CN105857309A - Automotive adaptive cruise control method taking multiple targets into consideration - Google Patents

Abstract

The invention discloses an automotive adaptive cruise control method taking multiple targets into consideration. A layer control strategy is utilized, upper layer control can decide an expected longitudinal acceleration according to a target vehicle and a current state of a controlled vehicle, and lower layer control can track the expected longitudinal acceleration through a reverse recursion method. The automotive adaptive cruise control method comprises the following steps: a mutual longitudinal dynamitic model between two vehicles is established, a design model predicating controller can obtain the expected distance between two vehicles according to a constant time headway strategy, a model prediction control algorithm is utilized to decide an expected longitudinal acceleration for tracking the expected distance between the two vehicles, vehicle control work conditions are divided into a driving work condition and a braking work condition, reverse longitudinal dynamic models for the two work conditions are established respectively according to a vehicle driving equation, an expected throttle percentage is obtained according to the vehicle reverse longitudinal dynamic model and the expected acceleration in the driving work condition, and an expected brake pedal openness is obtained according to the expected acceleration in the braking work condition.

Description

A kind of multiobject adaptive cruise control method of consideration

Technical field

The present invention relates to a kind of adaptive cruise control method, be specifically related to the multiobject vehicle adaptive cruise of a kind of consideration Control method.

Background technology

Adaptive learning algorithms (ACC) system is to combine Vehicle Safety Distance Keeping System on the basis of traditional cruise control system to drill Change.Detect whether the visual range at radar exists front truck by being positioned at the radar sensor of body forward structure, work as road When front is without vehicle, ACC vehicle can travel according to the speed being previously set, and detects that there is vehicle in front once onboard sensor, ACC system by adjusting this car speed, be allowed to front truck ensure a safety with following distance.The purpose of this system design is intended to Reduce the traffic accident caused because of the faulty operation of driver, improve driving safety, riding comfort etc..

ACC system mainly considers security and with the big target of car two when design at present, but during actual design also Have and what time still need to consider.First, according to American National Instrument Highway Traffic Safety Administration (NHTSA) in 2008 Pointing out the survey report of automobile ACC system, comfortableness is one of most concerned performance of driver, because comfort level can not get having Effect guarantee will result directly in passenger and refuses to use ACC system.Visible, in the design process carrying out ACC system, comfortableness is not Obtain one of irrespective performance.Secondly, along with the increase day by day of environmental pressure, energy problem becomes one of focus of people's care. Therefore, can whether ACC system have higher fuel economy and also determine its key factor promoted on road.

It is simple that Model Predictive Control (MPC) method that the present invention uses has algorithm design, strong robustness, and can process excellent Multiple control targets in change problem and the feature of multiple constraint.Briefly Model Predictive Control (MPC) is a kind of optimized algorithm, The tracing control to expectation input is realized by the thought rolling optimizing and feedback compensation.Work as according to system in each sampling instant Front measured state, utilizes Model forecast system output in future, asks by solving the optimization comprising object function and system restriction Topic, obtains optimization, in order to reduce the impact of external disturbance and model mismatch, is arrived by the first of optimization element interaction System, just completes a step and controls input.Said process is repeated in next sampling instant.

Summary of the invention

The invention provides a kind of multiobject adaptive cruise control method of consideration, the control strategy of employing layering: upper strata Controlling the state current according to target vehicle and this car, the multiple objective decisions considered in vehicle travel process go out desired Longitudinal acceleration, lower floor controls to realize the tracking to expectation longitudinal acceleration by the method for backstepping.

The present invention is achieved by the following technical solutions:

A kind of multiobject adaptive cruise control method of consideration, uses muti-layer control tactics: top level control is according to target carriage And the current state of controlled vehicle, decision-making goes out desired longitudinal acceleration；Lower floor controls to be realized the phase by the method for backstepping Hope the tracking of longitudinal acceleration；The method comprises the following steps:

Step one, set up two car twisting movement models: according to the kinematic relation between controlled vehicle-to-target vehicle, Set up two car twisting movement models, simultaneously using front truck acceleration information as disturbing signal；

Step 2, the design of host controller: the two car twisting movement models set up based on step one, design a model pre- Survey controller, obtain desired two following distances according to constant time headway strategy, according to the real-time status of vehicle, utilize model pre- Survey control algolithm decision-making to go out to follow the tracks of the expectation longitudinal acceleration required for this expectation following distance；

Step 3, set up vehicle against Longitudinal Dynamic Model: wagon control operating mode is divided into driving operating mode and damped condition, to two Kind operating mode is set up vehicle according to vehicle equation respectively and is used for institute against Longitudinal Dynamic Model against Longitudinal Dynamic Model, vehicle State the instruction of the expectation acceleration that host controller calculates to be changed into desired air throttle by vehicle against Longitudinal Dynamic Model and open Degree or desired brake pedal aperture；

Step 4, the design of lower level controller: according to vehicle against Longitudinal Dynamic Model, desirably accelerate under driving operating mode Degree tries to achieve desired throttle opening, and under damped condition, acceleration desirably tries to achieve desired brake pedal aperture；To obtain The control signal obtained exports to controlled vehicle, completes the tracing control to expectation following distance.

