CN103879401A - Coordinated control technology capable of effectively controlling automobile steering/braking system - Google Patents

Abstract

The invention relates to a coordinated control technology capable of effectively controlling automobile steering/braking system, aiming to solve the problems that an existing automobile steering/braking system is not subjected to coordination control through subsystems, no consideration is given to laterodeviation and vertical load parameters of tires, nonlinear uncertain changes of ground adhesion and coupling interference between the steering/braking system, and force closed-loop control is omitted. The steering/braking consistency coordinated control system is designed by defining coordinated errors. The problems about uncertainty changes of tire parameters and coupling interference can be well solved through development designed on the basis of events, nonlinear robust self-adaption and a force and position double closed-loop control technique. Braking force of wheels is well allocated, so that uniformity of tracking errors can rapidly decay to zero. An automobile can be always steered and braked according to expected yaw velocity, laterodeviation and sideslip never occur, and steering capability never loses; the coordinated control technology is good in the robust self-adaption and is applicable to various dynamic stability control system of different types of automobiles.

Description

The strong motor turning of controlling and the consistent Collaborative Control technology of brake system

Technical field: the present invention relates to the unify Collaborative Control technology of brake system of a kind of new automotive steering.Adopt based on the two Closed loop Control in event design, Robust Adaptive Control technology, power and position.Belong to automobile technical field.

Background technology: the control of existing motor turning/ABS (Anti-lock Braking System) does not have consistent Collaborative Control, lacks the total tune to subsystems.Can not distribute the lock torque of each wheel, each wheel braking force unbalance.Automobile usually produces lateral deviation moment of torsion, makes vehicle body occur to break away and loses insensitivity.The nonlinear uncertain that braking force and side direction adhesive ability are longitudinally adhered in cornering behavior parameter, vertical load parameter, the ground of not considering tire changes, can not high precision control.What turn to/braking control system of existing design adopted is that two degrees of freedom linearity turns to definite model, single wheel brake model and tire linearization to determine model.Only have position or speed to follow the tracks of and control, there is no power closed loop control.Do not consider motor tire strong nonlinearity uncertainty, do not study and consider to act on the interrelation between braking force and the longitudinal adhesive ability of tire, side direction adhesive ability and the yaw moment on tire.Turning efficiency and the deceleration and stopping performance impact of these power relations on automobile is very large.

Summary of the invention: the object of the invention is coupled interference and uncertain parameter perturbation problem for motor tire strong nonlinearity, steering swivel system and brake system, design real-time is good, robust adaptive is good, the two closed loops in strong and position of high conformity turn to/brake non-linear consistent cooperative control system based on event design vehicle.

The present invention is divided into the non-linear consistent cooperative control system of two turn to/ABS (Anti-lock Braking System) of step design vehicle.

One, the motor turning of first step research and design based on event design/brake consistent Collaborative Control algorithm, the good non linear robust adaptive control system of design real-time.

Motor turning/brake consistent cooperative control system as shown in Figure 1.β _d, γ _dby expecting turning track

obtain.Wherein to be that true value is determined known for each parameter, δ _dit is the deflection angle that is input to front-wheel by steering handwheel.Automobile in figure in braking procedure is expected car load vehicle velocity V _dchanging Pattern is by based on event design calculation out.

1, based on event design calculation V _d

First choose motion reference variable s and be defined as the stopping distance of passing by automobile barycenter place.

The model of each wheel on time t territory can be write as:

$\frac{\mathrm{ds}}{\mathrm{dt}}=v,$ $\frac{\mathrm{dv}}{\mathrm{dt}}=a,$

By converting acquisition speed and acceleration/accel about the function expression v=V (s) in distance domain, a=A (s).Definition:

W=v ²and $u=\frac{\mathrm{da}}{\mathrm{ds}},$ Can obtain: $\frac{\mathrm{dw}}{\mathrm{ds}}=2a,$ $\frac{\mathrm{da}}{\mathrm{ds}}=u$

Wherein above formula meets following constraint: | w|≤w _lim, | a|≤a _lim, | u|≤u _lim.

Control based on event knows that fundamental point is exactly to seek to meet the v=V of above-mentioned constraint (s), a=A (s).In braking procedure control, performance index function is to make stopping distance the shortest.That is:

$J=\underset{s_0}{\overset{s}{\∫}}\mathrm{ds}\→\min .$

Wherein start braking, s (0)=s when t=0 ₀.T=t _ftime braking stops, s (t _f)=s.

