\beta -\frac{l_{r}R}{v_0}$ wherein δ is a rotation angle of a front wheel, l.sub.f and l.sub.r are respectively a distance from a center of the front wheel and of a rear wheel to a center of mass, v.sub.0 is a longitudinal speed of the vehicle, β is a side slip angle of the vehicle, α.sub.f and α.sub.r are respectively a slip angle of the front wheel and of the rear wheel, and R is the radius of the wheel.

6. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 5, wherein in the step 2), an expression for estimating the peak adhesion coefficient of the road surface under the tire aligning torque and the lateral force excitation is as follows:
{circumflex over (M)}.sub.k=A{dot over (δ)}.sub.w+B{umlaut over (δ)}.sub.w+i.sub.s(δ.sub.w)M.sub.s+i.sub.m(δ.sub.w)M.sub.m
{circumflex over (M)}.sub.k=f(α,F.sub.z)
{circumflex over ({dot over (θ)})}.sub.y=k.sub.1 sgn({circumflex over (M)}.sub.k).Math.(M.sub.k−{circumflex over (M)}.sub.k)+k.sub.2 sgn(â.sub.y).Math.(a.sub.y−â.sub.y) wherein α is a slip angle of the wheel, δ.sub.w is a rotation angle of a steering wheel, i.sub.s(δ.sub.w) is a torque-to-rotation ratio of a booster motor to a master pin, i.sub.m(δ.sub.w) is a torque-to-rotation ratio of the steering wheel to the master pin, M.sub.m is a torque applied to the steering wheel, M.sub.s is a torque of the booster motor, A and B are fitting parameters, M.sub.k is a fitting total aligning torque, {circumflex over (M)}.sub.k is an estimated value of an aligning torque calculated based on the vertical load of the wheel and the slip angle of the wheel, F.sub.z is the vertical load applied on the wheel, â.sub.y is an estimated value of a lateral acceleration of the vehicle, a.sub.y is an actual value of the lateral acceleration of the vehicle, k and k.sub.2 are gains of estimators, {circumflex over (θ)}.sub.y is an estimated value of the peak adhesion coefficient of the road surface under the lateral force excitation, and {circumflex over ({dot over (θ)})}.sub.y is a derivative of {circumflex over (θ)}.sub.y with respect to time.

7. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 6, wherein the step 3) comprises: 31) obtaining a vehicle excitation state by the fuzzy inference; 32) performing the adaptive estimation of the peak adhesion coefficient of the road surface under a complex excitation.

8. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 7, wherein the step 31) is as follows: inputting a membership function, taking a slip rate reference λ/C.sub.λ and a slip angle reference α/C.sub.α as input quantities, wherein C.sub.λ and C.sub.α are catastrophe points at which tire characteristics enter a nonlinear zone and are respectively taken as a corresponding slip rate and the slip angle at which the peak adhesion coefficient is reached, and taking Ĉ.sub.1, Ĉ.sub.2 of different estimators as output quantities; setting [0,1] as a domain of both the input quantities and the output quantities; and dividing the domain into corresponding intervals respectively having small, medium and large fuzzy membership degrees.

9. The method of the adaptive estimation of the adhesion coefficient of the vehicle road surface considering complex excitation conditions according to claim 8, wherein in the step 32), an expression for performing the adaptive estimation of the peak adhesion coefficient of the road surface under a complex excitation is as follows:
{circumflex over ({dot over (θ)})}={circumflex over ({dot over (θ)})}.sub.x+{circumflex over ({dot over (θ)})}.sub.y
{circumflex over ({dot over (θ)})}.sub.x=γ[θ.sub.x(λ,{circumflex over (F)}.sub.x)−C.sub.1.Math.{circumflex over (θ)}]
{circumflex over ({dot over (θ)})}.sub.y=k.sub.1 sgn({circumflex over (M)}.sub.k).Math.(M.sub.k−Ĉ.sub.2{circumflex over (M)}.sub.k)+k.sub.2 sgn(â.sub.y).Math.(a.sub.y−Ĉ.sub.2â.sub.y) wherein Ĉ.sub.1 is a representative value of a longitudinal sliding degree of the wheel, Ĉ.sub.2 is a representative value of a side slip degree of the wheel, and {circumflex over (θ)} is an estimated value of the peak adhesion coefficient of the road surface.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0030] FIG. 1 is a flow chart of a method according to the present invention;

[0031] FIG. 2 is a schematic diagram of a single wheel dynamical model according to an embodiment;

[0032] FIG. 3 is a schematic diagram of a two-degree-of-freedom kinematic model of a whole vehicle according to an embodiment; and

[0033] FIG. 4 is a schematic diagram of estimation of an aligning torque according to an embodiment.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

EMBODIMENTS

[0035] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. Apparently, the described embodiments are some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all of other embodiments obtained by a person of ordinary skill in the art without any creative effort shall belong to the protection scope of the present invention.