The invention have the benefit that

1. the design philosophy levels function of the hierarchy that the present invention uses is concentrated and controls with clearly defined objective, and intermodule only transmits must The limit signal wanted and being independent of each other, is beneficial to debug system and can improve to a certain extent robustness and the reliability of system.

2. the present invention is during setting up the two mutual Longitudinal Dynamic Models of car, takes into full account the impact of front truck acceleration, should Signal is as disturbance, and this model is not related to the use of vehicle dynamics and parameter thereof simultaneously, it is adaptable to test beyond main car The transplanting of other vehicles, beneficially control algolithm.

3. the present invention has taken into full account during following the tracks of desired following distance, and main car needs the multiple traveling target met, bag Include security, with car, comfortableness and fuel economy etc..

Accompanying drawing explanation

Fig. 1 is adaptive cruise tracking control system block diagram；

Fig. 2 is two car twisting movement model schematic；

Fig. 3 is engine torque characteristic map schematic diagram；

Fig. 4 is for accelerating to control two following distance schematic diagrames；

Fig. 5 is for accelerating to control two vehicle speed schematic diagrames；

Fig. 6 is for accelerating to control controlled quentity controlled variable change schematic diagram；

Fig. 7 controls two following distance schematic diagrames for slowing down；

Fig. 8 controls two vehicle speed schematic diagrames for slowing down；

Fig. 9 controls controlled quentity controlled variable change schematic diagram for slowing down.

Detailed description of the invention

Technical scheme it is discussed in detail below in conjunction with accompanying drawing:

The invention provides and a kind of consider multiobject adaptive cruise control method, the method includes following step:

Step one, set up two car twisting movement models, as shown in Figure 2.

The kinematic relation met in the process of moving according to front and back's car, it is possible to obtain equation below:

$\left\{\begin{array}{c}v(k+1)=v\left(k\right)+a_f\left(k\right)T_s\\ v_{ref}(k+1)=v_{ref}\left(k\right)+a_l\left(k\right)T_s-a_f\left(k\right)T_s\\ \mathrm{\Δ}x(k+1)=\mathrm{\Δ}x\left(k\right)+v_{ref}\left(k\right)T_s+\frac{1}{2}\left(a_l\right(k)-a_f(k\left)\right)T_s^2\end{array}\right.---\left(1\right)$

Wherein, the longitudinal driving speed of carving copy car, unit m/s when v (k) represents k；a_l(k)、a_fK () is before and after two car k respectively The acceleration information in moment, unit m/s²；v_refK () represents the relative velocity of k moment two car, unit m/s, meet v_ref(k)=v_l(k)-v (k), v_lK () is the longitudinal velocity in front truck k moment, unit m/s；Δ x (k) is k moment two following distance, single Position m；T_sIt is the sampling period of system, unit s.

Choose state vector x (k)=[Δ x (k), v_ref(k),v(k)]^T, controlled vehicle acceleration is inputted, i.e. as the control of system U (k)=a_fK (), ordinary circumstance can utilize the acceleration transducer of vehicle to obtain very easily for the longitudinal acceleration of controlled vehicle , but owing to truck traffic not yet realizes, therefore current time goes for the acceleration of front truck and there is also bigger difficulty. Based on considerations above, it is believed that front truck (target vehicle) acceleration is the disturbance of ACC system, i.e. w (k)=a_lK (), owing to controlling Final goal be make two car actual pitch level off to expectation that spacing policy calculation goes out with following distance, therefore, the output choosing of system Select two car actual pitch.Above-mentioned equation thus can be described as the form of state-space expression as follows:

$\{\begin{array}{c}x(k+1)=Ax\left(k\right)+Bu\left(k\right)+Gw\left(k\right)\\ y\left(k\right)=Cx\left(k\right)\end{array}---\left(2\right)$

Wherein,

C=[1 0 0]

Step 2, the design of host controller: the two car twisting movement models set up based on step one, design a model pre- Survey controller, obtain desired two following distances according to the most widely used constant time headway strategy, according to the real-time shape of vehicle State, utilizes Model Predictive Control Algorithm decision-making to go out to follow the tracks of the desired longitudinal acceleration required for this expectation following distance.This designed Journey specifically comprises the following steps that

1) formation of optimization problem

The main target of ACC system has following 4 points: security, tracking performance, comfortableness and fuel economy.But these are several Point is conflicting.To meet cost-effectiveness requirement, it would be desirable that vehicle is the most steady, does not the most exist and adds Speed situation jumpy, this will certainly affect tracking performance.On the contrary, if design controller during only consider with , the most inevitably there is unnecessary acceleration and the generation of brake hard situation in track performance, this not only affects fuel economy, To a certain extent, if driver can not well adapt to adaptive cruise control system, then so-called trust crisis can be produced, Thus the consequence brought is exactly driver carries out pro-active intervention frequently, this not only with the design original intention of adaptive cruise control system Disagree, also can bring extra mental burden.In sum, in the design process being controlled system, only consider it In any one be irrational, it is necessary under same framework, take into account multiple target simultaneously.In order to quantify the ACC system proposed The performance indications of system, we reanalyse above-mentioned target.