In the time using control technology based on event to problem solving, be defined as follows:

x ₁＝w，x ₂＝a，c ₁＝x ₁-w _lim，c ₂＝-x ₁-w _lim，c ₃＝x ₂-a _lim，c ₄＝-x ₂-a _lim

The equation of state that can draw system is:

$\stackrel{\·}{X}=\mathrm{AX}+\mathrm{Bu}$

Wherein: $X=\left[\begin{array}{c}{x}_1\\ x_2\end{array}\right],$ $A=\left[\begin{array}{cc}0& 0\\ 0& 2\end{array}\right],$ $B=\left[\begin{array}{c}0\\ 1\end{array}\right],$ $\stackrel{\·}{X}=\frac{\mathrm{dX}}{\mathrm{ds}},$ C=[ _c1 c ₂c ₃c ₄] ^tand meet constraint C≤0.Initial condition (IC): X (s ₀)=X ₀.Braking terminal: X (s)=0.By greatly-minimal principle, problem is solved.Hamiltonian function about s is:

$H=1+{\mathrm{\λ}}^T(\mathrm{AX}+\mathrm{Bu})+{\mathrm{\μ}}_1{\stackrel{\·\·}{c}}_1+{\mathrm{\μ}}_2{\stackrel{\·\·}{c}}_2+{\mathrm{\μ}}_3{\stackrel{\·\·}{c}}_3+{\mathrm{\μ}}_4{\stackrel{\·\·}{c}}_4$

Wherein: λ=[λ ₁λ ₂] ^t, ${\stackrel{\·\·}{c}}_1=\frac{d^2{c}_1}{{\mathrm{ds}}^2}=2u,$ $\frac{d^2{c}_2}{{\mathrm{ds}}^2}=-2u,$ ${\stackrel{\·}{c}}_3=\frac{{\mathrm{dc}}_3}{\mathrm{ds}}=u,$ ${\stackrel{\·}{c}}_4=\frac{{\mathrm{dc}}_4}{\mathrm{ds}}=-u$

${\mathrm{\&mu;}}_1=\left\{\begin{array}{cc}0,& \mathrm{if}{x}_1<w_{\lim }\\ >0,& \mathrm{if}{x}_1=w_{\lim }\end{array}\right.,$ ${\mathrm{\&mu;}}_2=\left\{\begin{array}{cc}0,& {\mathrm{ifx}}_1>-w_{\lim }\\ >0,& \mathrm{if}{x}_1=-w_{\lim }\end{array}\right.,$

${\mathrm{\&mu;}}_3=\left\{\begin{array}{cc}0,& \mathrm{if}{x}_2<a_{\lim }\\ >0,& \mathrm{if}{x}_2=a_{\lim }\end{array}\right.,$ ${\mathrm{\&mu;}}_4=\left\{\begin{array}{cc}0,& {\mathrm{ifx}}_2>-a_{\lim }\\ >0,& \mathrm{if}{x}_2=-a_{\lim }\end{array}\right.$

Optimal solution is:

$u=\left\{\begin{array}{c}-u_{\lim },s_0<s<s_1\\ 0,s_1<s<s_2\\ u_{\lim },s_2<s\end{array}\right.,$ $a=\left\{\begin{array}{c}{-u}_{\lim }s+u_{\lim }{s}_0,s_0<s<s_1\\ -a_{\lim },s_1<s<s_2\\ u_{\lim }s-u_{\lim }{s}_2,s_2<s\end{array}\right.,$

$w=\left\{\begin{array}{c}{-u}_{\lim }{s}^2+{2u}_{\lim }{s}_{0}s+w_{\lim }-u_{\lim }{s}_0^2,s_0<s<s_1\\ {-a}_{\lim }s+{2a}_{\lim }{s}_1-u_{\lim }{s}_1^2+{2u}_{\lim }{s}_0{s}_1^2+w_{\lim }-w_{\lim }{s}_0^2,s_1<s<s_2\\ u_{\lim }{s}^2-{2u}_{\lim }{s}_{2}s-w_{\lim }+u_{\lim }{s}_2^2,s_2<s\end{array}\right.$

Wherein: s ₀the distance that while being braking, wheel has been passed by,

Wheel velocity v can be by

formula is obtained.Concerning each wheel of vehicle, as long as measure in real time formula above the distance s substitution that wheel passes by and can obtain the desired speed V of each wheel _lf, V _lr, V _rf, V _rr.Concerning car load, as long as measure in real time formula above the distance s substitution that barycenter passes by and can obtain the desired speed V of car load _d.V _dcan regard the functional of s (t) as.

2, the consistent cooperative control system design of Vehicle Nonlinear robust adaptive

Motor turning/braking is controlled and is subject to Wheel slip parameter k ₁, k ₂longitudinally adhere to braking force F _xithe factor such as nonlinear uncertain variation, coupled interference affect greatly.Disturb the consistent collaborative controller of design non linear robust self adaptation in order to compensate these.The good steering swivel system of Collaborative Control and brake system, accurate tracking turning track, the desired braking speed of a motor vehicle.Allow tracking error conformability rapidly decay to zero.

