CN104773173A - Autonomous driving vehicle traveling status information estimation method - Google Patents

Abstract

The invention discloses an autonomous driving vehicle traveling status information estimation method. The method comprises the steps that 1, a vehicle traveling status modular estimator is designed based on a simplified vehicle model and used for estimating other traveling status information of a vehicle according to measurement signals of a vehicle sensor, wherein the first step comprises the substeps that a longitudinal tire force observer is designed and used for estimating the longitudinal tire force of the vehicle, a lateral tire force estimator is designed and used for calculating the lateral tire force of the vehicle, a vehicle speed nonlinear full-dimensional status estimator is designed and used for estimating the longitudinal speed and the lateral speed of the vehicle, and the designed estimators designed in the substeps are integrated to obtain the vehicle traveling status modular estimator; 2, the signals measured by the sensor of the vehicle are input into the vehicle traveling status modular estimator obtained in the first step, and the estimation information of the tire force and speed of the vehicle is estimated through the vehicle traveling status modular estimator.

Description

A kind of autonomous land vehicle running condition information method of estimation

Technical field

The present invention relates to a kind of car status information method of estimation, be specifically related to a kind of autonomous land vehicle running condition information method of estimation, not only can be used for the theoretical investigation that car status information is estimated, be more suitable for the various operating modes that actual vehicle is run.

Background technology

Blocking up and accident problem for traffic, automobile controls development and is day by day tending towards information-based and intelligent, and wherein autonomous land vehicle is the development product realizing intelligent traffic network.Because automobile electric control system especially autonomous land vehicle is implement corresponding control logic according to vehicle running state mostly.Vehicle acceleration, slow down, to cruise and course changing control all depends on the size of traveling friction force between road and tire and the speed of operation of vehicle, in order to improve the traveling active safety of vehicle, accurately and in real time know that the tire force of vehicle and speed information are that autonomous land vehicle realizes motion control, completes particular path and travel the important foundation with dangerous situation early warning, and above-mentioned information generally can not utilize sensor directly to measure obtains, i.e. enable measurement, usual measuring instrument also exist measured error large with productive costs height etc. problem.Therefore, theoretical according to vehicle state estimation, the motoring condition parameter that cannot directly measure improves the study hotspot of autonomous land vehicle security technical capability to utilize common standard configuration onboard sensor metrical information to estimate.

Summary of the invention

For the problems referred to above, consider the strong nonlinearity of Vehicular system, each state variable has extremely strong coupling, the present invention is based on simplification auto model, Observer method is utilized to provide a kind of simply real-time, modular autonomous land vehicle running condition information method of estimation, then the signal of vehicle self-carrying sensor measurement is inputted integrated estimator, obtain the estimated information of vehicle tyre power and the speed of a motor vehicle.

The object of the invention is to be realized by following scheme:

A kind of autonomous land vehicle running condition information method of estimation, comprises the following steps:

Step one, based on simplification auto model, design vehicle motoring condition modularization estimator, it, for estimating vehicle other running condition information according to onboard sensor measurement signal, comprises following process:

1) longitudinal tire force observer is designed, for estimating longitudinal direction of car tire force;

2) side direction tire force estimator is designed, for calculating vehicle side to tire force;

3) the non-linear full dimension state estimator of design speed, for estimating longitudinal direction of car and the side direction speed of a motor vehicle;

4) each estimator designed by above-mentioned steps is carried out integrated, after integrated, obtain vehicle running state modularization estimator;

Step 2, the vehicle running state modularization estimator input of the signal of vehicle self-carrying sensor measurement obtained by step one, estimate the estimated information of vehicle tyre power and the speed of a motor vehicle by vehicle running state modularization estimator.

Described a kind of autonomous land vehicle running condition information method of estimation, the process 1 of step one) process of design longitudinal tire force observer is:

In conjunction with wheel rolling kinetics equation, build Unknown Input Observer to each wheel of vehicle and estimate longitudinal tire force, the interference estimator of longitudinal direction of car tire force is shown with following formula table:

$\left\{\begin{array}{c}{\hat{F}}_{\mathrm{xi}}=-\frac{J}{R_e}({\mathrm{\χ}}_i+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)\\ {\stackrel{\·}{\mathrm{\χ}}}_i=-{\mathrm{\ρ}}_i(\frac{T_i}{J}+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)-{\mathrm{\ρ}}_i{\mathrm{\χ}}_i\end{array}\right.$

Wherein, be the longitudinal tire force estimated valve of i-th wheel, unit N, J are vehicle wheel rotation inertia, units/kg * m ², R _efor tire effective radius, unit m, ω _ibe i-th wheel wheel speed, unit m/s, T _ibe that i-th wheel drives the summation with lock torque, unit N*m, k, ρ _ifor feedback gain.

Described a kind of autonomous land vehicle running condition information method of estimation, the process 2 of step one) process of design side direction tire force estimator is:

Utilize Dugoff tire model calculation side to tire force, the Dugoff tire model under vertical cunning-lateral deviation combinational acting can be represented with the nonlinear function of following formula:

$\left\{\begin{array}{c}{F}_{\mathrm{xi}}=f_{\mathrm{xi}}(V_x,V_y,a_x,a_y,r,{\mathrm{\ω}}_i,\mathrm{\δ})\\ F_{\mathrm{yi}}=f_{\mathrm{yi}}(V_x,V_y,a_x,a_y,r,{\mathrm{\ω}}_i,\mathrm{\δ})\end{array}\right.$

Wherein, F _xi/ F _yifor longitudinal direction/side direction tire force, unit N, V _x/ V _yfor longitudinal direction/side direction speed of a motor vehicle, unit m/s, a _x/ a _yfor longitudinal direction/lateral acceleration, unit m/s ², r is yaw velocity, unit rad/s, ω _ifor wheel speed, unit m/s, δ are steering wheel angle, unit rad.

