This action might not be possible to undo. Are you sure you want to continue?

**Stefan Fuchshumer, Kurt Schlacher and Thomas Rittenschober
**

Abstract—The central issue of this contribution is the discus-

sion of the differential ﬂatness of the planar holonomic bicycle

model. The components of a ﬂat output are given as the lateral

and the longitudinal velocity component of a distinguished point

located on the longitudinal axis of the vehicle. This property

is shown for the front-, rear- and all-wheel driven vehicle,

without referring to particular representatives of the functions

modelling the lateral tire forces. The clear physical meaning

of the ﬂat output is regarded as particularly useful for the

control design task. The vehicle dynamics control design is

accomplished following the ﬂatness based control theory.

I. INTRODUCTION

This contribution proposes a novel approach for the non-

linear vehicle dynamics control which is essentially based on

the observation that the dynamics of the planar holonomic

bicycle model depicted in Fig. 1 are differentially ﬂat. The

contact between the tires and the road is modelled in terms

of contact forces, which implies that the tires are enabled

to slip and slide on the road. The steering angle and the

longitudinal tire forces are regarded as control inputs.

F

lv

F

sv

(v, β, r, δ)

F

lh

F

sh

(v, β, r)

δ

v

β

ψ

r, M

d

α

h

α

v

Ξ

x

y

C

X

Y

Fig. 1. The planar (holonomic) bicycle model.

While the ﬂatness property of the non-holonomic kine-

matic vehicle, which is based on the arrangement of ideal

rolling contact of the wheels, is well-known in the literature

and numerously exploited for tracking applications of vehi-

cles (with trailers), cf. e.g. [4], the differential ﬂatness of the

S. Fuchshumer is with the Christian Doppler Laboratory for Automatic

Control of Mechatronic Systems in Steel Industries, Institute of Automatic

Control and Control Systems Technology, Johannes Kepler University of

Linz, Austria. stefan.fuchshumer@jku.at

K. Schlacher is the head of the Christian Doppler Laboratory for

Automatic Control of Mechatronic Systems in Steel Industries and the

Institute of Automatic Control and Control Systems Technology, Johannes

Kepler University of Linz, Austria. kurt.schlacher@jku.at

T. Rittenschober is with Profactor Produktionsforschungs GmbH, Steyr,

Austria. thomas.rittenschober@profactor.at

(holonomic) bicycle model of Fig. 1 has not been reported

yet.

The bicycle model as introduced in [9] emerges from

the four-wheel vehicle by gluing together the front and the

rear wheels to a single (mass-less) front and rear wheel,

respectively, located at the longitudinal axis of the car. This

planar model, known as a well-established basis for the

design of vehicle dynamics control systems, see, e.g., [1],

[3], is capable of rendering the longitudinal, lateral and yaw

dynamics of the vehicle. The pitch and roll dynamics of a

vehicle are clearly not involved in the scope of this model.

Vehicle dynamics control systems as e.g. ESP (electronic

stability program) acting on the brakes and the traction con-

trol system are implemtented in series-production vehicles.

These systems support the driver in emergency situations by

producing a (counter-) yaw torque due to individually con-

trolled braking of all four wheel in the case of exceedance of

a certain yaw rate. Accordingly, the potential of involving the

steering system to handle emergency situations is very rich.

The automotive industry aims at utilizing these possibilities

e.g. by means of supporting the driver for the difﬁcult task

of counter-steering by application of a respective artiﬁcial

torque to the steering wheel, i.e., as a haptic recommendation

for the driver. As a very recent advance, active front steering

(AFS) [5] has been implemented in series-production vehi-

cles. The basic function of AFS is to mechanically add an

additional steering angle (adjusted e.g. by an electric drive) to

the steering angle given by the driver. Besides the objective to

enable a velocity-dependent gain of the steering mechanism,

the AFS system can be used for vehicle dynamics control,

too.

This paper is organized as follows: After a brief revisit

of the bicycle dynamics in Section II, the discussion is

concerned with the system analysis in paragraph III yielding

a physically relevant representative for the ﬂat output of the

bicycle model as the main result. This representative of the

ﬂat output does not depend on the particular choice of the

vehicle’s actuation, i.e., it holds for the rear-, front- and all-

wheel driven car equivalently. Additionally, the ﬂat output

is attached with a clear physical meaning. Up to a certain

family of functions which have to be excluded, the system

analysis does not refer to particular representatives for the

functions describing the lateral tire forces.