First, no matter taking which kind of algorithm, security is the primary goal that system the most at every moment will meet. That is in order to meet security requirement, at any time two following distances will more than the following distance of a safety, as following about Shown in Shu Fangcheng (3).

Constraint 1: Δ x (k) >=d_c (3)

Wherein, d_cRepresent two following distances of safety.

Secondly, for tracking performance, driver expects actual two following distances tracking upper desired two following distances during stable state.

Target 1: Δ x (k) → Δ x_desWhen k → ∞ (4)

Wherein, Δ x_desRepresent desired two following distances.

Finally, for riding comfort and the requirement of fuel economy, in vehicle traveling process, embody the finger of riding comfort This parameter of longitudinal acceleration of mark mainly vehicle, the least riding comfort of absolute value of acceleration is the highest, smooths simultaneously Dynamic response curve is also beneficial to the raising of fuel economy.

Constraint 2:a_fmin≤a_f(k)≤a_fmax (5)

In addition, it is contemplated that the restriction of vehicle self-ability, vehicle travel process also needs to meet following constraint of velocity:

Constraint 3:v_min≤v(k)≤v_max (6)

In sum, under the framework of MPC, consider that the control of multiobject vehicle ACC system can be summarized as following optimization and ask Topic:

Problem one:

$\underset{u\left(k\right)}{min}J\left(y\right(k),u(k),m,p)---\left(7\right)$

Satisfied two car twisting movements:

$\left\{\begin{array}{c}v(k+1)=v\left(k\right)+a_f\left(k\right)T_s\\ v_{ref}(k+1)=v_{ref}\left(k\right)+a_l\left(k\right)T_s-a_f\left(k\right)T_s\\ \mathrm{\Δ}x(k+1)=\mathrm{\Δ}x\left(k\right)+v_{ref}\left(k\right)T_s+\frac{1}{2}\left(a_l\right(k)-a_f(k\left)\right)T_s^2\end{array}\right.$

Meet inequality constraints simultaneously:

$\left\{\begin{array}{c}\mathrm{\Δ}x\left(k\right)\≥d_c\\ a_{fmin}\≤a_f\left(k\right)\≤a_{fmax}\\ v_{\min }\≤v\left(k\right)\≤v_{max}\end{array}\right.$

Wherein,

$\begin{array}{c}J\left(y\right(k),u(k),m,p),\\ =\underset{i=1}{\overset{p}{\Σ}}{||{\mathrm{\Γ}}_{y,i}\left(y_c\right(k+i|k)-r(k+i\left)\right)||}^2+\underset{i=1}{\overset{m}{\Σ}}{||{\mathrm{\Γ}}_{u,i}(k+i-1)||}^2\end{array}$

In formula, p is the prediction time domain of system, and m is to control time domain and m≤p.

2) the solving of optimization problem

Assume that all of state all can be measured to obtain, for predictive equation of deriving, also need to do hypothesis below:

(1) controlling outside time domain, controlled quentity controlled variable is constant, i.e. u (k+i)=u (k+m-1), i=m, m+1 ... p-1.

(2) interference keeps constant after the k moment, i.e. w (k+i)=w (k), i=1,2 ... p-1. Solving for the ease of controller, the expression-form of predictive equation of first deriving, derivation is as follows:

X (k+1 | k)=Ax (k)+Bu (k)+Gw (k)

X (k+2 | k)=Ax (k+1 | k)+Bu (k+1)+Gw (k+1)

=A²x(k)+ABu(k)+Bu(k+1)+(AG+G)w(k)

X (k+3 | k)=Ax (k+2 | k)+Bu (k+2)+Gw (k+2)

=A³x(k)+A²Bu(k)+ABu(k+1)+Bu(k+2)+(A²G+AG+G)w(k)

Analogize and can obtain:

X (k+m | k)=Ax (k+m-1 | k)+Bu (k+m-1)+Gw (k+m-1)

=A^mx(k)+A^m-1Bu(k)+A^m-2Bu(k+1)+…+ABu(k+m-2)+Bu(k+m-1)

+(A^m-1G+A^m-2G+…+AG+G)w(k)

X (k+p | k)=Ax (k+p-1 | k)+Bu (k+p-1)+Gw (k+p-1)