Described a kind of autonomous land vehicle running condition information method of estimation, the process 3 of step one) the non-linear full dimension state estimator of design speed comprises following process:

The auto model of 1 carries out force analysis by reference to the accompanying drawings, obtains vehicle dynamic model to be:

$\left\{\begin{array}{c}\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{F}_{\mathrm{xi}}={\mathrm{ma}}_x=m({\stackrel{\·}{V}}_x-r{V}_y)\\ \underset{i=1}{\overset{4}{\mathrm{\Σ}}}{F}_{\mathrm{yi}}={\mathrm{ma}}_y=m({\stackrel{\·}{V}}_y+r{V}_x)\\ \underset{i=1}{\overset{4}{\mathrm{\Σ}}}{M}_{\mathrm{zi}}=I_z\stackrel{\·}{r}\end{array}\right.$

In above formula,

$\begin{array}{c}\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{F}_{\mathrm{xi}}=(F_{x_1}+F_{x_2})\cos {\mathrm{\δ}}_w-(F_{y_1}+F_{y_2})\sin {\mathrm{\δ}}_w+(F_{x_3}+F_{x_4})\\ \underset{i=1}{\overset{4}{\mathrm{\Σ}}}{F}_{\mathrm{yi}}=(F_{x_1}+F_{x_2})\sin {\mathrm{\δ}}_w+(F_{y_1}+F_{y_2})\cos {\mathrm{\δ}}_w+(F_{y_3}+F_{y_4})\\ \underset{i=1}{\overset{4}{\mathrm{\Σ}}}{M}_{\mathrm{zi}}=l_F(F_{x_1}+F_{x_2})\sin {\mathrm{\ϵ}}_w+l_F(F_{y_1}+F_{y_2})\cos {\mathrm{\δ}}_w+\frac{b_F}{2}(F_{x_2}-F_{x_1})\cos {\mathrm{\δ}}_w\\ +\frac{b_R}{2}(F_{x_4}-F_{x_3})-\frac{b_F}{2}(F_{y_2}-F_{y_1})\sin {\mathrm{\δ}}_w-l_R(F_{y_3}+F_{y_4})\end{array}$

Wherein, m is complete vehicle quality, units/kg, V _x/ V _yfor longitudinal direction of car/side velocity under bodywork reference frame, unit m/s, l _f/ l _rfor vehicle centroid is to the distance of front/rear axle, unit m, b _f/ b _rfor vehicle wheel base, unit m, front wheel steering angle δ _w=δ _w1=δ _w2=δ/I _sw, δ is steering wheel angle, unit rad, I _swfor steering gear ratio, rear-axle steering angle δ _w3=δ _w4=0, unit rad, r are Vehicular yaw cireular frequency, and unit rad/s, g are gravity constant, F _xi/ F _yiwhat be respectively four tires indulges/side direction tire force, unit N, I _zfor car body is around the rotor inertia of z-axis, units/kg * m ², M _zifor the rotating torque that each tire rotates around z-axis, unit N*m.

To sum up, non-linear full dimension speed of a motor vehicle estimator is:

$\left\{\begin{array}{c}{\stackrel{\·}{\hat{V}}}_x=r{\hat{V}}_y+a_x+\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{K}_i(v_{\mathrm{xi}}-{\hat{V}}_x)\\ {\stackrel{\·}{\hat{V}}}_y=-r{\hat{V}}_x+a_y+K_y({\mathrm{ma}}_y-{\hat{F}}_y)\\ \stackrel{\·}{\hat{r}}=\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{M}_{\mathrm{zi}}/I_z+K_r(r-\hat{r})\end{array}\right.$

Wherein,

${\stackrel{\·}{\hat{V}}}_x=r{\hat{V}}_y+a_x+\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{K}_i(v_{\mathrm{xi}}-{\hat{V}}_x)$ For longitudinal direction of car speed of a motor vehicle estimator;

${\stackrel{\·}{\hat{V}}}_y=-r{\hat{V}}_x+a_y+K_y({\mathrm{ma}}_y-{\hat{F}}_y)$ For vehicle side is to speed of a motor vehicle estimator;

for Vehicular yaw angular velocity estimator;

K _i, K _y, K _rbe respectively the estimator gain of the longitudinal direction of car speed of a motor vehicle, the side direction speed of a motor vehicle and yaw velocity.

In order to verify validity and the practicality of vehicle running state estimator of the present invention, take off data under the different operating modes adopting autonomous land vehicle red flag HQ430 car test to collect is as estimator input and comparative information, wherein, steering wheel angle, gear, engine speed information adopt the measurement of existing vehicle-mounted CAN bus to obtain, speed of a motor vehicle employing model is costly that the vehicle speedometer measurement of DEWE-VGPS-200C obtains, and the gyroscope survey that longitudinal direction of car/lateral acceleration and yaw velocity are then NAV420CA-100 by model obtains.Because these measurement signals derive from different measuring instruments, application is difficult to keep synchronous consistent, therefore, the present invention has carried out some conversion process while the above-mentioned three kinds of apparatus measures signals of collection, make above-mentioned signal can simultaneously synchronism output, whole acquisition of signal schematic diagram as shown in Figure 2.

Compared with prior art, the present invention has the following advantages:

(1) a kind of vehicle tyre power based on nonlinear observer and speed of a motor vehicle modularization estimation scheme is proposed;

(2) the alternative specific onboard sensor of estimator, reduces vehicle production cost;

(3) for the different driving cycle of real vehicle, the estimated information of vehicle tyre power, longitudinal direction and the side direction speed of a motor vehicle can reliably be obtained in real time.

Accompanying drawing explanation

Fig. 1 is that stressed schematic diagram overlooked by vehicle.

Fig. 2 is acquisition of signal schematic diagram.

Fig. 3 is wheel kinetic model.

Fig. 4 is longitudinal tire force estimator structure.

Fig. 5 is vehicle running state modularization estimator structured flowchart.

Fig. 6 is the estimation contrast block diagram based on HQ430 real vehicle data.

Fig. 7 is snakelike bar operating mode input information. (a) steering wheel angle (b) wheel speed (c) wheel driving torque.

Fig. 8 is that under snakelike bar working condition measuring information, output state estimates correlation curve figure. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Fig. 9 is output state evaluated error diagram of curves under snakelike bar working condition measuring information. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Figure 10 is stable state circumference operating mode input information. (a) steering wheel angle (b) wheel speed (c) wheel driving torque.

Figure 11 is that under stable state circumference working condition measuring information, output state estimates correlation curve figure. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Figure 12 is output state evaluated error diagram of curves under stable state circumference working condition measuring information. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Figure 13 is bearing circle angle step operating mode input information. (a) steering wheel angle (b) wheel speed (c) wheel driving torque.

Figure 14 is that under the step working condition measuring information of bearing circle angle, output state estimates correlation curve figure. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Figure 15 is output state evaluated error diagram of curves under the step working condition measuring information of bearing circle angle. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

District's operating mode input information centered by Figure 16. (a) steering wheel angle (b) wheel speed (c) wheel driving torque.

Centered by Figure 17, under district's working condition measuring information, output state estimates correlation curve figure. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Output state evaluated error diagram of curves under district's working condition measuring information centered by Figure 18. (a) the longitudinal speed of a motor vehicle (b) yaw velocity (c) longitudinal acceleration (d) lateral acceleration.