Specific design process is as follows:

Consider two front-wheel steering car models, the comprehensive non-linear dynamic model of automobile three degree of freedom automobile is as follows:

Vehicle complete vehicle braking longitudinal force equation of equilibrium is:

$m(\stackrel{\&CenterDot;}{v}-\mathrm{\&gamma;\&beta;v})=-((F_{x1}+F_{x2})\cos \mathrm{\&delta;}+(F_{x3}+F_{x4})+(F_{y1}+F_{y2})\sin \mathrm{\&delta;})---\left(1\right)$

The transverse mechanical equation of equilibrium of car load is:

$\mathrm{mv}(\stackrel{\&CenterDot;}{\mathrm{\&beta;}}+\mathrm{\&gamma;})=(F_{x1}+F_{x2})\sin \mathrm{\&delta;}+(F_{y1}+F_{y2})\cos \mathrm{\&delta;}+(F_{y3}+F_{y4})---\left(2\right)$

The weaving mechanical balance equation of car load is:

$I\stackrel{\&CenterDot;}{\mathrm{\&gamma;}}=\frac{w}{2}(F_{x1}-F_{x2})\cos \mathrm{\&delta;}+a(F_{x1}+F_{x2})\sin \mathrm{\&delta;}+\frac{w}{2}(F_{y1}-F_{y2})\sin \mathrm{\&delta;}+a(F_{y1}+F_{y2})\cos \mathrm{\&delta;}+\frac{w}{2}(F_{x4}-F_{x3})-b(F_{y3}+F_{4r})---\left(3\right)$

In formula, m is complete vehicle quality.F _xifor being subject to ground, each wheel longitudinally adheres to braking force, F _yifor the side direction adhesive ability that is subject to ground of each wheel.Subscript i=1,2,3,4 represent left front, right front, left back, right back wheel.V is the speed at automobile barycenter place.δ is front-wheel active steering angle.Automobile side slip angle β.γ is vehicle complete vehicle yaw velocity.A be barycenter to front wheel spindle distance, b is that barycenter is to hind axle distance.W is the wheelspan of lateral direction of car left and right two wheels, and I is car load rotation around center of mass inertia.

Here longitudinal adhesive ability F of single tire _xibe treated to nonlinear uncertain variable quantity.And side direction adhesive ability F _yidescribe the friction force between tire and ground, there is strong nonlinear characteristic.It has determined the dynamic property of vehicle body and wheel to adopt non-linear fraction tire model ^[57]tyre slip angle and the tyre side characteristic to adhesive ability is described:

$F_{\mathrm{yi}}\&ap;\frac{\mathrm{\&mu;}{F}_{\mathrm{zi}}}{{\mathrm{\&mu;}}_0{F}_{\mathrm{zi}0}}\frac{{\mathrm{\&eta;}}_z}{{\mathrm{\&eta;}}_{\mathrm{\&lambda;}}{\mathrm{\&lambda;}}_i^2+1}\frac{C_a}{{\mathrm{\&eta;}}_{\mathrm{\&alpha;}}{\mathrm{\&alpha;}}_i^2+1}{\mathrm{\&alpha;}}_i$

Wherein μ is the adhesion value on current road surface, F _zirespectively analysis of wheel vertical load, λ _ifor wheel slip, α _ifor tyre slip angle, other parameters are tire characteristics parameter, and implication is shown in ^[57].

Be uncertain processing by above formula non-linear conversion, the side force F of wheel _yican be write as: F _yi=k _iα _i

Wherein: wheel lateral deviation characteristic parameter, important uncertain variable parameter.And then can obtain:

Car load transverse force equation of equilibrium is:

$\mathrm{mv}(\stackrel{\&CenterDot;}{\mathrm{\&beta;}}+\mathrm{\&gamma;})=(F_{x1}+F_{x2})\sin \mathrm{\&delta;}-(k_1+k_2)\mathrm{\&beta;}-\frac{({\mathrm{ak}}_1-{\mathrm{bk}}_2)}{v}\mathrm{\&gamma;}+k_1\mathrm{\&delta;}$

K in formula ₁, k ₂respectively front-wheel, trailing wheel lateral deviation characteristic parameter.Consider the impact of steering swivel system and suspension system, k ₁, k ₂important Parameters and uncertain variation.

Car load weaving moment-equilibrium equation is:

$I\stackrel{\&CenterDot;}{\mathrm{\&gamma;}}=-({\mathrm{ak}}_1-{\mathrm{bk}}_2)\mathrm{\&beta;}-\frac{(a^2{k}_1+b^2{k}_2)}{v}\mathrm{\&gamma;}+{\mathrm{ak}}_1\mathrm{\&delta;}+M_z$

M in formula _zcar load yaw moment.