The task of (real-time) trajectory planning amounts to

mapping the inputs supplied by the driver, i.e., the current

position of the throttle/brake pedal and the angle of the

steering wheel, to suitable trajectories for the ﬂat output. By

virtue of the free parametrization of the nonlinear vehicle

dynamics by means of the ﬂat output, these trajectories can

Proceedings of the

44th IEEE Conference on Decision and Control, and

the European Control Conference 2005

Seville, Spain, December 12-15, 2005

ThA06.4

0-7803-9568-9/05/$20.00 ©2005 IEEE

6492

be shaped arbitrarily by the design engineer in order to

entail the desired dynamic behavior/handling of the vehicle.

The physical meaning of the ﬂat output is regarded to

facilitate this objective signiﬁcantly. Informations gathered

by scanning the environment, e.g., regarding the conditions

of the road, or the position of a detected obstacle, can also be

incorporated for the real-time shaping of the trajectories. The

trajectory shaping task, which inherently involves subjective

sensations of the driver, is not addressed in this contribution.

Finally, to illustrate the proposed control approach, simula-

tion results given in Section IV conclude this contribution.

II. THE BICYCLE MODEL

To start with, let us ﬁrst revisit the modelling assumptions

and the dynamics of the bicycle model depicted in Fig. 1. A

detailed discussion on these issues can be found e.g. in [9]

or [3]. The global position (X, Y ) of the center of gravity

C and the orientation ψ of the longitudinal axis represent

the degrees of freedom of the bicycle. The parameters are

given as follows: m denotes the vehicle mass, J is the

moment of inertia with respect to the yaw axis ﬁxed at C,

and l

v

, l

h

denote the distances between C and the front

and the rear wheel, respectively. The steering angle δ as

well as the longitudinal tire forces F

lv

and F

lh

, which are

due to the motor (or the braking) torque, are regarded as

control inputs. The torque M

d

denotes a disturbance acting

w.r.t. the yaw axis. The inputs are collected to the vector

u

T

= [δ, F

lv

, F

lh

, M

d

].

Let v represent the (magnitude of the) velocity of C, and

let β denote the angle between the longitudinal axis and the

velocity vector

˙

X∂

X

+

˙

Y ∂

Y

at C, i.e.,

v =

˙

X

2

+

˙

Y

2

, β = arctan

˙

Y /

˙

X

−ψ. (1)

Thus, we have

˙

X = v cos (β + ψ) ,

˙

Y = v sin (β + ψ) ,

˙

ψ = r, (2)

with r denoting the yaw rate. Notice that in the scope of

vehicle dynamics control (to assist the driver in emergency

situations), only the case v > 0 is considered.

In order to describe the tire/road contact via suitable

mathematical models for the lateral tire forces F

sv

and F

sh

,

the literature offers a wide variety of modelling assumptions.

A very common and well-established assumption, cf. e.g.

[2], is that the lateral tire forces, which allow for changes

of the vehicle’s orientation, are regarded as functions of the

respective side-slip angles of the wheels. The side-slip angles

α

h

and α

v

represent the angles between the velocity vector at

the rear/front wheel and the associated tire plane, see Fig. 1,

α

h

(v, β, r) = −arctan

v sin β −l

h

r

v cos β

,

α

v

(v, β, r, δ) = δ −arctan

v sin β + l

v

r

v cos β

.

(3)

However, in the scope of the system analysis to be

given in Section III, we do not a-priori select a particular

representative for the functions of the lateral tire forces, i.e.,

we will regard the functions F

sh

(v, β, r), F

sv

(v, β, r, δ) as

arbitrary smooth functions

1

. The dynamics of the bicycle

model read

˙

¯ x = f (¯ x, u) , ¯ x

T

=

v β r

(4)

with the components of the vector ﬁeld f given as

f

1

=

1

m

(F

sv

(v, β, r, δ) sin (β −δ) + (5)

+ F

lv

cos (β −δ) + F

sh

(v, β, r) sin β + F

lh

cos β) ,

f

2

= −r +

1

mv

(F

sv

(v, β, r, δ) cos (β −δ) − (6)

−F

lv

sin (β −δ) + F

sh

(v, β, r) cos β −F

lh

sin β)

and

f

3

=

1

J

(l

v

(F

sv

(v, β, r, δ) cos δ + F

lv

sin δ) −

−l

h

F

sh

(v, β, r) + M

d

) . (7)

The reason for selecting the representation (4)-(7) is twofold:

First, this particular choice for the coordinates ¯ x implies that

the functions f

α

do not depend on the vehicle’s orientation

ψ (and the global position X, Y as well). Thus, these coordi-

nates are often referred to as vehicle coordinates. Secondly,

as the vehicle dynamics control under consideration is not

concerned with position control of the car, but with control

objectives depending on ¯ x only, the subsystem (2) has been

dropped from the entire bicycle dynamics.