Detailed description of the invention

Below in conjunction with accompanying drawing, the present invention is further described with implementing operating mode.

The invention provides a kind of modular autonomous land vehicle motoring condition method of estimation, comprise the following steps:

Step one, based on simplification auto model, design vehicle motoring condition modularization estimator:

1. longitudinal tire force Design of Observer:

In conjunction with single-wheel rolling kinetic model, as shown in Figure 3.The kinetics equation of wheel can represent with formula (1):

$J{\stackrel{\·}{\mathrm{\ω}}}_i=T_i-R_e{F}_{\mathrm{xi}},(i=1,...,4)---\left(1\right)$

Wherein, J is vehicle wheel rotation inertia, units/kg * m ², ω _ifor wheel speed, unit m/s, T _ifor wheel drives the summation with lock torque, unit N*m, R _efor tire effective radius, unit m, F _xibe the longitudinal tire force of i-th wheel, unit N.

In the design of longitudinal tire force estimator, choose real-time, calculated amount is little, the Unknown worm nonlinear estimator of strong interference immunity, groundwork produces state estimation error, in equation of state, carry out decoupling zero interference by reconstruct unknown input disturbances (or noise) variable, eliminate the impact of unknown disturbances.

Longitudinal tire force estimator structure as shown in Figure 4.According to formula (1), choose wheel driving/braking moment and wheel speed signal as estimator input.Wheel torque signal can calculate according to the pressure signal of measured engine torque and rotating speed and wheel cylinder.Here for the ease of analyzing and describe vehicle tyre power estimation problem, suppose that wheel driving/braking torque signals can directly obtain.

Unknown Input Observer is built to each wheel of vehicle and estimates longitudinal tire force:

$\left\{\begin{array}{c}{\stackrel{\·}{\hat{F}}}_{\mathrm{xi}}=K({\stackrel{\·}{\mathrm{\ω}}}_i-{\stackrel{\·}{\widehat{\mathrm{\ω}}}}_i)\\ {\stackrel{\·}{\widehat{\mathrm{\ω}}}}_i=\frac{T_i}{J}-\frac{R_e}{J}{\hat{F}}_{\mathrm{xi}}\end{array}\right.---\left(2\right)$

According to the Direct dynamics relation of the differential of longitudinal tire force and wheel speed, by the differential of wheel speed as correcting value, feedback gain is K, and namely this Interference Estimation amount is proportional to the error between estimator state and system time of day.

For avoiding measurement signal ω _icarry out the evaluated error that differential brings, if arithmetic number ρ _i, ordinary differential equation variable χ _i, order then the estimator of longitudinal direction of car tire force can be converted into such as formula shown in (3):

${\hat{F}}_{\mathrm{xi}}=-\frac{J}{R_e}({\mathrm{\χ}}_i+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)---\left(3\right)$

Longitudinal tire force as an exogenous disturbances item, variable χ _idynamic characteristics as follows:

$\begin{array}{c}{\stackrel{\·}{\mathrm{\χ}}}_i=-{\mathrm{\ρ}}_i{\stackrel{\·}{\widehat{\mathrm{\ω}}}}_i\\ =-{\mathrm{\ρ}}_i(\frac{T_i}{J}-\frac{R_e}{J}{\hat{F}}_{\mathrm{xi}})\\ =-{\mathrm{\ρ}}_i\frac{T_i}{J}+{\mathrm{\ρ}}_i\frac{R_e}{J}[-\frac{J}{R_e}({\mathrm{\χ}}_i+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)]\\ =-{\mathrm{\ρ}}_i\frac{T_i}{J}-{{\mathrm{\ρ}}^2}_i{\mathrm{\ω}}_i-{\mathrm{\ρ}}_i{\mathrm{\χ}}_i\\ =-{\mathrm{\ρ}}_i(\frac{T_i}{J}+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)-{\mathrm{\ρ}}_i{\mathrm{\χ}}_i\end{array}---\left(4\right)$

In conjunction with wheel rolling kinetics equation, the dynamic characteristics of longitudinal tire force estimated valve can be expressed as formula (5):

${\stackrel{\·}{\hat{F}}}_{\mathrm{xi}}={\mathrm{\ρ}}_i(F_{\mathrm{xi}}-{\hat{F}}_{\mathrm{xi}})---\left(5\right)$

In sum, the final expression-form of the interference estimator of longitudinal direction of car tire force is such as formula shown in (6):

$\left\{\begin{array}{c}{\hat{F}}_{\mathrm{xi}}=-\frac{J}{R_e}({\mathrm{\χ}}_i+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)\\ {\stackrel{\·}{\mathrm{\χ}}}_i=-{\mathrm{\ρ}}_i(\frac{T_i}{J}+{\mathrm{\ρ}}_i{\mathrm{\ω}}_i)-{\mathrm{\ρ}}_i{\mathrm{\χ}}_i\end{array}\right.---\left(6\right)$

In this exogenous disturbances state estimator, as the state variable not track channel state of distracter, namely Interference Estimation amount is dynamically dynamic independent of system state estimation, and the structure of this interference estimator can systematically be expanded, and can expand higher order term more needed for multisystem in future.Choose liapunov function when estimating system is stablized, have convolution (5), the boundary obtaining the evaluated error of longitudinal tire force is:

$|F_{\mathrm{xi}}-{\hat{F}}_{\mathrm{xi}}|\≤\frac{{\left|{\stackrel{\·}{F}}_{\mathrm{xi}}\right|}_{\max }}{{\mathrm{\ρ}}_i}---\left(7\right)$

2. side direction tire force estimator design:

For the tire force of Dugoff tire model under vertical cunning-lateral deviation combinational acting, this power is about tire coefficient of friction, straight skidding rate, longitudinal direction and lateral rigidity, and the nonlinear function of the primary variables such as vertical load, is shown below:

$\left\{\begin{array}{c}{F}_{\mathrm{xi}}=f_{\mathrm{xi}}^{\′}(S_{\mathrm{xi}},C_x,C_y,{\mathrm{\α}}_i,F_{\mathrm{zi}},\mathrm{\μ},...)\\ F_{\mathrm{yi}}=f_{\mathrm{yi}}^{\′}(S_{\mathrm{xi}},C_x,C_y,{\mathrm{\α}}_i,F_{\mathrm{zi}},\mathrm{\μ},...)\end{array}\right.---\left(8\right)$

Wherein, S _xirepresent tire straight skidding rate, v _xifor tire longitudinal velocity, unit m/s, C _x, C _yrepresent tire longitudinal direction and lateral rigidity, unit N/rad, α _irepresent tyre slip angle, unit rad, F _zifor each tire vertical force, unit N, μ are tire surface friction coefficient.