$M_z=\frac{w}{2}(F_{x1}-F_{x2})\cos \mathrm{\&delta;}+a(F_{x1}+F_{x2})\sin \mathrm{\&delta;}+\frac{w}{2}(F_{y1}-F_{y2})\sin \mathrm{\&delta;}+\frac{w}{2}(F_{x4}-F_{x3})$

Vehicle complete vehicle braking longitudinal force equation of equilibrium is:

$m(\stackrel{\&CenterDot;}{v}-\mathrm{\&gamma;\&beta;v})=-((F_{x1}+F_{x2})\cos \mathrm{\&delta;}+(F_{x3}+F_{x4})+(F_{y1}+F_{y2})\sin \mathrm{\&delta;})$

Control system overall design philosophy is for the uncertain time-varying dynamics model of above-mentioned three degree of freedom automobile multivariable nonlinearity, introduces Liapunov stability function.Under the prerequisite of the consistent Asymptotic Stability convergence of Guarantee control system, controlled input and parameter learning rule.Completely different from the classical adaptive control of real-time extreme difference, just introduce and uncertain parameters k here ₁, k ₂, F _xicorresponding adjustable parameter, guarantees the dynamic quality of whole closed loop system by this parameter of online correction.The adjustable parameter of sort controller is the state variable of controller.Designed Robust adaptive controller is a kind of dynamic controller, is particularly suitable for the occasion that requirement of real-time is high.

Design procedure:

First design active front δ controller.Controlling target is to allow

$\stackrel{\&CenterDot;}{\stackrel{\&OverBar;}{\mathrm{\&beta;}}}=-\frac{k_1+k_2}{\mathrm{mv}}\mathrm{\&beta;}-(\frac{{\mathrm{ak}}_1-{\mathrm{bk}}_2}{{\mathrm{mv}}^2}+1)\mathrm{\&gamma;}+\frac{(k_1+F_{x1}+F_{x2})}{\mathrm{mv}}\mathrm{\&delta;}$

K in formula ₁, k ₂becoming attaches most importance to wants uncertain variable parameter.The longitudinal adhesive ability F of tire ground _xialso be important uncertain variable parameter.Getting uncertain variable quantity is designated as:

$k_{11}=\frac{k_1}{k_1+F_{x1}+F_{x2}}>0,$ $k_{21}=\frac{k_2}{k_1+F_{x1}+F_{x2}}>0,$ $k_{31}=\frac{1}{k_1+F_{x1}+F_{x2}}>0,$

k ₄₁＝k ₁+F _x1+F _x2＞0

Have: $k_{31}\stackrel{\&CenterDot;}{\stackrel{\&OverBar;}{\mathrm{\&beta;}}}=\frac{\mathrm{\&beta;v}+\mathrm{a\&gamma;}}{{\mathrm{mv}}^2}{k}_{11}+\frac{\mathrm{v\&beta;}-\mathrm{b\&gamma;}}{{\mathrm{mv}}^2}{k}_{21}+k_{31}\mathrm{\&gamma;}-\frac{1}{\mathrm{mv}}\mathrm{\&delta;}$

Get Lyapunov function:

$V_1=\frac{1}{2}{q}_1{k}_{31}{\stackrel{\&OverBar;}{\mathrm{\&beta;}}}^2+\frac{1}{2}{({\hat{k}}_{11}-k_{11})}^2+\frac{1}{2}{({\hat{k}}_{21}-k_{21})}^2+\frac{1}{2}{({\hat{k}}_{31}-k_{31})}^2>0$

Q in formula ₁> 0 constant.

respectively true value k ₁₁, k ₂₁, k ₃₁estimated valve.

${\stackrel{\&CenterDot;}{V}}_1=q_1\stackrel{\&OverBar;}{\mathrm{\&beta;}}{k}_{31}\stackrel{\&CenterDot;}{\stackrel{\&OverBar;}{\mathrm{\&beta;}}}+({\hat{k}}_{11}-k_{11}){\stackrel{\&CenterDot;}{\hat{k}}}_{11}+({\hat{k}}_{21}-k_{21}){\stackrel{\&CenterDot;}{\hat{k}}}_{21}+({\hat{k}}_{31}-k_{31}){\stackrel{\&CenterDot;}{\hat{k}}}_{31}$

Get:

$\mathrm{\&delta;}=\mathrm{mv}(k_{e1}\stackrel{\&OverBar;}{\mathrm{\&beta;}}+\frac{\mathrm{\&beta;v}+\mathrm{a\&gamma;}}{{\mathrm{mv}}^2}{\hat{k}}_{11}+\frac{\mathrm{v\&beta;}-\mathrm{b\&gamma;}}{{\mathrm{mv}}^2}{\hat{k}}_{21}+{\hat{k}}_{31}\mathrm{\&gamma;})---\left(4\right)$

Self study rule:

${\stackrel{\&CenterDot;}{\hat{k}}}_{11}=q_1\stackrel{\&OverBar;}{\mathrm{\&beta;}}\frac{\mathrm{\&beta;v}+\mathrm{a\&gamma;}}{{\mathrm{mv}}^2},$ ${\stackrel{\&CenterDot;}{\hat{k}}}_{21}=q_1\stackrel{\&OverBar;}{\mathrm{\&beta;}}\frac{\mathrm{v\&beta;}-\mathrm{b\&gamma;}}{{\mathrm{mv}}^2},$ ${\stackrel{\&CenterDot;}{\hat{k}}}_{31}=q_1\stackrel{\&OverBar;}{\mathrm{\&beta;}}\mathrm{\&gamma;}---\left(5\right)$

Wherein k _e1> 0 constant.