=A^px(k)+A^p-1Bu(k)+A^p-2Bu(k+1)+…+A^p-mBu(k+m-1)+

A^p-m-1Bu(k+m-1)+…+ABu(k+m-1)+Bu(k+m-1)+

(A^p-1G+A^p-2G+…+AG+G)w(k)

Due to:

Y (k)=Cx (k)

So:

Y (k+1 | k)=Cx (k+1 | k)

=CAx (k)+CBu (k)

Y (k+2 | k)=CA²x(k)+CABu(k)+CBu(k+1)+(CAG+CG)w(k)

Analogize and can obtain:

Y (k+m | k)=CA^mx(k)+CA^m-1Bu(k)+CA^m-2Bu(k+1)+…+CABu(k+m-2)+

CBu(k+m-1)+(CA^m-1G+CA^m-2G+…+CAG+CG)w(k)

Y (k+p | k)=CA^px(k)+CA^p-1Bu(k)+CA^p-2Bu(k+1)+…+CA^p-mBu(k+m-1)

+CA^p-m-1Bu(k+m-1)+…+CABu(k+m-1)+CBu(k+m-1)

+(CA^p-1G+CA^p-2G+…+CAG+CG)w(k)

Definition p step prediction output vector and m step input vector are as follows:

$\begin{array}{ccc}{Y}_p(k+1|k)\stackrel{def}{=}\left[\begin{array}{c}y(k+1|k)\\ y(k+2|k)\\ .\\ .\\ .\\ y(k+p|k)\end{array}\right]& U\left(k\right)\stackrel{def}{=}\left[\begin{array}{c}u\left(k\right)\\ u(k+1)\\ .\\ .\\ .\\ u(k+m-1)\end{array}\right]& W\left(k\right)=\left[\begin{array}{c}w\left(k\right|k)\\ w\left(k\right|k)\\ .\\ .\\ .\\ w\left(k\right|k)\end{array}\right]\end{array}$

The expression formula that can obtain predictive equation is as follows:

Y_p(k+1 | k)=S_xx(k)+S_wW(k)+S_uU(k) (8)

Wherein,

According to CTH spacing policy module, it is desirable to two following distances should meet following relation with this vehicle speed:

R (k+i)=t_h·v(k+i)+r₀ (9)

In formula, t_hRepresent headway, r₀It it is relevant with a security constant.

DefinitionOwing to r (k+i) is relevant with input u (k), so handle 3rd state v (k) of system is defined as the output v of system_b, then

v_b(k+1)=v_b(k)+T_su(k) (10)

Above-mentioned output is write as the form of matrix

V_b(k+1 | k)=V_xx(k)+V_uU(k) (11)

Wherein,

DefinitionThen R (k+1)=t_h[V_xx(k)+V_uU(k)]+R₀, bring above-mentioned expression-form into mesh Scalar functions (7), and define E_p(k+1 | k)=(S_x-t_hV_x)x(k)+S_wW(k)-R₀, rearrange this object function and just can obtain as follows Form:

$\begin{array}{c}J={||{\mathrm{\Γ}}_y\left(E_p\right(k+1|k)+(S_u-t_h{V}_u\left)U\right(k\left)\right)||}^2+{||{\mathrm{\Γ}}_{u}U\left(k\right)||}^2\\ =U{\left(k\right)}^T\[{(S_u-t_h{V}_u)}^T{{\mathrm{\Γ}}_y}^T{\mathrm{\Γ}}_y(S_u-t_h{V}_u)+{{\mathrm{\Γ}}_u}^T{\mathrm{\Γ}}_u\]U\left(k\right)\\ +2E_p{(k+1|k)}^T{{\mathrm{\Γ}}_y}^T{\mathrm{\Γ}}_y(S_u-t_h{V}_h)U\left(k\right)\\ +E_p{(k+1|k)}^T{{\mathrm{\Γ}}_y}^T{\mathrm{\Γ}}_y{E}_p(k+1|k)\end{array}---\left(12\right)$

Due to E_p(k+1|k)^TΓ_y ^TΓ_yE_p(k+1 | k) is unrelated with optimized variable, the form being so written as by object function (12):

J=U (k)^THU(k)+G(k+1|k)^TU(k) (13)

Wherein,

H=(S_u-t_hV_u)^TΓ_y ^TΓ_y(S_u-t_hV_u)+Γ_u ^TΓ_u

G(k+1|k)^T=2E_p(k+1|k)Γ_y ^TΓ_y(S_u-t_hV_u)

The constraints conversion by optimization problem is needed to become C for the ease of solving of controller_uThe form of z >=b.