Following the usual tensor notation, coordinates and com-

ponent functions of vector ﬁelds are indicated with super-

script indices, and the Einstein convention for summation on

repeated indices is arranged. Additionally, let ∂

α

= ∂/∂¯ x

α

,

∂

δ

= ∂/∂δ, and let L

f

c = f

α

∂

α

c denote the Lie derivative

of the function c (¯ x) along the vector ﬁeld f = f

α

∂

α

.

III. THE FLATNESS PROPERTY OF THE BICYCLE

DYNAMICS

Let M

d

= 0 in the scope of the following system analysis.

We will commence by investigating the general case of the

so-called all-wheel driven vehicle, which means that the

motor (or braking) torque can be supplied to the front and

the rear wheel with a given transmission ratio. To this end,

let us introduce the distribution of the aggregate longitudinal

force F

l

to the front and the rear wheel as

F

lh

= γF

l

, F

lv

= (1 −γ) F

l

, (8)

with γ ∈ [0, 1] referred to as the transmission ratio. This

ratio γ is regarded either as a function of the time, γ (t) :

R → [0, 1], or as a function of the vehicle’s state ¯ x. The ﬁrst

point of view amounts to regarding γ as an input given by

the driver, the second viewpoint, however, is understood as

an action of a vehicle control system.

The central result for this type of vehicle is given in the

following proposition.

1

This C

∞

assumption is arranged to avoid mathematical subtleties. Thus,

it is a sufﬁcient condition in the scope of the following discussions.

6493

Proposition 1: Let the lateral tire forces F

sh

(v, β, r),

F

sv

(v, β, r, δ) be arbitrary (smooth) functions up to the

requirements

F

sh

(v, β, r) =

(ml

v

)

2

J (l

v

+ l

h

)

v

2

sin β cos β+

+κ

h

v cos β, v sin β −

J

ml

v

r

,

(9)

and

F

sv

=

κ

v

(v, β, r) −(1 −γ) (γ sin δ + (1 −γ) δ) F

l

1 −γ (1 −cos δ)

, (10)

with κ

h

, κ

v

as arbitrary (smooth) functions. Then, the

bicycle dynamics (4)-(8), u

T

= [δ, F

l

], M

d

= 0, with

γ ∈ [0, 1] regarded either as a function of the time or the state

¯ x, are differentially ﬂat (for v = 0). Moreover, this system

is exact input/state linearizable via static state feedback. An

output y =

y

1

, y

2

**entailing the exact linearization with
**

relative degree (1, 2), i.e., a ﬂat output, is given as

y

1

= c

1

(¯ x) = v cos β (11)

and

y

2

= c

2

(¯ x) = v sin β −

J

ml

v

r. (12)

Proof: (sketch) The exact linearizability is straight-

forwardly veriﬁed with (11), (12). For the (local) coordinates

transformation z = ϕ(¯ x) =

c

1

(¯ x) , c

2

(¯ x) , L

f

c

2

(¯ x)

T

,

⎡

⎣

z

1

z

2

z

3

⎤

⎦

=

⎡

⎣

v cos β

v sin β −Jr/ (ml

v

)

L

f

c

2

(¯ x)

⎤

⎦

, (13)

with

L

f

c

2

=

l

v

+ l

h

ml

v

F

sh

(v, β, r) −vr cos β, (14)

to qualify as a diffeomorphism, the requirement

B(v, β, r) −v

2

cos β/ (l

v

+ l

h

) = 0 (15)

with

B =

Jv (∂

v

F

sh

) sin β + J (∂

β

F

sh

) cos β + ml

v

v (∂

r

F

sh

)

m

2

l

2

v

has to be fulﬁlled due to the implicit function theorem.

The requirement (15) amounts to excluding the family (9)

of functions for the lateral rear tire force F

sh

. Remark 5

provides a comment on the condition (9) to show that it does

not imply a practically relevant restriction on the choice of

F

sh

.