${\stackrel{\&CenterDot;}{V}}_1=-q_1{k}_{e1}{e}_1^2<0.$ Have $\underset{t\&RightArrow;\&infin;}{\lim }\stackrel{\&OverBar;}{\mathrm{\&beta;}}=0.$

Then design yaw moment controller and car load speed of a motor vehicle brake controller.The collaborative error of definition conformability:

Definition: conformability is worked in coordination with error

$e_{c1}=\stackrel{\&OverBar;}{\mathrm{\&beta;}}-k_3\stackrel{\&OverBar;}{\mathrm{\&gamma;}};$ $e_{c2}=\stackrel{\&OverBar;}{\mathrm{\&beta;}}-k_4\stackrel{\&OverBar;}{v};$ K in formula ₃> 0, k ₄> 0 undetermined constant.

Wherein $\stackrel{\&OverBar;}{\mathrm{\&gamma;}}=\mathrm{\&gamma;}-{\mathrm{\&gamma;}}_d.$ $\stackrel{\&OverBar;}{v}=v-v_d.$ Had by above-mentioned definition and formula:

${\stackrel{\&CenterDot;}{e}}_{c1}=-\frac{k_1+k_2}{\mathrm{mv}}\mathrm{\&beta;}-(\frac{{\mathrm{ak}}_1-{\mathrm{bk}}_2}{{\mathrm{mv}}^2}+1)\mathrm{\&gamma;}+\frac{(k_1+F_{x1}+F_{x2})}{\mathrm{mv}}\mathrm{\&delta;}+k_3\frac{{\mathrm{ak}}_1-{\mathrm{bk}}_2}{I}\mathrm{\&beta;}+k_3\mathrm{\&gamma;}\frac{a^2{k}_1+b^2{k}_2}{\mathrm{Iv}}-\frac{{\mathrm{ak}}_3{k}_1}{I}\mathrm{\&delta;}-\frac{k_3}{I}{M}_z+k_3{\stackrel{\&CenterDot;}{\mathrm{\&gamma;}}}_d$

${\stackrel{\&CenterDot;}{e}}_{c2}=\frac{k_1+k_2}{\mathrm{mv}}\mathrm{\&beta;}-(\frac{{\mathrm{ak}}_1-{\mathrm{bk}}_2}{{\mathrm{mv}}^2}+1)\mathrm{\&gamma;}+\frac{(k_1+F_{x1}+F_{x2})}{\mathrm{mv}}\mathrm{\&delta;}-k_4\mathrm{\&beta;\&gamma;v}+\frac{k_1{k}_4}{m}(\mathrm{\&beta;}+\frac{\mathrm{a\&gamma;}}{v})\mathrm{\&delta;}+k_4\frac{\mathrm{\&Sigma;}{F}_{x1}}{m}+k_4{\stackrel{\&CenterDot;}{v}}_d$

Above formula is write as: ${\stackrel{\&CenterDot;}{E}}_c=-E_1{k}_1-E_2{k}_2-E_3{k}_{41}+E_0+\mathrm{DU}$

In formula: column vector E _c, E ₁, E ₂, E ₃, E ₀, U ∈ R ^{2 × 1}, matrix D ∈ R ^{2 × 2}, k ₁, k ₂, k ₄₁be real number.

$E_c=\left(\begin{array}{c}{e}_{c1}\\ e_{c2}\end{array}\right),$ $E_1=\left[\begin{array}{c}\frac{\mathrm{av}+\mathrm{a\&gamma;}}{\mathrm{mv}}+\frac{\mathrm{a\&beta;v}+a^2\mathrm{\&gamma;}-\mathrm{av\&delta;}}{\mathrm{Iv}}{k}_3\\ \frac{\mathrm{\&beta;v}+\mathrm{a\&gamma;}}{{\mathrm{mv}}^2}+\frac{k_4}{m}(\mathrm{\&beta;}+\frac{a}{v}\mathrm{\&gamma;})\mathrm{\&delta;}\end{array}\right],$ $E_2=\left[\begin{array}{c}\frac{\mathrm{\&beta;v}-\mathrm{b\&gamma;}}{{\mathrm{mv}}^2}+\frac{\mathrm{b\&beta;v}+b^2\mathrm{\&gamma;}}{\mathrm{Iv}}{k}_3\\ \frac{\mathrm{\&beta;v}-\mathrm{b\&gamma;}}{{\mathrm{mv}}^2}\end{array}\right]$