Can be to be converted into the form of the output constraint of system for security constraint equation (3):

$\mathrm{\Δ}x\left(k\right)\≥d_c\⇒S_{u}U\left(k\right)\≥D_c-S_{x}x\left(k\right)-S_{w}W\left(k\right)---\left(14\right)$

Wherein S_u,S_x,S_wExpression formula is identical with above-mentioned, D_c=[d_c d_c … d_c]^T。

Can be to be converted into following form for control constraints (5):

$a_{f\min }\≤a_f\left(k\right)\≤a_{f\max }\⇒\left[\begin{array}{c}-I_{m*m}\\ I_{m*m}\end{array}\right]U\left(k\right)\≥\left[\begin{array}{c}-U_{\max }\\ U_{\min }\end{array}\right]---\left(15\right)$

Wherein, U_max=[a_fmax a_fmax … a_fmax]^T,U_min=[a_fmin a_fmin … a_fmin]^T。

This vehicle speed of definition is previously noted for state constraint (6) export as a constraint, and the predictive equation of output of deriving, this The expression-form that above-mentioned constraint just can be written as by sample:

$\begin{array}{c}-V_{u}U\left(k\right)\≥V_{x}x\left(k\right)-V_{\max }\\ V_{u}U\left(k\right)\≥V_{\min }-V_{x}x\left(k\right)\end{array}---\left(16\right)$

Wherein, V_x,V_uExpression formula as it was previously stated,

V_max=[v_max v_max … v_max]^T,V_min=[v_min v_min … v_min]^T。

So far, just the constraint of system is all changed complete.So optimization problem one may finally change into problem two:

Problem two:

Wherein,

H=(S_u-t_hV_u)^TΓ_y ^TΓ_y(S_u-t_hV_u)+Γ_u ^TΓ_u

G(k+1|k)^T=2E_p(k+1|k)Γ_y ^TΓ_y(S_u-t_hV_u)

$\begin{array}{cc}{C}_u=\left[\begin{array}{c}-I_{m*m}\\ I_{m*m}\\ S_u\\ V_u\\ -V_u\end{array}\right]& b(k+1|k)=\left[\begin{array}{c}-U_{\max }\\ U_{\min }\\ D_c-S_{x}x\left(k\right)-S_{w}W\left(k\right)\\ V_{\min }-V_{x}x\left(k\right)\\ V_{x}x\left(k\right)-V_{\max }\end{array}\right]\end{array}$

In Calling MATLAB, solver quadprog just can complete solving of host controller, it is thus achieved that desired longitudinal acceleration.

Step 3, set up vehicle against Longitudinal Dynamic Model: the autonomous vehicle studied herein is all based on the simulation software of high-fidelity Self shifter vehicle in veDYNA, say, that without considering the impact of gear during vehicle travels.So affect The factor of autonomous vehicle longitudinal driving is mainly throttle opening and two factors of brake pedal aperture.Running according to vehicle Understand, when, after given throttle opening input, generation is exported moment of torsion by engine accordingly, and this moment of torsion is through hydraulic moment changeable Device sends the speed changer of vehicle to, acts on wheel eventually through bearing arrangement, produces corresponding driving moment.Damped condition Same, when after given brake pressure input, brake fluid system the moment produced is applied directly on wheel, compels Vehicle deceleration is made to travel.According to above-mentioned analysis, host controller the expectation acceleration calculated instruction must be inverse longitudinally by vehicle Kinetic model is changed into the position of desired throttle opening and brake pedal, then this control signal is applied to controlled vehicle, To control the acceleration of vehicle, deceleration and uniform motion, it is achieved the function of self-adaption cruise system.So needing be divided into acceleration and subtract Two kinds of operating modes of speed are set up against Longitudinal Dynamic Model:

A. accelerate to control (driving operating mode)

After logic switch, if switching to acceleration control, then must the requirement of desirably acceleration, through being calculated expectation Motor torque, then check in desired throttle opening by the reverse model of engine.First vapour is set up according to Newton's second law Car traveling equation:

$\mathrm{\δ}ma=\frac{i_g{i}_0{\mathrm{\η}}_T}{r_{eff}}{T}_e-mgf-\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2-mgsin\mathrm{\θ}---\left(18\right)$

The output moment of torsion of engine can be obtained according to above formula:

$T_e=\frac{(mgf+\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2+mgsin\mathrm{\θ}+\mathrm{\δ}ma)r_{eff}}{i_g{i}_0{\mathrm{\η}}_T}---\left(19\right)$

In formula, the implication of each symbol is as follows:

T_eIt is that engine expects moment of torsion, i_gIt is the gearratio of speed changer, i₀Represent main step-down ratio, η_TRepresent the machine of power train Tool efficiency, r_effBeing the effective radius of wheel, m is complete vehicle quality, and f is coefficient of rolling resistance, C_dIt is coefficient of air resistance, A Being front face area, ρ is atmospheric density, and v is longitudinal direction of car travel speed, and θ represents road grade, and δ is vehicle rotary quality Conversion coefficient, a is the longitudinal acceleration of vehicle, and g is acceleration of gravity.