The static state feedback entailing the linear and time-

invariant dynamics outlined in the coordinates z, with the

new input w

T

=

w

1

, w

2

**, is obtained (locally) as the
**

solution of

L

f

c

1

(¯ x, u) = w

1

,

L

2

f

c

2

(¯ x, u) = w

2

(16)

with respect to u

T

= [δ, F

l

]. Following the implicit function

theorem, (local) solvability is guaranteed iff the condition

γF

sv

sin δ −(1 −γ (1 −cos δ)) (∂

δ

F

sv

+ (1 −γ) F

l

) = 0,

(17)

as well as (15) is met. Thus, additionally to (9), we have to

impose the requirement (10). Again, Remark 5 provides a

comment on this restriction.

A. Some Key Observations Associated with the Flatness

Property of the Bicycle Dynamics

Remark 1: Note that, except for the restrictions (9) and

(10), the property of exact linearizability for (11), (12)

does not depend on the particular choice of the functions

F

sv

(v, β, r, δ) and F

sh

(v, β, r). Additionally, note that the

restrictions (9), (10) are associated with the particular choice

for c

1

(¯ x), c

2

(¯ x) as given in (11), (12).

Remark 2: The function v sin β −J/ (ml

v

) r occurring in

(12) depicts the y-component v

Ξ

y

of the velocity of the point

Ξ located on the vehicle axis with the (vehicle) coordinates

(x, y) = (−J/ (ml

v

) , 0), see Fig. 1. Thus, the output c

2

of

(12) is attached with a clear physical meaning.

Remark 3: The function c

1

(¯ x) of (11) depicts the x-

component of the velocity of points located on the vehicle’s

longitudinal axis.

Remark 4: The location of the point Ξ as introduced

above only depends on the parameters J, m and l

v

which

are known (rather) accurately in applications.

Remark 5: Typically, as sketched in Section II, the lateral

tire forces F

sh

and F

sv

are considered as functions of the

respective side-slip angles α

h

, α

v

, cf. e.g. [2] for a very

well-established approach. In view of this, the conditions

(9) and (10) do not impose practically relevant restrictions.

Particularly, to this end, note that the (arbitrary, smooth)

function κ

v

of the condition (10) on F

sv

must not depend

on the steering angle δ. Additionally, note that the function

κ

h

involved in the restriction (9) for F

sh

is a function of the

velocity components v

Ξ

x

, v

Ξ

y

of the point Ξ.

Remark 6: Note that the transmission ratio γ, regarded as

γ (t) or γ (v, β, r), does not explicitly appear in ˙ y

2

= z

3

=

L

f

c

2

, see (13) and (14).

The point Ξ is interesting also from another point of view.

To this end, let us calculate the (x, y)-decomposition of the

acceleration of a point Q located at the x-axis at the distance

x = ζ from the center of gravity C. Let R = SO(2), i.e.,

the group of rotations in the plane,

R(ψ) =

¸

cos ψ sin ψ

−sin ψ cos ψ

, 0 ≤ ψ < 2π

denote the mapping from the inertial frame (X, Y ) to the

vehicle coordinates (x, y). Variables augmented with the

subscripts x, y are related to the vehicle coordinate system,

whereas the subscripts X, Y indicate the respective inertial

frame representation. The velocity of Q is given as

¸

v

Q

x

v

Q

y

=

¸

v cos β

v sin β + ζr

,

¸

v

Q

X

v

Q

Y

= R

−1

¸

v

Q

x

v

Q

y

,

and the acceleration reads

¸

a

Q

x

a

Q

y

= R

d

dt

¸

v

Q

X

v

Q

Y

=

¸

˙ v

Q

x

˙ v

Q

y

+

¸

−v

Q

y

v

Q

x

˙

ψ.

6494

Thus, we have

a

Q

x

=

1

m

(F

lv

cos δ + F

lh

−F

sv

(v, β, r, δ) sin δ) −ζr

2

for the longitudinal component and

a

Q

y

=

1

m

1 −ζ

ml

h

J

F

sh

(v, β, r) +

+

1 + ζ

ml

v

J

(F

lv

sin δ + F

sv

(·) cos δ)

+

ζ

J

M

d

for the lateral acceleration, with (·) = (v, β, r, δ).

Remark 7: The lateral acceleration a

y

at the point Ξ does

not explicitly depend on the contact forces of the front wheel,

i.e, F

sv

(v, β, r, δ) and F

lv

.