Then the computing formula of side direction tire force is:

$F_{\mathrm{yi}}=\frac{C_y\tan {\mathrm{\α}}_i}{1+S_{\mathrm{xi}}}f\left({\mathrm{\λ}}_i\right)---\left(9\right)$

Wherein, variable λ _iand function f (λ _i) be expressed as:

$\begin{array}{c}{\mathrm{\&lambda;}}_i=\frac{\mathrm{\&mu;}{F}_{\mathrm{zi}}(1+S_{\mathrm{xi}})}{2\sqrt{{\left(C_x{S}_{\mathrm{xi}}\right)}^2+{\left(C_y{\mathrm{\&alpha;}}_i\right)}^2}}\\ f\left({\mathrm{\&lambda;}}_i\right)=\left\{\begin{array}{cc}1,& {\mathrm{\&lambda;}}_i\&GreaterEqual;1\\ (2-{\mathrm{\&lambda;}}_i){\mathrm{\&lambda;}}_i,& {\mathrm{\&lambda;}}_i<1\end{array}\right.\end{array}---\left(10\right)$

Each tire vertical force F _zicomputing formula be:

$\begin{array}{c}{F}_{z1}=\frac{m(l_{R}g-{\mathrm{ha}}_x)}{2(l_F+l_R)}-\frac{\mathrm{mg}{a}_y(l_{R}g-{\mathrm{ha}}_x)}{(l_F+l_R)b_{F}g}\\ F_{z2}=\frac{m(l_{R}g-{\mathrm{ha}}_x)}{2(l_F+l_R)}+\frac{{\mathrm{mha}}_y(l_{R}g-{\mathrm{ha}}_x)}{(l_F+l_R)b_{F}g}\\ F_{z3}=\frac{m(l_{F}g+{\mathrm{ha}}_x)}{2(l_F+l_R)}-\frac{{\mathrm{mha}}_y(l_{F}g+{\mathrm{ha}}_x)}{(l_F+l_R)b_{F}g}\\ F_{z4}=\frac{m(l_{F}g+{\mathrm{ha}}_x)}{2(l_F+l_R)}+\frac{{\mathrm{mha}}_x(l_{F}g+{\mathrm{ha}}_x)}{(l_F+l_R)b_{F}g}\end{array}---\left(11\right)$

Wherein, m is complete vehicle quality, units/kg, and h is vehicle centroid height, unit m, l _f/ l _rfor vehicle centroid is to the distance of front/rear axle, unit m, b _ffor vehicle rear wheel distance, unit m, a _x/ a _yfor longitudinal direction/lateral acceleration, unit m/s ², g is acceleration due to gravity, tyre slip angle α _ibe expressed as:

$\begin{array}{c}{\mathrm{\&alpha;}}_1={\mathrm{\&delta;}}_w-\arctan \left(\frac{V_y+l_{F}r}{V_x-\frac{b_F}{2}r}\right)\\ {\mathrm{\&alpha;}}_2={\mathrm{\&delta;}}_w-\arctan \left(\frac{V_y+l_{F}r}{V_x+\frac{b_F}{2}r}\right)\\ {\mathrm{\&alpha;}}_3={\mathrm{\&delta;}}_w-\arctan \left(\frac{V_y-l_{R}r}{V_x-\frac{b_R}{2}r}\right)\\ {\mathrm{\&alpha;}}_4={\mathrm{\&delta;}}_w-\arctan \left(\frac{V_y-l_{R}r}{V_x+\frac{b_R}{2}r}\right)\end{array}---\left(12\right)$

In formula, δ _wfor tire corner, unit rad, V _x/ V _yfor longitudinal direction/side direction speed of a motor vehicle, unit m/s, r are yaw velocity, unit rad/s, b _f/ b _rfor vehicle wheel base, unit m.

Finally, the Dugoff tire model (8) under vertical cunning-lateral deviation combinational acting can be represented with the nonlinear function of formula (13) again, δ is steering wheel angle, unit rad.

$\left\{\begin{array}{c}{F}_{x_i}=f_{x_i}(V_x,V_y,a_x,a_y,r,{\mathrm{\&omega;}}_i,\mathrm{\&delta;})\\ F_{y_i}=f_{y_i}(V_x,V_y,a_x,a_y,r,{\mathrm{\&omega;}}_i,\mathrm{\&delta;})\end{array}\right.---\left(13\right)$

3. the non-linear omnidirectional vision design of the speed of a motor vehicle:

Stressed schematic diagram overlooked by the seven freedom vehicle of 1 with reference to the accompanying drawings, and force analysis obtains vehicle dynamic model such as formula shown in (14).

$\left\{\begin{array}{c}\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{xi}}={\mathrm{ma}}_x=m({\stackrel{\&CenterDot;}{V}}_x-r{V}_y)\\ \underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{yi}}={\mathrm{ma}}_y=m({\stackrel{\&CenterDot;}{V}}_y+r{V}_x)\\ \underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}=I_z\stackrel{\&CenterDot;}{r}\end{array}\right.---\left(14\right)$

In formula, the longitudinal direction of four tires is made a concerted effort, resulting side force and the conjunction rotating torque expression formula of rotating around z-axis be such as formula shown in (15).

$\begin{array}{c}\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{xi}}=(F_{x_1}+F_{x_2})\cos {\mathrm{\&delta;}}_w-(F_{y_1}+F_{y_2})\sin {\mathrm{\&delta;}}_w+(F_{x_3}+F_{x_4})\\ \underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{yi}}=(F_{x_1}+F_{x_2})\sin {\mathrm{\&delta;}}_w+(F_{y_1}+F_{y_2})\cos {\mathrm{\&delta;}}_w+(F_{y_3}+F_{y_4})\\ \underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}=l_F(F_{x_1}+F_{x_2})\sin {\mathrm{\&epsiv;}}_w+l_F(F_{y_1}+F_{y_2})\cos {\mathrm{\&delta;}}_w+\frac{b_F}{2}(F_{x_2}-F_{x_1})\cos {\mathrm{\&delta;}}_w\\ +\frac{b_R}{2}(F_{x_4}-F_{x_3})-\frac{b_F}{2}(F_{y_2}-F_{y_1})\sin {\mathrm{\&delta;}}_w-l_R(F_{y_3}+F_{y_4})\end{array}---\left(15\right)$

Wherein, m is complete vehicle quality, units/kg, V _x/ V _yfor longitudinal direction of car/side velocity under bodywork reference frame, unit m/s, l _f/ l _rfor vehicle centroid is to the distance of front/rear axle, unit m, b _f/ b _rfor vehicle wheel base, unit m, front wheel steering angle δ _w=δ _w1=δ _w2=δ/I _sw, δ is steering wheel angle, unit rad, I _swfor steering gear ratio, rear-axle steering angle δ _w3=δ _w4=0, unit rad, r are Vehicular yaw cireular frequency, and unit rad/s, g are gravity constant, F _xi/ F _yiwhat be respectively four tires indulges/side direction tire force, unit N, I _zfor car body is around the rotor inertia of z-axis, units/kg * m ², M _zifor the rotating torque that each tire rotates around z-axis, unit N*m.