Engine torque characteristic map schematic diagram in veDYNA auto model, counter tabling look-up is utilized to can be obtained by correspondence under this moment of torsion The size of throttle opening, as shown in Figure 3.According to T_eWith engine speed ω_e, utilize engine air throttle open degree characteristic curve Figure, can be in the hope of desired throttle opening α_desFor

α_des=f (T_e,ω_e) (20)

B. control for brake (damped condition)

After logic switch, as switched to control for brake, must desirably acceleration, try to achieve desired braking moment, then Desired brake pedal aperture is tried to achieve, by β by the reverse model of brake_desPut on controlled vehicle by actuator to be braked Control.

When braking travels, the equation of vehicle is as follows:

$\mathrm{\δ}ma=F_b+mgf+\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2+mgsin\mathrm{\θ}---\left(21\right)$

Engine braking moment expression formula can be obtained according to above formula:

$T_b=(\mathrm{\δ}ma-mgf-\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2-mgsin\mathrm{\θ})r_{eff}---\left(22\right)$

Think that the four wheels of vehicle is identical herein, say, that the braking moment of car load by four wheels mean allocation, this Sample just can be with the braking moment of each wheelOwing to the calculating of the braking moment on each wheel in veDYNA is public Formula meets equation below:

M_b=2 (P_b·A_b·r_b·μ_b) (23)

In formula,

P_bRepresent the brake pressure on each wheel, unit: Pa

μ_bRepresent the coefficient of friction between brake block and brake disc

A_bRepresent the contact area of friction plate and brake disc, unit: m²

r_bRepresent braking radius, unit: m

Above-mentioned parameter is brought into the relation of brake pressure and the braking moment that can be obtained by each wheel,

M_b=0.1323 × 10^-3·P_b (24)

Obtain desired brake pressure P_bAfter, convert thereof into desired brake pedal aperture and just complete the deceleration under damped condition Control.In veDYNA auto model, the aperture relation of braking moment and brake pedal meets equation below:

${\mathrm{\β}}_{des}=\frac{P_b}{P_{bmax}}\×100\%---\left(25\right)$

In formula, P_bmaxRepresenting maximum brake pressure, value is 2 × 10⁷Pa。

Step 4, the design of lower level controller: the expectation acceleration that lower level controller solves according to host controller, first warp Cross logic judgment module decision-making and go out in order to the requirement following the tracks of this expectation acceleration takes to drive module or brake module, herein I Take simplest changing method based on threshold value, it is believed that when acceleration more than zero when use drive control, acceleration is little Control for brake is taked in zero when.As control need to be driven, obtain desired driving moment according to formula (19), according to the most anti- The engine speed information of feedback, and equation (20) is obtained with corresponding throttle opening, this control signal effect is given Controlled vehicle, completes to drive the tracing control under operating mode.In like manner, as needed control for brake, first obtain according to formula (22), (23) Desired driving moment, is obtained with the aperture of corresponding brake pedal further according to equation (24), (25), and this is controlled letter Number effect to controlled vehicle, complete the tracing control under damped condition.

The off-line simulation checking of of the present invention consideration multiobject adaptive cruise control method is given below.

The validity based on multiobject adaptive cruise control method proposed for the checking present invention, during choosing cruise Two kinds of typical conditions are verified, concrete experimental result and analysis are given below.

(1) Control release result is accelerated

Before setting in experiment, 20s is operated vehicle by the Virtual drivers of veDYNA, makes vehicle accelerate to 120km/h, before observing At side 100m, front truck travels with the speed of 100km/h, is 40.33m through calculating desired two following distances, and actual two following distances are 100m, under controller action, the several states in vehicle travel process are as shown in Fig. 4, Fig. 5, Fig. 6.As can be seen from the results, The initial time of controller action, actual two following distances are more than desired two following distances, and first ACC system controls this car and accelerate, To shorten the spacing between two cars, two following distances are made to level off to desired Safety distance, when two following distances shorten to a certain degree, ACC system controls this car and relatively evenly slows down, and makes the speed speed close to preceding vehicle of ACC vehicle, meanwhile, makes two cars Between spacing progressively narrow down to the desired Safety distance in two workshops of default.In this process, two following distances are the biggest Change in the following distance vehicle acceleration simultaneously of safety is also at rational scope.

(2) deceleration Control release result

Before setting in experiment, 20s is operated vehicle by the Virtual drivers of veDYNA, makes vehicle accelerate to 120km/h, before observing At side 35m, front truck travels with the speed of 100km/h, is 40.33m through calculating desired two following distances, and actual two following distances are 35m, under controller action, the several states in vehicle travel process are as shown in Fig. 7, Fig. 8, Fig. 9.As can be seen from the figure at t= During 20s, two car initial separation are less than safe spacing, and the traveling of following between two cars has certain insecurity, and ACC system is straight Connect this car of control and carry out the deceleration of some strength, improve two cars and follow the security of traveling, when main vehicle speed is reduced to a certain degree, The slightly smaller than speed of front truck, then ACC system controls vehicle and carries out suitable acceleration, makes the speed of ACC vehicle level off to front truck Speed, makes the spacing between two cars level off to desired Safety distance simultaneously.