EMBODIMENTS

[0036] As shown in FIG. 1, the present invention provides a method of adaptive estimation of an adhesion coefficient of a vehicle road surface considering complex excitation conditions, the method including the following steps:

[0037] Step 1, designing an estimator based on a single-wheel dynamical model, and estimating a peak adhesion coefficient of the road surface under a longitudinal tire force and longitudinal excitation. The process includes:

[0038] 1.1 establishing a single-wheel dynamical model of a whole vehicle.

[0039] First, obtaining a wheel angular velocity and a wheel slip rate:

[00005]$\dot{\omega }=\frac{1}{I_{\omega }}\left[T_m-{\mu }_x\left({\theta }_x,\lambda \right).Math.F_z.Math.R\right]\lambda =\{\begin{array}{c}\frac{\omega .Math.R-v_x}{\omega .Math.R};v_x<\omega .Math.R\\ \frac{v_x-\omega .Math.R}{v_x};v_x\ge \omega .Math.R\end{array}$

[0040] wherein ω is the angular velocity of the wheel, R is a radius of the wheel, T.sub.m is a driving/braking torque acting on the wheel, F.sub.z is a vertical load acting on the wheel, I.sub.ω is a rotational inertia of the wheel, λ is a slip rate of the wheel, v.sub.x is a longitudinal speed at a center of the wheel, and μ.sub.x (θ.sub.x,λ) is a current utilization adhesion coefficient of the tire to the road surface obtained based on a tire model;

[0041] Then, expressing the tire model as:

[00006]$\mu \left(\theta ,\lambda \right)=\theta -\theta e^{-\frac{c_1}{\theta }\left(\lambda +c_2{\lambda }^2\right)}-c_3\lambda \mathrm{sgn}\left(\lambda \right)+c_4{\lambda }^2$

[0042] wherein θ is the peak adhesion coefficient of the road surface, i.e., the peak adhesion coefficient of the road surface corresponding to a highest point of a μ−λ curve, λ is the slip rate of the wheel, c.sub.1 is a longitudinal slip stiffness of the tire, i.e., a slope of the μ−λ curve at an origin, and c.sub.2, c.sub.3, and c.sub.4 are respectively control parameters for a descending section of the curve of the peak adhesion coefficient of the road surface versus the slip rate.

[0043] 1.2 An expression of an estimation algorithm of the peak adhesion coefficient of the road surface under the longitudinal tire force and longitudinal excitation is as follows:

[00007]${\hat{F}}_x=\frac{I_w}{r}\left(y+K\omega \right)-F_z.Math.{\mu }_x\left({\hat{\theta }}_x,\lambda \right)\stackrel{.}{y}=-\frac{K}{I_w}\left(T_m+R{\hat{F}}_x\right)+\frac{R}{I_w}.Math.\frac{\partial {\mu }_x\left({\hat{\theta }}_x,\lambda \right)}{\partial {\theta }_x}.Math.{\stackrel{\stackrel{.}{ˆ}}{\theta }}_x{\stackrel{\stackrel{.}{ˆ}}{\theta }}_x=\gamma \left[{\theta }_x\left(\lambda ,{\hat{F}}_x\right)-{\hat{\theta }}_x\right]$

[0044] wherein {circumflex over (F)}.sub.x is an estimated value of the tire longitudinal force, μ.sub.x({circumflex over (θ)}.sub.x,λ) is the utilization adhesion coefficient calculated based on an estimated value of the adhesion coefficient of the road surface and the slip rate, K is a gain of a longitudinal force estimator, θ.sub.x(λ,{circumflex over (F)}.sub.x) is the peak adhesion coefficient of the road surface calculated by a numerical calculation method from the curve described by the tire model based on a current longitudinal force and slip rate, {circumflex over (θ)}.sub.x is an estimated value of the peak adhesion coefficient of the road surface under longitudinal excitation, and γ is a gain of an adhesion coefficient estimator of the road surface.

[0045] Step 2, designing an estimator based on a two-degree-of-freedom kinematic model of the whole vehicle, and estimating the peak adhesion coefficient of the road surface under a tire aligning torque and lateral force excitation. The process includes:

[0046] 2.1 Establishing the two-degree-of-freedom kinematic model of the whole vehicle.