Remark 8: For ζ = J/ (ml

h

), the lateral acceleration

does not explicitly depend on the lateral force F

sh

(v, β, r)

of the rear wheel. This point, with coordinates (x, y) =

(J/ (ml

h

) , 0), occurs in the analysis of Ackermann leading

to the “robustly decoupling control law of DLR” [1], [3].

Remark 9: There is an interesting relation to prior work

on ﬂatness, in particular, on conﬁguration ﬂatness of La-

grangian systems underactuated by one control [8]. E.g.,

the planar rigid body actuated by two body-ﬁxed forces

allows for a conﬁguration-ﬂat output (Huyghens oscillation

center), see, e.g., [6], with the PVTOL being a well-known

representative of this class. However, in contrast to the ﬂat-

ness based PVTOL (position) control, the proposed vehicle

dynamics control involves static state feedback only, which

is a consequence of the control objective involving functions

of ¯ x instead of conﬁguration coordinates.

B. The Front- and the Rear-Wheel Driven Bicycle

The ﬂatness property of the rear-wheel driven vehicle (γ =

1) and the front-wheel driven vehicle (γ = 0) follow as

special cases from Proposition 1. For the rear-wheel driven

car, the restriction (17) yields F

sv

tan δ − ∂

δ

F

sv

= 0, and,

thus, the requirement

F

sv

(v, β, r, δ) = κ

v

(v, β, r) cos

−1

δ,

see (10). In the case of the front-wheel driven bicycle, the

restriction

F

sv

(v, β, r, δ) = κ

v

(v, β, r) −F

l

δ

has to be obeyed, see (10).

IV. SCHEME OF A FLATNESS BASED VEHICLE

DYNAMICS CONTROL

Given the trajectory of the ﬂat output y, the associated

trajectories of the system variables ¯ x

T

= [v, β, r] and the

control inputs u

T

= [δ, F

l

] are obtained as the solution of

F =

¸

c

1

(¯ x) = y

1

, c

2

(¯ x) = y

2

, L

f

c

2

(¯ x) = ˙ y

2

,

L

f

c

1

(¯ x, u) = ˙ y

1

, L

2

f

c

2

(¯ x, u) = ¨ y

2

¸

.

(18)

By obeying the restrictions regarding the lateral tire forces

F

sv

, F

sh

given in Proposition 1 and Section III-B, respec-

tively, the implicit function theorem guarantees the local

solvabiltiy of (18).

Following [4], [10], [11], Fig. 2 depicts the scheme of

a ﬂatness based control, which is applied to the bicycle

model. The nonlinear state feedback u

i

= χ

i

(¯ x, w), i = 1, 2,

derived from (16) entails the exact linearization of the bicycle

dynamics, ˙ z

1

= w

1

, ˙ z

2

= z

3

, ˙ z

3

= w

2

, see also (13).

This implies the linearity and time-invariance of the tracking

error dynamics, with the tracking error e = y − y

d

. Here,

the subscript d is used to indicate the desired values. The

asymptotic stabilization of the tracking error dynamics

˙ e

1

= ˙ y

1

− ˙ y

1

d

= ˙ z

1

− ˙ y

1

d

= w

1

− ˙ y

1

d

¨ e

2

= ¨ y

2

− ¨ y

2

d

= ˙ z

3

− ¨ y

2

d

= w

2

− ¨ y

2

d

(19)

can be obtained by means of the control law

w

1

= ˙ y

1

d

−µe

1

− ¯ µξ

1

= ˙ y

1

d

−µ

z

1

−y

1

d

− ¯ µξ

1

w

2

= ¨ y

2

d

−ν

1

e

2

−ν

2

˙ e

2

− ¯ νξ

2

=

= ¨ y

2

d

−ν

1

z

2

−y

2

d

−ν

2

z

3

− ˙ y

2

d

− ¯ νξ

2

(20)

with the integral part

˙

ξ

1

= e

1

,

˙

ξ

2

= e

2

(21)

and µ > 0, ¯ µ > 0, and ν

1

, ν

2

, ¯ ν such that the characteristic

polynomials are Hurwitz. Thus, the error dynamics read ˙ e

1

=

−µe

1

− ¯ µξ

1

, ¨ e

2

= −ν

1

e

2

−ν

2

˙ e

2

− ¯ νξ

2

with (21).