Claims (4)

1. one kind considers multiobject adaptive cruise control method, it is characterised in that employing muti-layer control tactics: upper strata Controlling the state current according to target vehicle and controlled vehicle, decision-making goes out desired longitudinal acceleration；Lower floor controls to pass through backstepping Method realize to expectation longitudinal acceleration tracking；The method comprises the following steps:

Step one, set up two car twisting movement models: according to the kinematic relation between controlled vehicle-to-target vehicle, Set up two car twisting movement models, simultaneously using front truck acceleration information as disturbing signal；

Step 2, the design of host controller: the two car twisting movement models set up based on step one, design a model pre- Survey controller, obtain desired two following distances according to constant time headway strategy, according to the real-time status of vehicle, utilize model pre- Survey control algolithm decision-making to go out to follow the tracks of the expectation longitudinal acceleration required for this expectation following distance；

Step 3, set up vehicle against Longitudinal Dynamic Model: wagon control operating mode is divided into driving operating mode and damped condition, to two Kind operating mode is set up vehicle according to vehicle equation respectively and is used for institute against Longitudinal Dynamic Model against Longitudinal Dynamic Model, vehicle State the instruction of the expectation acceleration that host controller calculates to be changed into desired air throttle by vehicle against Longitudinal Dynamic Model and open Degree or desired brake pedal aperture；

Step 4, the design of lower level controller: according to vehicle against Longitudinal Dynamic Model, desirably accelerate under driving operating mode Degree tries to achieve desired throttle opening, and under damped condition, acceleration desirably tries to achieve desired brake pedal aperture；To obtain The control signal obtained exports to controlled vehicle, completes the tracing control to expectation following distance.

2. the multiobject adaptive cruise control method of a kind of consideration as claimed in claim 1, it is characterised in that described The car twisting movement model that step one is set up is:

$\left\{\begin{array}{c}x(k+1)=Ax\left(k\right)+Bu\left(k\right)+Gw\left(k\right)\\ y\left(k\right)=Cx\left(k\right)\end{array}\right.$

Wherein,

$\begin{array}{cccc}A=\left[\begin{array}{ccc}1& T_s& 0\\ 0& 1& 0\\ 0& 0& 1\end{array}\right]& B=\left[\begin{array}{c}-\frac{1}{2}{T_s}^2\\ -T_s\\ T_s\end{array}\right]& G=\left[\begin{array}{c}\frac{1}{2}{T_s}^2\\ T_s\\ 0\end{array}\right]& C=\left[\begin{array}{ccc}1& 0& 0\end{array}\right]\end{array}$

U (k)=a_f(k), w (k)=a_l(k), a_l(k)、a_fK () is the acceleration information in before and after two car k moment respectively, unit m/s²；X (k)=[Δ x (k), v_ref(k),v(k)]^T, v_refK () represents the relative velocity of k moment two car, unit m/s, meet v_ref(k)=v_l(k)-v (k), v_lK () is the longitudinal velocity in front truck k moment, unit m/s；Δ x (k) is k moment two following distance, Unit m；T_sIt is the sampling period of system, unit s.

3. the multiobject adaptive cruise control method of a kind of consideration as claimed in claim 1, it is characterised in that described The design process of step 2 host controller specifically includes following steps:

1) optimization problem is proposed:

$\underset{u\left(k\right)}{min}J\left(y\right(k),u(k),m,p)$

Wherein,

$\begin{array}{c}J\left(y\right(k),u(k),m,p)\\ =\underset{i=1}{\overset{p}{\mathrm{\Σ}}}\left|\right|{\mathrm{\Γ}}_{y,i}\left(y_c\left(k+i|k\right)-r\left(k+i\right)\right)|{|}^2+\underset{i=1}{\overset{m}{\mathrm{\Σ}}}|\left|{\mathrm{\Γ}}_{u,i}u\left(k+i-1\right)\right|{|}^2\end{array}$

P is the prediction time domain of system, and m is to control time domain and m≤p；

Satisfied two car twisting movements:

$\left\{\begin{array}{c}v(k+1)=v\left(k\right)+a_f\left(k\right)T_s\\ v_{ref}(k+1)=v_{ref}\left(k\right)+a_l\left(k\right)T_s-a_f\left(k\right)T_s\\ \mathrm{\Δ}x(k+1)=\mathrm{\Δ}x\left(k\right)+v_{ref}\left(k\right)T_s+\frac{1}{2}\left(a_l\right(k)-a_f(k\left)\right)T_s^2\end{array}\right.$