[0047] Obtaining the slip angle of the wheel:

[00008]${\alpha }_f=\beta +\frac{l_{f}R}{v_0}-\delta {\alpha }_r=\beta -\frac{l_{r}R}{v_0}$

[0048] wherein δ is a rotation angle of a front wheel, l.sub.f and l.sub.r are respectively a distance from a center of the front wheel and of a rear wheel to a center of mass, v.sub.0 is a longitudinal speed of the vehicle, β is a side slip angle of the vehicle, and α.sub.f and α.sub.r are respectively a slip angle of the front wheel and of the rear wheel.

[0049] 2.2 The estimation algorithm of the adhesion coefficient of the road surface under longitudinal tire force and longitudinal excitation.

[0050] An expression is as follows:

[00009]${\hat{F}}_x=\frac{I_w}{r}\left(y+K\omega \right)-F_z.Math.{\mu }_x\left({\hat{\theta }}_x,\lambda \right)\stackrel{.}{y}=-\frac{K}{I_w}\left(T_m+R{\hat{F}}_x\right)+\frac{R}{I_w}.Math.\frac{\partial {\mu }_x\left({\hat{\theta }}_x,\lambda \right)}{\partial {\theta }_x}.Math.{\stackrel{.}{\stackrel{ˆ}{\theta }}}_x{\stackrel{\stackrel{.}{ˆ}}{\theta }}_x=\gamma \left[{\theta }_x\left(\lambda ,{\hat{F}}_x\right)-{\hat{\theta }}_x\right]$

[0051] wherein {circumflex over (F)}.sub.x is an estimated value of the tire longitudinal force, μ.sub.x({circumflex over (θ)}.sub.x,λ) is the utilization adhesion coefficient calculated based on an estimated value of the adhesion coefficient of the road surface and the slip rate, K is a gain of a longitudinal force estimator, θ.sub.x(λ,{circumflex over (F)}.sub.x) is the peak adhesion coefficient of the road surface calculated by a numerical calculation method from the curve described by the tire model based on a current longitudinal force and slip rate, {circumflex over (θ)}.sub.x is an estimated value of the peak adhesion coefficient of the road surface under longitudinal excitation, and γ is a gain of an adhesion coefficient estimator of the road surface.

[0052] Step 3, determining the excitation conditions met by the vehicle from vehicle state parameters, obtaining limits that the current longitudinal and lateral tire forces can reach by fuzzy inference, and thereby designing a fusion observer to fuse estimation results. The process includes:

[0053] 3.1 Fuzzy inference of vehicle excitation states.

[0054] Inputting a membership function, taking a slip rate reference λ/C.sub.λ and a slip angle reference α/C.sub.α as input quantities, wherein C.sub.λ and C.sub.α are catastrophe points at which tire characteristics enter a nonlinear zone and are respectively taken as the corresponding slip rate and slip angle at which the peak adhesion coefficient is reached, and both of the two items are obtained in real time through numerical calculation based on {circumflex over (θ)}; and taking Ĉ.sub.1, Ĉ.sub.2 of different estimators as output quantities. Setting [0,1] as a domain of both the input quantities and the output quantities; and dividing the domain into corresponding intervals respectively having S, M and B (respectively corresponding to small, medium and large) fuzzy membership degrees.

[0055] 3.2 An adaptive estimation algorithm of the peak adhesion coefficient of the road surface under complex excitations.

[0056] An expression is as follows:

{circumflex over ({dot over (θ)})}.sub.x=γ[θ.sub.x(λ,{circumflex over (F)}.sub.x)−C.sub.1.Math.{circumflex over (θ)}]

{circumflex over ({dot over (θ)})}.sub.y=k.sub.1 sgn({circumflex over (M)}.sub.k).Math.(M.sub.k−Ĉ.sub.2{circumflex over (M)}.sub.k)+k.sub.2 sgn(â.sub.y).Math.(a.sub.y−Ĉ.sub.2â.sub.y)

{circumflex over ({dot over (θ)})}={circumflex over ({dot over (θ)})}.sub.x+{circumflex over ({dot over (θ)})}.sub.y

[0057] A hardware device of the present invention requires sensors, including GPS, inertial elements and steering wheel rotation angle and torque sensors, and uses mass-produced electric controllers for the whole vehicle for data sampling, so as to implement on-line estimation by the algorithms designed in Steps 1 and 2. The fuzzy logic designed in Step 3 is burned into a controller in the form of a query table to obtain final fusion estimation results.

PARAMETER DESCRIPTION OF THE EMBODIMENTS

[0058] The superscript {circumflex over ( )} represents an estimated value, the superscript .Math. represents a first derivative, the subscript x represents a longitudinal direction, and the subscript y represents a lateral direction.

[0059] The above are merely specific embodiments of the present invention, however, the protection scope of the present invention is not limited thereto. Anyone who familiar with the technical field can easily conceive various equivalent modifications or substitutions within the technical scope revealed by the present invention. These modifications or substitutions should be included within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.