In order to illustrate the proposed vehicle dynamics control

approach, a close-to-reality modelled sports car which is pro-

vided by the multi-body simulation program MSC.ADAMS

[7], is used as a testrig. This demo vehicle comprises fully-

detailed models of the suspension, the powertrain and the

steering system, while the bodywork is considered as a rigid

body, spring-mounted on the chassis. The total number of

degrees of freedom amounts to 96.

The parameters of the corresponding bicycle model, which

have been extracted from the MSC.ADAMS vehicle, are

given by m = 1529 kg, J = 1344 kgm

2

, l

h

= 1.08 m and

l

v

= 1.481 m. The point Ξ, which is attached with particular

meaning regarding the ﬂatness property, see Remark 2, is

located at (x, y) = (−0.594 m, 0).

The lateral forces F

sh

, F

sv

acting on the tires are modelled

following the widely-used model of Pacejka [2],

F

sk

(α

k

) = 2D

s

sin (C

s

arctan (B

s

α

k

−

− E

s

(B

s

α

k

−arctan (B

s

α

k

)))) ,

k ∈ {v, h}, often referred to as “Pacejka’s magic formula” in

the literature. Thus, the lateral forces are given as functions

z = ϕ(¯ x)

tracking

controller

bicycle

dynamics

nonlinear

state feedback

(13)

(4)-(8) (16) (20),(21)

y u w

¯ x

y

d

, ˙ y

d

, .

z

Fig. 2. Scheme of a ﬂatness based vehicle dynamics control.

6495

0

0.5

1

1.5

2

2.5

3

3.5

0 20 40 60 80 100 120 140 160

Y

(

m

)

X (m)

Fig. 3. Single lane change maneuver due to (22), (23): Trajectory of the

vehicle’s center of gravity C.

of the respective side-slip angles, see also Remark 5. For

instance, the parameters of the rear tire model are chosen to

be B

s

= 13 rad

−1

, C

s

= 1.65, D

s

= 4789 N, E

s

= 0.68.

Clearly, for an application, the shaping of the trajectories

is to be done in real-time, based on the inputs supplied by

the driver. To this end, as the position of the throttle/brake

pedal is thought of as to reﬂect the driver’s demands on the

longitudinal dynamics, this input (supplied at the current time

t) might be used to deﬁne the desired trajectory for v

Ξ

x

for

a future time horizon. Accordingly, the angle of the steering

wheel may be regarded as the driver’s demand on the lateral

dynamics, so, this input might be used to adjust a desired

trajectory for v

Ξ

y

.

However, for the current sake of illustration, the desired

trajectory y

d

(t) : [0, T] → R

2

of the ﬂat output is chosen as

y

1

d

(t) = v

0

+

3t

2

T −2t

3

T

3

(v

T

−v

0

) (22)

for the longitudinal component of the velocity at Ξ, and

y

2

d

(t) =

⎧

⎨

⎩

p (t −t

1

, a

1

, t

2

−t

1

) , t ∈ [t

1

, t

2

)

p (t −t

2

, a

2

, t

3

−t

2

) , t ∈ [t

2

, t

3

)

0, else

(23)

for the lateral component, with

p (t, a, τ) = −a

t

3

(τ −t)

3

τ

6

.

Here, v

0

and v

T

denote the x-component of the velocity at

the time t = 0 and t = T, respectively. Let t

1

= 1.5 s, t

2

=

2.5 s, t

3

= 3.5 s, T = 5 s, v

0

= 27.7 m/s, v

T

= 33.3 m/s,

a

1

= 50 m/s, a

2

= −57 m/s, then this trajectory y

d

(t) of

(22), (23) implies a single lane change maneuver, associated

with an acceleration of the vehicle, see Fig. 3. Fig. 4 ﬁnally

depicts the simulation results of the proposed ﬂatness based

control. The coefﬁcients of the tracking controller (20) are

chosen as µ = 10 s

−1

, ¯ µ = 10 s

−2

, and ν

1

= 1200 s

−2

,

ν

2

= 60 s

−1

, ¯ ν = 8000 s

−3

.

Remark 10: The difference between the actual rear tire

longitudinal force F

lh

and the predicted value (dashed line)

is due to the fact that the MSC.ADAMS model involves

aerodynamic forces and the road resistance, which have to

be overcome to follow the desired trajectory.