The non-linear full dimension state estimator design method of the described speed of a motor vehicle is as follows:

Choose the correcting value of coarse value as longitudinal speed of a motor vehicle estimator of the vehicular longitudinal velocity that can be obtained by wheel speed indirect calculation, ignore tyre side inclination angle now, suppose that its value is zero.Then for tire i (i=1 ..., 4) core wheel speed can be expressed as at the component of x-axis:

v _i,x＝R _eω _icosδ _w(16)

According to the algebraic relation between tire core wheel speed and car speed, the coarse value v of vehicular longitudinal velocity can be obtained _xishown in (17),

$\begin{array}{c}{v}_{x1}=R_e{\mathrm{\&omega;}}_1\cos {\mathrm{\&delta;}}_w+\frac{b_F}{2}r\\ v_{x2}=R_e{\mathrm{\&omega;}}_2{\cos \mathrm{\&delta;}}_w-\frac{b_F}{2}r\\ v_{x3}=R_e{\mathrm{\&omega;}}_3+\frac{b_R}{2}r\\ v_{x4}=R_e{\mathrm{\&omega;}}_4-\frac{b_R}{2}r\end{array}---\left(17\right)$

In conjunction with above-mentioned simplification vehicle dynamic model, by the longitudinal speed of a motor vehicle observed reading v by calculating _xiwith estimated valve do difference, deviation, as the correction term of longitudinal speed of a motor vehicle estimator, can the longitudinal speed of a motor vehicle estimator of design vehicle be:

${\stackrel{\&CenterDot;}{\hat{V}}}_x=r{\hat{V}}_y+a_x+\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{K}_i(v_{\mathrm{xi}}-{\hat{V}}_x)---\left(18\right)$

In formula, longitudinal speed of a motor vehicle estimator gain K _i, (i=1,2,3,4) are zonal cooling and bounded, and its value reflects observed reading v _xiwith estimated valve degree of closeness.

After designing longitudinal speed of a motor vehicle estimator, by longitudinal vehicle velocity V _xestimated valve as system known quantity, be input in side direction speed of a motor vehicle estimator, utilize lateral acceleration information a simultaneously _ycalculate vehicle body lateral force ma _yas observed reading, utilize tire force estimated valve to calculate, have the vehicle body resulting side force of Direct dynamics relation with the side direction speed of a motor vehicle as estimated valve, and by the two deviation as correction term, side direction speed of a motor vehicle estimator can be constructed:

${\stackrel{\&CenterDot;}{\hat{V}}}_y=-r{\hat{V}}_x+a_y+K_y({\mathrm{ma}}_y-{\hat{F}}_y)---\left(19\right)$

Wherein K _yit is the gain of adjustable side direction speed of a motor vehicle estimator.

Although the value of yaw velocity directly can be measured by sensor and obtain, as can be seen from vehicle dynamics, the yaw velocity r of vehicle and longitudinal vehicle velocity V _xand side direction vehicle velocity V _yclose-coupled, is necessary to estimate in the lump yaw velocity.

By yaw velocity observed reading r and estimated valve between deviation as correction term, yaw velocity estimator can be constructed:

$\stackrel{\&CenterDot;}{\hat{r}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}/I_z+K_r(r-\hat{r})---\left(20\right)$

To sum up designing non-linear full dimension speed of a motor vehicle estimator is

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{\hat{V}}}_x=r{\hat{V}}_y+a_x+\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{K}_i(v_{\mathrm{xi}}-{\hat{V}}_x)\\ {\stackrel{\&CenterDot;}{\hat{V}}}_y=-r{\hat{V}}_x+a_y+K_y({\mathrm{ma}}_y-{\hat{F}}_y)\\ \stackrel{\&CenterDot;}{\hat{r}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}/I_z+K_r(r-\hat{r})\end{array}\right.---\left(21\right)$

Wherein K _i, K _y, K _rbe respectively the estimator gain of the longitudinal direction of car speed of a motor vehicle, the side direction speed of a motor vehicle and yaw velocity.

4. vehicle running state modularization estimator is integrated:

To above vehicle tyre power and the modeling in MATLAB/simulink of speed of a motor vehicle observer integrated, the structure of the vehicle running state modularization estimator after integrated is as shown in Figure 5.

Step 2, the vehicle running state modularization estimator input of the signal of vehicle self-carrying sensor measurement obtained by step one, estimate the estimated information of vehicle tyre power and the speed of a motor vehicle by vehicle running state modularization estimator.

The input of longitudinal tire force estimator is wheel torque signal and wheel speed signal, estimates the longitudinal force obtaining each tire; The incoming signal of side direction tire force estimator comprises steering wheel angle, wheel speed, longitudinal direction and lateral acceleration, the speed of a motor vehicle and yaw velocity, obtains front and back wheel side direction tire and makes a concerted effort; Speed of a motor vehicle estimator is with tire force estimated valve for input information, and adopt cascade structure, the sensor signal that longitudinal speed of a motor vehicle estimator needs comprises steering wheel angle, wheel speed, longitudinal direction and lateral acceleration and yaw velocity information; The sensor signal that side direction speed of a motor vehicle estimator needs comprises steering wheel angle, lateral acceleration and yaw velocity.For the automobile of band ESP configuration, the input measurement signal that estimator of the present invention needs is rational for real train test.The mutual transmission that there is output signal between longitudinal direction of car, side direction tire force and speed of a motor vehicle estimator module be coupled, final export vehicle travel required for speed of a motor vehicle estimated information.

The process of real train test data:

Ignore the inclination of vehicle and the impact of luffing, suppose that in the process of estimation vehicle running state, vehicle travels on straight road, the sensor input signal obtained meets survey precision requirement, and the friction coefficient of road surface/tire is known.The red flag HQ430 car parameter values of vehicle running state modularization estimator design ap-plication is as shown in table 1.