Meet inequality constraints simultaneously:

$\left\{\begin{array}{c}\mathrm{\Δ}x\left(k\right)\≥d_c\\ a_{fmin}\≤a_f\left(k\right)\≤a_{fmax}\\ v_{\min }\≤v\left(k\right)\≤v_{\max }\end{array}\right.$

a_l(k)、a_fK () is the acceleration information in before and after two car k moment respectively, unit m/s²； X (k)=[Δ x (k), v_ref(k),v(k)]^T, v_refK () represents the relative velocity of k moment two car, unit m/s, meet v_ref(k)=v_l(k)-v (k), v_lK () is the longitudinal velocity in front truck k moment, unit m/s；Δ x (k) is k moment two following distance, Unit m；T_sIt is the sampling period of system, unit s；

2) solving-optimizing problem: by described step 1) optimization problem that proposes is converted into:

$\underset{U\left(k\right)}{min}U{\left(k\right)}^{T}HU\left(k\right)+G{(k+1|k)}^{T}U\left(k\right)$

s.t. C_uU(k)≥b(k+1|k)

Wherein,

H=(S_u-t_hV_u)^TΓ_y ^TΓ_y(S_u-t_hV_u)+Γ_u ^TΓ_u

G(k+1|k)^T=2E_p(k+1|k)Γ_y ^TΓ_y(S_u-t_hV_u)

$\begin{array}{cc}{C}_u=\left[\begin{array}{c}-I_{m*m}\\ I_{m*m}\\ S_u\\ V_u\\ -V_u\end{array}\right]& b\left(k+1|k\right)=\left[\begin{array}{c}-U_{\max }\\ U_{\min }\\ D_c-S_{x}x\left(k\right)-S_{w}W\left(k\right)\\ V_{\min }-V_{x}x\left(k\right)\\ V_{x}x\left(k\right)-V_{\max }\end{array}\right]\end{array}$

In Calling MATLAB, solver quadprog just can complete solving of host controller, it is thus achieved that desired longitudinal acceleration.

4. the multiobject adaptive cruise control method of a kind of consideration as claimed in claim 1, it is characterised in that described Step 3 is set up vehicle and is specifically included against Longitudinal Dynamic Model:

1) drive the vehicle of operating mode against Longitudinal Dynamic Model:

The equation formula of vehicle is set up according to Newton's second law:

$\mathrm{\δ}ma=\frac{i_g{i}_0{\mathrm{\η}}_T}{r_{eff}}{T}_e-mgf-\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2-mgsin\mathrm{\θ}$

The output moment of torsion of engine can be obtained according to above formula:

$T_e=\frac{(mgf+\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2+mgsin\mathrm{\θ}+\mathrm{\δ}ma)r_{eff}}{i_g{i}_0{\mathrm{\η}}_T}$

In formula, T_eIt is that engine expects moment of torsion, i_gIt is the gearratio of speed changer, i₀Represent main step-down ratio, η_TRepresent transmission The mechanical efficiency of system, r_effBeing the effective radius of wheel, m is complete vehicle quality, and f is coefficient of rolling resistance, C_dIt it is air drag Coefficient, A is front face area, and ρ is atmospheric density, and v is longitudinal direction of car travel speed, and θ represents road grade, and δ is automobile Correction coefficient of rotating mass, a is the longitudinal acceleration of vehicle, and g is acceleration of gravity；

Engine torque characteristic map schematic diagram in veDYNA auto model, counter tabling look-up is utilized to obtain air throttle corresponding under this moment of torsion The size of aperture, according to T_eWith engine speed ω_e, utilize engine air throttle open degree characteristic curve map, try to achieve desired solar term Door aperture α_desFor:

α_des=f (T_e,ω_e)

2) vehicle of damped condition is against Longitudinal Dynamic Model:

When braking travels, the equation of vehicle is as follows:

$\mathrm{\δ}ma=F_b+mgf+\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2+mgsin\mathrm{\θ}$

Engine braking moment expression formula can be obtained according to above formula:

$T_b=(\mathrm{\δ}ma-mgf-\frac{1}{2}{C}_d{\mathrm{A\ρv}}^2-mg\sin \mathrm{\θ})r_{eff}$

The braking moment of each wheel

The computing formula of the braking moment on each wheel meets below equation:

M_b=2 (P_b·A_b·r_b·μ_b)

In formula,

P_bRepresent the brake pressure on each wheel, unit: Pa；μ_bRepresent the coefficient of friction between brake block and brake disc；A_b Represent the contact area of friction plate and brake disc, unit: m²；r_bRepresent braking radius, unit: m；

Above-mentioned parameter is brought into the relation of brake pressure and the braking moment obtaining each wheel:

M_b=0.1323 × 10^-3·P_b

Obtain desired brake pressure P_bAfter, convert thereof into the deceleration control that desired brake pedal aperture i.e. completes under damped condition System.