Remark 11: The deviation of the yaw rate and the steering

input results from unequal loading of the inner and the outer

track during cornering. This unequal loading clearly affects

the lateral and longitudinal tire forces via the normal force.

V. CONCLUSIONS AND FUTURE WORKS

Individually controlled braking of all four wheels (ESP)

to maintain the vehicle’s stability and steering response has

proven as very valuable to increase safety. Comparably,

making available the steering system to the vehicle dynamics

control offers additional potential to support the driver in

emergency situations. To this end, the differential ﬂatness

property of the bicycle model might offer new perspectives

to cope with the vehicle dynamics control problem involving

the longitudinal forces of the tires and the steering angle as

control inputs. The ﬂat output revealed in this contribution

could be identiﬁed as the longitudinal and the lateral compo-

nent of the velocity of a certain point located on the vehicle’s

longitudinal axis. The location of this distinguished point

is determined in terms of the mass, the moment of inertia

and the distance between the front wheel and the center of

gravity, which can be regarded as well-known parameters in

practical applications. Additionally, this ﬂat output does not

depend on the particular actuation of the vehicle, i.e., it holds

for the rear-, front- and all-wheel driven car equivalently.

Besides the objective of real-time trajectory shaping for

the ﬂat output, our future research will be concerned with

observer design and parameter identiﬁcation of the tire char-

acteristics. While the ﬁrst issue deals with driver’s experience

during demanding maneuvers, the second point focuses on

the fact that the automotive industry, in general, is strongly

reluctant to implement side-slip angle sensors. Finally, vehi-

cle dynamics control is required to work seamlessly under

various road conditions.

The authors wish to thank the reviewers for the useful

comments.

REFERENCES

[1] J. Ackermann, T. B¨ unte and D. Odenthal, ”Advantages of active steer-

ing for vehicle dynamics control”, in 32nd Int. Symp. on Automotive

Technology and Automation, Vienna, Austria, pp. 263-270, 1999.

[2] E. Bakker, H. Pacejka and L. Lidner, A new tire model with an

application in vehicle dynamics studies, SAE Paper No. 890087, pp.

101-113, 1989.

[3] T. B¨ unte, Beitr¨ age zur robusten Lenkregelung von Personenkraftwa-

gen, PhD thesis, Technische Hochschule Aachen, 1998.

[4] M. Fliess, J. L´ evine, P. Martin and P. Rouchon, Flatness and defect of

non-linear systems: Introductory theory and examples, Int. J. Control,

61:1327-1361, 1995.

[5] P. K¨ ohn, G. Baumgarten, ”Die Aktivlenkung – das neue fahrdynamis-

che Lenksystem von BMW”, in 11. Aachener Kolloquium Fahrzeug-

und Motorentechnik, 2002.

[6] P. Martin, R.M. Murray, P. Rouchon, Flat systems, equivalence and

trajectory generation, CDS Technical Report, CDS 2003-008, 2003.

[7] MSC.ADAMS, www.mscsoftware.com.

6496

27.0

28.0

29.0

30.0

31.0

32.0

33.0

34.0

y

1

(

m

/

s

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

β

(

◦

)

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

y

2

(

m

/

s

)

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

r

(

r

a

d

/

s

)

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

α

h

(

◦

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

δ

(

◦

)

-8.0

-6.0

-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

0 1 2 3 4 5

a

y

(

m

/

s

2

)

t (s)

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0 1 2 3 4 5

F

l

h

(

k

N

)

t (s)

Fig. 4. Simulation results of the proposed ﬂatness based vehicle dynamics control approach, using a close-to-reality modelled sports car provided by the

multi-body simulation program MSC.ADAMS. The desired trajectories are indicated with dashed lines.

[8] M. Rathinam, R.M. Murray, Conﬁguration ﬂatness of Lagrangian sys-

tems underactuated by one control, SIAM J. Control Optim., 36(1):164-

179, 1998.

[9] P. Riekert, T. Schunck, Zur Fahrmechanik des gummibereiften Kraft-

fahrzeugs, Ingenieur Archiv, 11:210-224, 1940.

[10] R. Rothfuß, Anwendung der ﬂachheitsbasierten Analyse und Regelung

nichtlinearer Mehrgr¨ oßensysteme, VDI Verlag, 1997.

[11] J. Rudolph, Flatness based control of distributed parameter systems,

Shaker, Aachen, 2003.

[12] S. Sastry, Nonlinear systems, analysis, stability and control, Springer,

New York, 1999.