$E_3=\left(\begin{array}{c}-\frac{\mathrm{\&delta;}}{\mathrm{mv}}\\ -\frac{\mathrm{\&delta;}}{\mathrm{mv}}\end{array}\right),$ $E_0=\left(\begin{array}{c}{k}_3{\stackrel{\&CenterDot;}{\mathrm{\&gamma;}}}_d-\mathrm{\&gamma;}\\ k_4{\stackrel{\&CenterDot;}{v}}_d-k_4\mathrm{\&beta;\&gamma;v}-\mathrm{\&gamma;}\end{array}\right),$ $D=\left(\begin{array}{cc}-\frac{k_3}{I}& \\ & \frac{k_4}{m}\end{array}\right),$ $U=\left(\begin{array}{c}{M}_z\\ \mathrm{\&Sigma;}{F}_{\mathrm{xi}}\end{array}\right)$

∑ F in formula _xifor being subject to total ground, automobile four-wheel adheres to braking force.M _zit is car load yaw rotating torque.

Get Lyapunov function: $V_2=\frac{1}{2}{E}_c^T{\mathrm{QE}}_c+\frac{1}{2}{\stackrel{\&OverBar;}{\mathrm{\&theta;}}}^{T}R\stackrel{\&OverBar;}{\mathrm{\&theta;}}>0$

Q in formula, the permanent matrix of the R > suitable dimension of 0 positive definite.

it is the estimated valve of true value θ. $\widehat{\mathrm{\&theta;}}={\left(\begin{array}{ccc}{\hat{k}}_1& {\hat{k}}_2& {\hat{k}}_{41}\end{array}\right)}^T,$

respectively true value k ₁, k ₂, k ₄₁estimated valve.Ф ^T＝[E ₁ E ₂ E ₃]∈R ^2×3

Have in the above and in formula, to get control inputs amount and be:

$U=D^{-1}({-E}_0+E_1{\hat{k}}_1+E_2{\hat{k}}_2+E_3{\hat{k}}_{41}-K_{e0}{E}_{c1})---\left(6\right)$

Self study rule:

$\stackrel{\&CenterDot;}{\widehat{\mathrm{\&theta;}}}=R^{-1}\mathrm{\&Phi;Q}{E}_C---\left(7\right)$

Have: ${\stackrel{\&CenterDot;}{V}}_2=-E_c^{T}Q{K}_{e0}{E}_c<0.$

In formula: positive definite diagonal angle constant feedback gain matrix K _e0=diag[λ ₁λ ₂], λ _i> 0.

Under formula (4), (5), (6), (7) effect, motor turning error changing Pattern and speed of a motor vehicle error

changing Pattern meet coherence request and have:

When turning to tracking error and unanimously collaborative error E _c, there is longitudinal braking error at → 0 o'clock

thereby automobile have simultaneously good have consistent synergisticing performance and good turn to deceleration and stopping performance and stopping distance the shortest.

Two, second step is for motor tire strong nonlinearity ambiguous model problem, the non-linear consistent cooperative control system of the motor turning of the strong and position pair closed loops that design real-time is good/brake.

Fig. 2 is the non-linear consistent cooperative control system of motor turning/brake of the two closed loops in power and position.Different from Fig. 1, control automobile steering system, brake system and four tires together with the two closed loops of the feedforward of employing power here, force feedback and position feedback.The power closed loop accurate tracking of interior ring is expected to rotate yaw moment, is expected longitudinally to adhere to braking force.The position closed loop accurate tracking of outer shroud is expected turning track and is expected the car load braking speed of a motor vehicle.Tracking error conformability rapidly decays to zero.Keep automobile to have and well turn to and deceleration and stopping performance, stopping distance is short as far as possible.Control inputs amount comprises active front steering angle sigma and is applied to the braking force F on each wheel _bi.

The design of brakig force distribution controller is divided into three aspects.

1, power feedforward and Force Feedback Controller design

Due to the very strong kinetics relation that has between longitudinal adhesive ability, side direction adhesive ability and the yaw moment of the braking force on motor tire strong nonlinearity uncertainty and tire, sideslip angle and tire, their turning efficiency and deceleration and stopping performancies to automobile play a part very crucial.So must design the two closed loop mixture controls in power and position, accurate tracking car load expects to rotate yaw moment M _zdadhere to braking force ∑ F with expectation ground _xid.Wherein M _zd, ∑ F _xidcalculated by formula (6) (7).

The controlling quantity of power feedforward controller is M _zdwith ∑ F _xid.