Table 1

Parameter	Symbol	Value	Unit
				Complete vehicle quality	m	2160	[kg]
Vehicle centroid is to front axle distance	l F	1.5	[m]
				Vehicle centroid is to rear axle distance	l R	1.35	[m]
Front wheel tread	b F	1.535	[m]
				Rear track	b R	1.535	[m]
Vehicle centroid is to ground distance	h	0.568	[m]
				Vehicle wheel rotation inertia	J	1.3	[kg*m 2]
Vehicle is around the rotor inertia of z-axis	I z	3411.52	[kg*m 2]
				The effective radius of tire	R e	0.32	[m]
Longitudinal cornering stiffness	C x	137920	[N/rad]
				Horizontal cornering stiffness	C y	87594	[N/rad]

In order to verify validity and the practicality of vehicle running state estimator of the present invention, take off data under the different operating modes adopting autonomous land vehicle red flag HQ430 car test to collect is as estimator input and comparative information, wherein, steering wheel angle, Das Gaspedal aperture, gear, engine speed information adopts the measurement of existing vehicle-mounted CAN bus to obtain, speed of a motor vehicle employing model is costly that the vehicle speedometer measurement of DEWE-VGPS-200C obtains, the gyroscope survey that longitudinal direction of car/lateral acceleration and yaw velocity are then NAV420CA-100 by model obtains.The present invention has carried out some conversion process while the above-mentioned three kinds of apparatus measures signals of collection, makes above-mentioned signal can simultaneously synchronism output, and whole acquisition of signal schematic diagram as shown in Figure 2.

In Fig. 2, the vehicle power (storage battery) of 12V provides power line voltage for DEWE-VGPS-200C type vehicle speedometer and NAV420CA-100 type gyroscope, but the speed of a motor vehicle obtained due to the measurement of DEWE-VGPS-200C type vehicle speedometer is analog signal, and the longitudinal acceleration that NAV420CA-100 type gyroscope survey obtains, the steering wheel angle that lateral acceleration and yaw velocity and CAN measurement obtain, gear, accelerator open degree, wheel speed and brake-pressure are digital signal, in order to synchronism output metrical information, the analog signal speed of a motor vehicle is inputed to homemade data actuation to be converted to digital signal and to be sent in CAN, CAN card output afterwards through being integrated with CAN translation-protocol obtains the speed of a motor vehicle, bearing circle, gear, accelerator open degree, wheel speed and these 6 numerical informations of brake-pressure.Finally, in conjunction with acceleration signal, all numerical informations measured are realized real-time synchronization collection and monitoring by the LabviewSKY system on computer, here, the power line voltage of LabviewSKY system should be 220V, and this voltage is provided through inverter conversion by the vehicle power of 12V.So, the metrical information needed for real vehicle realistic model of the present invention is all that synchronous acquisition obtains.

Utilize professional Origin software to the test figures smoothing filtration treatment again after process, then the data after application smoothly calculate accordingly, special version, the shake of steering wheel angle signal is here quantizing error (sampled by A/D and bring), without the need to processing.

Vehicle running state modularization estimator compliance test result:

Fig. 6 describes the estimation effect contrast block diagram of vehicle running state modularization estimator of the present invention based on HQ430 real vehicle data.Because vehicle test data are limited, when verifying the estimation effect of vehicle state estimation device in the present invention, measure the HQ430 realistic model output signal that incoming signal combines red flag HQ430 car actual tests data and builds based on vehicle emulation dynamics software veDYNA.Wheel speed ω _iwith wheel torque signal T _iemulate kinetic model output by HQ430 vehicle to obtain.Test condition is limited, the side direction speed of a motor vehicle and tire force test figures directly wouldn't be measured by real vehicle and obtain, in conjunction with before describe vehicle dynamic model relational expression, here the comparative information choosing the longitudinal acceleration with Direct dynamics relation in real vehicle data represents the estimated result of longitudinal tire force, the comparative information of lateral acceleration represents the estimated result of the side direction speed of a motor vehicle and side direction tire force, and experimental result equally can show the estimation effect of vehicle running state modularization estimator.

Match with real vehicle operating mode when measuring test figures for making emulation operating mode, the coupling of more than 2000 group parameters is carried out in veDYNA parameter gui interface, acquisition channel comprises: directly read bibliography correlation parameter, carry out the stand tests such as suspension K/C characteristic test, tire characteristics test and obtain fore suspension and rear suspension K/C characteristic parameter and tire characteristics parameter, utilize test figures identified parameters etc., obtain the complete emulation kinetic model of red flag HQ430 car.The steering wheel angle that mode input adopts HQ430 real train test to record and vehicle speed data, by the speed information inputted adopt the unified throttle thought in pilot model to carry out acceleration pedal that modeling obtains cooking up and brake pedal signal.

Known to respective country test criteria, the test of various reflection vehicle side to performance and the collection of inputoutput data has been carried out in conjunction with actual, comprise: the Pylon course slalom test evaluating the road-holding property, steering effort size etc. of vehicle, the stable state circumference test of reflection vehicle stable state yaw response, the bearing circle a step input test of vehicle transient state response under reflection mensuration time domain, reflection vehicle is in the autocentre district maneuvering test etc. of the maneuvering performance of center.Arrange the model accuracy obtaining HQ430 emulation auto model as shown in table 2.

Table 2

Data show that set up vehicle dynamic model is high-fidelity, substantially can simulate autonomous land vehicle red flag HQ430 car performance for various researchs such as estimation, controls, the improvement of its result of study to true red flag vehicle also has certain directive significance and using value.

Respectively by Pylon course slalom test, the test of stable state circumference, the test of bearing circle a step input, maneuvering test data in autocentre district are input in vehicle-state modularization estimator, obtain the estimation effect checking curve of the vehicle running state modularization estimator designed by several groups.

Through debugging, choosing integrated estimator gain is ρ _i=1500, K _i=0.75, K _y=-1/ (0.5*2160), K _r=50.

Vehicle travels at asphalt coating, and Fig. 7 gives the steering wheel angle of vehicle running state modularization estimator under the snakelike bar operating mode of vehicle, wheel speed and wheel driving torque and inputs information.It is 30s to 42s that test results intercepts the time, now vehicle speed of operation is stabilized in 60km/h, sideway movement enters snakelike bar operating mode, from curve, now the left and right directions dish corner of driver's operation is 50 degree, and four-wheel wheel speed is near 55m/s, back-wheel drive torque is 100-300N*m, now input real vehicle metrical information, obtain characteristic signal vehicle speed, yaw state and acceleration estimation curve as shown in Figure 8 through integrated state estimator, corresponding evaluated error curve as shown in Figure 9.Now lateral acceleration trial value reaches more than 0.4g, be in limit operating condition, and estimation curve and error effects are still more stable, the maximum estimated error obtaining the longitudinal direction of car speed of a motor vehicle is 0.25m/s, the maximum estimated error of yaw velocity is 2deg/s, the longitudinal acceleration evaluated error of reflection longitudinal tire force information is 0.7m/s^2 to the maximum, the lateral acceleration estimation curve of reflection side direction tire force information and side direction vehicle speed signal comparatively trial value curve shifts to an earlier date, evaluated error is 4m/s^2 to the maximum, but tracking performance is good.Error amount increases to some extent along with the increase of trial value, and estimation curve can follow the tracks of the test figures of reality well, shows that this integrated estimator has good estimated performance for vehicle maneuverability, steering effort size.