As the car load yaw moment M expecting _zadhere to braking force F with the ground of expecting _xiwhen unequal with corresponding actual value, produce error delta M _z, Δ ∑ F _xi.Force Feedback Controller produces control inputs amount by control algorithm, four wheel braking force F of fine adjustment _bieliminate error.Making vehicle complete vehicle yaw moment and ground adhere to braking force is all expectation value.

The Force Feedback Controller control algorithm of interior ring adopts good PI or the pid control algorithm of real-time.

2, F is set up in real-time online identification _xiand F _bibetween relation

F _xi(i=1,2,3,4) are that braking force is longitudinally adhered on the ground on each tire, with the braking force F applying on each tire _bi(i=1,2,3,4) difference.Braking force F is longitudinally adhered on the ground of each tire _xibe subject to the impact of the serious nonlinear uncertainty of tire, determined by tire vertical loading and ground attaching coefficient, usually uncertain variation.And braking force F _bibeing applied on each tire by stop mechanism, is control inputs amount.Therefore in order to obtain imposing on the braking force F of each wheel _bi, first set up F _biwith F _xibetween interrelation.

F _xiand F _bibetween have corresponding nonlinear uncertain relation.Here each tire being got to real-time estimation model is:

F _bi＝k _xiF _xi+k _xi0 (8)

F in formula _biit is the braking force that drg actuating unit produces.K _xi, k _xi0be that nonlinear uncertain changes, count for much from the condition of road surface of different roads and the vertical load of each wheel.Brake model for each wheel has:

$J\stackrel{\&CenterDot;}{\mathrm{\&omega;}}=F_{\mathrm{xi}}r-T_b---\left(9\right)$

Calculate this wheel angular acceleration in measurement

act on lock torque T on wheel _b, after tire radius r, braking force F is adhered on longitudinal ground that can obtain this wheel _xiestimated valve

wherein act on braking force on wheel $F_{\mathrm{bi}}=\frac{T_{\mathrm{bi}}}{r}.$

Due to k _xi, k _xi0that nonlinear uncertain changes.Need on-line identification.(8) formula is converted into down and is established an equation: y (k)=Ф ₁ ^t(k) θ ₁+ η (k) (10)

In formula: y (k)=F _bi(k), θ ₁=[k _xik _xi0] ^t, η (k) measures noise, Ф ₁ ^t=[F _xi1].

Parameter identification adopts the good band dead band projection algorithm of real-time ^[50]:

${\widehat{\mathrm{\&theta;}}}_1(k+1)={\widehat{\mathrm{\&theta;}}}_1\left(k\right)+\mathrm{\&alpha;}\left(k\right){\mathrm{\&Phi;}}_1\left(k\right){(c_{0}I+{{\mathrm{\&Phi;}}_1}^T{\mathrm{\&Phi;}}_1)}^{-1}e\left(k\right)$ $e\left(k\right)=y\left(k\right)-{{\mathrm{\&Phi;}}_1}^T\left(k\right){\widehat{\mathrm{\&theta;}}}_1\left(k\right)$

${\mathrm{\&alpha;}}_i\left(k\right)=\left\{\begin{array}{cc}1& \left|e\left(k\right)\right|>2\mathrm{\&Delta;}\\ 0& \left|e\left(k\right)\right|<2\mathrm{\&Delta;}\end{array}\right.$

α in formula (k)=diag{ α _i, i=1,2.C ₀> 0 prevents the non-existent constant of inverse matrix.

it is the estimated valve of the θ in k+1 moment.Δ permissible error.Can estimate k corresponding to this moment of each wheel with above-mentioned algorithm _xi, k _xi0.

Longitudinally adhere to braking force F by each wheel ground of known expectation _xi, calculate the braking force F of the drg actuating unit generation of obtaining expectation from formula (8) _bi.

3, wheel braking force distributing box

The car load yaw moment controlling quantity M being obtained by (6), (7) formula _zneed to be reasonably allocated to each wheel, be realized by the brake actuator on each wheel.Concerning car load, have:

$M_z=\frac{w}{2}(F_{x1}-F_{x2})\cos \mathrm{\&delta;}+a(F_{x1}+F_{x2})\sin \mathrm{\&delta;}+\frac{w}{2}(F_{y1}-F_{y2})\sin \mathrm{\&delta;}+\frac{w}{2}(F_{x4}-F_{x3})$

∑F _xi＝F _x1+F _x2+F _x3+F _x4 (11)

If four wheel braking forces are F _bi(i=1,2,3,4).First select master and slave brake wheel according to following principle.