Figure 10 gives the steering wheel angle of vehicle running state modularization estimator under vehicle stable state circumference operating mode, wheel speed and wheel driving torque and inputs information.It is 20s to 69.29s that test results intercepts the time, now vehicle speed of operation is low speed acceleration mode 10-40km/h, sideway movement enters stable state circumference operating mode, from curve, now the steering wheel angle of driver's operation is fixed as 159 degree, four-wheel wheel speed is 7-35m/s, back-wheel drive torque is 100-300N*m, now input real vehicle metrical information, characteristic signal vehicle speed is obtained through integrated state estimator, yaw state and acceleration estimation curve are as shown in figure 11, corresponding evaluated error curve as shown in figure 12, lateral acceleration maxim reaches 7m/s^2, the maximum estimated error obtaining the longitudinal direction of car speed of a motor vehicle is 0.3m/s, the maximum estimated error of yaw velocity is 4deg/s, the longitudinal acceleration evaluated error of reflection longitudinal tire force information is 0.8m/s^2 to the maximum, the evaluated error of the lateral acceleration of reflection side direction tire force information and side direction vehicle speed signal is 1.6m/s^2 to the maximum, lateral wind is travelled because take off data is subject to vehicle, the impact of the factors such as road grade, there is certain measurement noises, the single step evaluated error of yaw velocity and acceleration/accel is larger, and evaluated error increases to some extent along with the increase of simulation value, but the error of single step increase can not on estimation effect particularly speed of a motor vehicle estimation effect produce large impact, overall estimation curve can follow the tracks of the test figures of reality, show that this integrated estimator has good estimated performance for vehicle stable state yaw response.

Figure 13 gives the steering wheel angle of vehicle running state modularization estimator under the step operating mode of bearing circle angle, wheel speed and wheel driving torque and inputs information.It is 40s to 46.69s that test results intercepts the time, now vehicle speed of operation is stabilized conditions 80km/h, sideway movement approach axis dish angle step operating mode, from curve, now the steering wheel angle of driver's operation from the beginning 2 degree are increased to 60 degree, there is overshoot by a small margin centre, and four-wheel wheel speed is near 70m/s, and back-wheel drive torque is 50-200N*m.Now input real vehicle metrical information, characteristic signal vehicle speed is obtained through integrated state estimator, yaw state and acceleration estimation curve are as shown in figure 14, corresponding evaluated error curve as shown in figure 15, lateral acceleration maxim reaches 6.4m/s^2, yaw velocity and lateral acceleration become large in hand-wheel signal angle step overshoot place evaluated error, there is One-step error, the maximum estimated error of longitudinal speed of a motor vehicle is 0.4m/s, the maximum estimated error of yaw velocity is 6.9deg/s, the longitudinal acceleration evaluated error of reflection longitudinal tire force information is 0.4m/s^2 to the maximum, the evaluated error of the lateral acceleration of reflection side direction tire force information and side direction vehicle speed signal is 0.75m/s^2 to the maximum, and estimation curve tracking performance is good, meet performance requriements, show that this integrated estimator has satisfied estimation effect for the vehicle transient state response measured under time domain.

Figure 16 gives the steering wheel angle of vehicle running state modularization estimator under the operating mode of center, wheel speed and wheel driving torque and inputs information.It is 40s to 79.66s that test results intercepts the time, now vehicle speed of operation is stabilized conditions 50km/h, side direction motoring condition enters center operating mode, from curve, now the left and right directions dish corner of driver's operation is 40 degree, four-wheel wheel speed is near 44m/s, back-wheel drive torque is near 100N*m, now input real vehicle metrical information, obtain characteristic signal vehicle speed, yaw state and acceleration estimation curve as shown in figure 17 through integrated state estimator, corresponding evaluated error curve as shown in figure 18.Lateral acceleration maxim is 2.6m/s^2, vehicle is in stable sideway movement state, evaluated error curve is also more stable, the maximum estimated error of the longitudinal direction of car speed of a motor vehicle is 0.2m/s, the maximum estimated error of yaw velocity is 2deg/s, the longitudinal acceleration evaluated error of reflection longitudinal tire force information is 0.24m/s^2 to the maximum, the lateral acceleration evaluated error of reflection side direction tire force information and side direction vehicle speed signal is 0.2m/s^2 to the maximum, estimation curve and test figures are consistent, show that the maneuvering performance state estimation that this integrated estimator travels for center is good.

The mean and variance correspondingly calculating the estimated bias arranging the speed of a motor vehicle, yaw velocity and the acceleration signal that obtain above-mentioned four kinds of operating modes is as shown in table 3, and data show that the estimation effect that real train test is verified is good, have certain reliability.

Table 3

The present invention has alerting ability with portable for the method for designing of the vehicle running state modularization estimator of vehicle traveling tire power and the speed of a motor vehicle, and sensor resource and vehicle production cost can be saved, autonomous land vehicle research real-time stabilization is estimated that the method for estimation of vehicle running state has most important theories and instructs and practical application meaning.

Claims (4)

1. an autonomous land vehicle running condition information method of estimation, is characterized in that, comprises the following steps:

Step one, based on simplification auto model, design vehicle motoring condition modularization estimator, it, for estimating vehicle other running condition information according to onboard sensor measurement signal, comprises following process:

1) longitudinal tire force observer is designed, for estimating longitudinal direction of car tire force;

2) side direction tire force estimator is designed, for calculating vehicle side to tire force;

3) the non-linear full dimension state estimator of design speed, for estimating longitudinal direction of car and the side direction speed of a motor vehicle;

4) each estimator designed by above-mentioned steps is carried out integrated, after integrated, obtain vehicle running state modularization estimator;

Step 2, the vehicle running state modularization estimator input of the signal of vehicle self-carrying sensor measurement obtained by step one, estimate the estimated information of vehicle tyre power and the speed of a motor vehicle by vehicle running state modularization estimator.