In practice due to friction circle, front-wheel and trailing wheel apply respectively equate that braking force produces will have a great difference around barycenter yaw moment.In the time of oversteer, outside front-wheel apply braking force to correct oversteer the most effective.In like manner, on inner rear wheel, apply braking force for correcting the most effectively [25] of understeering.Therefore in the time of automobile oversteer, select two front-wheels of automobile as main brake wheel, two rear wheels are from brake wheel.In the time of automobile understeering, select two trailing wheels of automobile as main brake wheel, two front vehicle wheels are from brake wheel.If two front vehicle wheels are centre brake drum, two rear wheels are during from brake wheel, and braking force F is longitudinally adhered on the ground on two front vehicle wheels (main brake wheel) _xiobtain formula below by separating (11) formula.：

$F_{x1}=\frac{1}{2}(\mathrm{\&Sigma;}{F}_{\mathrm{xi}}-(F_{x3}+F_{x4}))+\frac{1}{w\cos \mathrm{\&delta;}}(M_z-a\sin \mathrm{\&delta;}(\mathrm{\&Sigma;}{F}_{\mathrm{xi}}-(F_{x3}+F_{x4}))-\frac{w\sin \mathrm{\&delta;}}{2}(F_{y1}-F_{y2})-\frac{w}{2}(F_{x4}-F_{x3}))$

$F_{x2}=\frac{1}{2}(\mathrm{\&Sigma;}{F}_{\mathrm{xi}}-(F_{x3}+F_{x4}))-\frac{1}{w\cos \mathrm{\&delta;}}(M_z-a\sin \mathrm{\&delta;}(\mathrm{\&Sigma;}{F}_{\mathrm{xi}}-(F_{x3}+F_{x4}))-\frac{w\sin \mathrm{\&delta;}}{2}(F_{y1}-F_{y2})-\frac{w}{2}(F_{x4}-F_{x3}))$

Two rear wheels (from brake wheel) F _x3, F _x4respectively by M _zand F _xiforce feedback PI control algorithm provide.Wherein: M _zwith ∑ F _xiobtained by formula (6) (7).δ can be obtained by actual measurement.Wheelspan w and a are constant.Wherein side force of tire F _yicalculated by estimator: and

corresponding wheel longitudinally adheres to braking force F this moment _xireal-time estimator, can obtain by estimating in formula (9).Finally calculate the braking force F on each tire by formula (8), (10) again _bi.

Accompanying drawing explanation:

Fig. 1 is the consistent cooperative control system schematic diagram of motor turning of the present invention/brake.

Fig. 2 is the consistent cooperative control system composition diagram of motor turning/brake of the two closed loops in the power of having of the present invention and position.

Fig. 3 is automobile hardware system of the present invention

The control of turn in Fig. 3/ABS (Anti-lock Braking System) shares same hardware system.This is a distributed and multi-layer control structure.Between the industrial computer execution design of one, upper strata and multi-controller, coordinate.Five one-chip computers of lower floor complete the Robust Adaptive Control to four wheels and a steering wheel angle.Brake actuator utilizes vehicle existing installation additional motor drg again.Steering wheel angle actuating unit is that additional motor speed reduction gearing drives.The sensor photoelectric code disk that tests the speed, the gyro inertial navigation combination of measuring speed and car body yaw velocity, induction type angular velocity sensor, acceleration pick-up, distance measuring sensor, power and torque sensor etc.

The real-time performance of the controller of design meets automobile control requirement.In all controlling quantitys, quantity of state all can be measured in real time or calculate.Thereby whole control system can realize completely.

Claims (3)

1. the desired braking speed based on automobile in event design calculation braking procedure.

First choose motion reference variable s and be defined as the stopping distance of passing by automobile barycenter place.Model on the time t territory of automobile speed is obtained to speed and acceleration/accel about the function expression v=V (s) on distance domain s, a=A (s) by conversion.Then get optimal performance index stopping distance the shortest, utilize the Hamiltonian function about s, solve optimal solution by greatly-minimal principle, can solve and expect vehicle complete vehicle braking desired speed.

2. in technical foundation according to claim 1, define conformability tracking error, the consistent cooperative control system of design vehicle non linear robust self adaptation.

Motor turning/braking is controlled and is subject to Wheel slip parameter k ₁, k ₂longitudinally adhere to braking force F _xithe factor such as nonlinear uncertain variation, coupled interference affect greatly.Disturb in order to compensate these, the present invention defines conformability tracking error and designs the consistent collaborative controller of non linear robust self adaptation.The good steering swivel system of Collaborative Control and brake system, accurate tracking turning track, the desired braking speed of a motor vehicle.Allow tracking error conformability rapidly decay to zero.

3. according to designing the non-linear consistent cooperative control system of the motor turning of the strong and position pair closed loops that real-time is good/brake on control technology basis described in claim 1 or 2.

Existing relevant automotive research only has position control, does not carry out power closed loop control.Turning efficiency and the deceleration and stopping performance impact of tire strong nonlinearity relation on automobile is very large.The present invention utilizes power feedforward, force feedback and the uncertain variation of position feedback double closed-loop control system dynamic compensation tire strong nonlinearity.And the two closed loop hybrid control systems in power and position can improve controller performance better, compensate better the strong nonlinearity uncertainty of tire.

Images