2., according to a kind of autonomous land vehicle running condition information method of estimation according to claim 1, it is characterized in that, the process 1 of described step one) process of design longitudinal tire force observer is:

In conjunction with wheel rolling kinetics equation, build Unknown Input Observer to each wheel of vehicle and estimate longitudinal tire force, the interference estimator of longitudinal direction of car tire force is shown with following formula table:

$\left\{\begin{array}{c}{\hat{F}}_{\mathrm{xi}}=-\frac{J}{R_e}({\mathrm{\&chi;}}_i+{\mathrm{\&rho;}}_i{\mathrm{\&omega;}}_i)\\ {\stackrel{\&CenterDot;}{\mathrm{\&chi;}}}_1=-{\mathrm{\&rho;}}_i(\frac{T_i}{J}+{\mathrm{\&rho;}}_i{\mathrm{\&omega;}}_i)-{\mathrm{\&rho;}}_i{\mathrm{\&chi;}}_i\end{array}\right.$

Wherein, be the longitudinal tire force estimated valve of i-th wheel, unit N; J is vehicle wheel rotation inertia, units/kg * m ²; R _efor tire effective radius, unit m; ω _ibe i-th wheel wheel speed, unit m/s; T _ibe that i-th wheel drives the summation with lock torque, unit N*m; k, ρ _ifor feedback gain;

In this exogenous disturbances state estimator, as the state variable not track channel state of distracter, namely Interference Estimation amount is dynamically dynamic independent of system state estimation.

3., according to a kind of autonomous land vehicle running condition information method of estimation according to claim 1, it is characterized in that, the process 2 of described step one) process of design side direction tire force estimator is:

Utilize Dugoff tire model calculation side to tire force, the Dugoff tire model indulged under cunning-lateral deviation combinational acting represented with following nonlinear function:

$\left\{\begin{array}{c}{F}_{\mathrm{xi}}=f_{\mathrm{xi}}(V_x,V_y,a_x,a_y,r,{\mathrm{\&omega;}}_i,\mathrm{\&delta;})\\ F_{\mathrm{yi}}=f_{\mathrm{yi}}(V_x,V_y,a_x,a_y,r,{\mathrm{\&omega;}}_i,\mathrm{\&delta;})\end{array}\right.$

Wherein, F _xi/ F _yifor longitudinal direction/side direction tire force, unit N; V _x/ V _yfor longitudinal direction/side direction speed of a motor vehicle, unit m/s; a _x/ a _yfor longitudinal direction/lateral acceleration, unit m/s ²; R is yaw velocity, unit rad/s; ω _ifor wheel speed, unit m/s; δ is steering wheel angle, unit rad.

4., according to a kind of autonomous land vehicle running condition information method of estimation according to claim 1, it is characterized in that, the process 3 of described step one) the non-linear full dimension state estimator of design speed comprises following process:

Carry out force analysis to vehicle, obtaining vehicle dynamic model is:

$\left\{\begin{array}{c}\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{xi}}={\mathrm{ma}}_x=m({\stackrel{\&CenterDot;}{V}}_x-r{V}_y)\\ \underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{yi}}={\mathrm{ma}}_y=m({\stackrel{\&CenterDot;}{V}}_y+r{V}_x)\\ \underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}=I_z\stackrel{\&CenterDot;}{r}\end{array}\right.$

In above formula,

$\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{xi}}=(F_{x_1}+F_{x_2})\cos {\mathrm{\&delta;}}_w-(F_{y_1}+F_{y_2})\sin {\mathrm{\&delta;}}_w+(F_{x_3}+F_{x_4})$

$\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{F}_{\mathrm{yi}}=(F_{x_1}+F_{x_2})\cos {\mathrm{\&delta;}}_w-(F_{y_1}+F_{y_2})\sin {\mathrm{\&delta;}}_w+(F_{y_3}+F_{y_4})$

$\begin{array}{c}\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}=l_F(F_{x_1}+F_{x_2})\sin {\mathrm{\&delta;}}_w+l_F(F_{y_1}+F_{y_2})\cos {\mathrm{\&delta;}}_w+\frac{b_F}{2}(F_{x_2}-F_{x_1})\cos {\mathrm{\&delta;}}_w\\ +\frac{b_R}{2}(F_{x_4}-F_{x_3})-\frac{b_F}{2}(F_{y_2}-F_{y_1})\sin {\mathrm{\&delta;}}_w-l_R(F_{y_3}+F_{y_4})\end{array}$

Wherein, m is complete vehicle quality, units/kg; V _x/ V _yfor longitudinal direction of car/side velocity under bodywork reference frame, unit m/s; l _f/ l _rfor vehicle centroid is to the distance of front/rear axle, unit m; b _f/ b _rfor vehicle wheel base, unit m; Front wheel steering angle δ _w=δ _w1=δ _w2=δ/I _sw, δ is steering wheel angle, unit rad; I _swfor steering gear ratio, rear-axle steering angle δ _w3=δ _w4=0, unit rad; R is Vehicular yaw cireular frequency, unit rad/s; G is gravity constant; F _xi/ F _yiwhat be respectively four tires indulges/side direction tire force, unit N; I _zfor car body is around the rotor inertia of z-axis, units/kg * m ²; M _zifor the rotating torque that each tire rotates around z-axis, unit N*m;

To sum up, non-linear full dimension speed of a motor vehicle estimator is:

$\left\{\begin{array}{c}{\stackrel{\&CenterDot;}{\hat{V}}}_x=r{\hat{V}}_y+a_x+\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{K}_i(v_{\mathrm{xi}}-{\hat{V}}_x)\\ {\stackrel{\&CenterDot;}{\hat{V}}}_y=-r{\hat{V}}_x+a_y+K_y({\mathrm{ma}}_y-{\hat{F}}_y)\\ \stackrel{\&CenterDot;}{\hat{r}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}/I_z+K_r(r-\hat{r})\end{array}\right.$

Wherein,

${\stackrel{\&CenterDot;}{\hat{V}}}_x=r{\hat{V}}_y+a_x+\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{K}_i(v_{\mathrm{xi}}-{\hat{V}}_x)$ For longitudinal direction of car speed of a motor vehicle estimator;

${\stackrel{\&CenterDot;}{\hat{V}}}_y=-r{\hat{V}}_x+a_y+K_y({\mathrm{ma}}_y-{\hat{F}}_y)$ For vehicle side is to speed of a motor vehicle estimator;

$\stackrel{\&CenterDot;}{\hat{r}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{M}_{\mathrm{zi}}/I_z+K_r(r-\hat{r})$ For Vehicular yaw angular velocity estimator;

K _i, K _y, K _rbe respectively the estimator gain of the longitudinal direction of car speed of a motor vehicle, the side direction speed of a motor vehicle and yaw velocity.