6497

which are due to the motor (or the braking) torque. u) . The physical meaning of the ﬂat output is regarded to facilitate this objective signiﬁcantly. and let β denote the angle between the longitudinal axis and the ˙ ˙ velocity vector X∂X + Y ∂Y at C. The global position (X. Fsv (v. The steering angle δ as well as the longitudinal tire forces Flv and Flh . the subsystem (2) has been ¯ dropped from the entire bicycle dynamics. is not addressed in this contribution. it is a sufﬁcient condition in the scope of the following discussions. 1]. ˙ Y = v sin (β + ψ) . We will commence by investigating the general case of the so-called all-wheel driven vehicle. lh denote the distances between C and the front and the rear wheel. in the scope of the system analysis to be given in Section III. is understood as an action of a vehicle control system. and the Einstein convention for summation on ¯ repeated indices is arranged. this particular choice for the coordinates x implies that ¯ the functions f α do not depend on the vehicle’s orientation ψ (and the global position X. The ﬁrst ¯ point of view amounts to regarding γ as an input given by the driver. In order to describe the tire/road contact via suitable mathematical models for the lateral tire forces Fsv and Fsh . THE FLATNESS PROPERTY OF THE BICYCLE DYNAMICS Let Md = 0 in the scope of the following system analysis. r. let us introduce the distribution of the aggregate longitudinal force Fl to the front and the rear wheel as Flh = γFl . ˙ ˙ β = arctan Y /X − ψ. 1 (Fsv (v. r). cf.g. the second viewpoint. 1 This C ∞ assumption is arranged to avoid mathematical subtleties. δ) sin (β − δ) + (5) m + Flv cos (β − δ) + Fsh (v. 6493 . i. β. A very common and well-established assumption. v cos β v sin β + lv r . Flh .e. β. J is the moment of inertia with respect to the yaw axis ﬁxed at C. which means that the motor (or braking) torque can be supplied to the front and the rear wheel with a given transmission ratio. or the position of a detected obstacle. The parameters are given as follows: m denotes the vehicle mass. The inputs are collected to the vector uT = [δ. these coordinates are often referred to as vehicle coordinates. r) sin β + Flh cos β) . r. Y as well). simulation results given in Section IV conclude this contribution. the literature offers a wide variety of modelling assumptions. (2) The reason for selecting the representation (4)-(7) is twofold: First. 1. see Fig.. let ∂α = ∂/∂ xα . ¯ x xT = ¯ v β r (4) with the components of the vector ﬁeld f given as f1 = 1 (Fsv (v. β. or as a function of the vehicle’s state x. r. f 2 = −r + and f3 = (7) Thus. the yaw axis.. Thus. β. A detailed discussion on these issues can be found e. 1] referred to as the transmission ratio. which inherently involves subjective sensations of the driver. x III. δ) cos (β − δ) − (6) mv −Flv sin (β − δ) + Fsh (v. r) + Md ) . v sin β − lh r . we do not a-priori select a particular representative for the functions of the lateral tire forces. The central result for this type of vehicle is given in the following proposition. e. The dynamics of the bicycle model read ˙ x = f (¯. and let Lf c = f α ∂α c denote the Lie derivative of the function c (¯) along the vector ﬁeld f = f α ∂α . is that the lateral tire forces. Notice that in the scope of vehicle dynamics control (to assist the driver in emergency situations). II. Let v represent the (magnitude of the) velocity of C. Thus. 1.t. [2]. γ (t) : R → [0. This ratio γ is regarded either as a function of the time. r. β. Md ]. e. respectively. coordinates and component functions of vector ﬁelds are indicated with superscript indices. in [9] or [3]. β. only the case v > 0 is considered. Following the usual tensor notation. are regarded as control inputs. i. ˙ ψ = r..be shaped arbitrarily by the design engineer in order to entail the desired dynamic behavior/handling of the vehicle.e. as the vehicle dynamics control under consideration is not concerned with position control of the car. The side-slip angles αh and αv represent the angles between the velocity vector at the rear/front wheel and the associated tire plane. ∂δ = ∂/∂δ.g. we have ˙ X = v cos (β + ψ) . Additionally. Flv .g. but with control objectives depending on x only. to illustrate the proposed control approach. β. Y ) of the center of gravity C and the orientation ψ of the longitudinal axis represent the degrees of freedom of the bicycle. let us ﬁrst revisit the modelling assumptions and the dynamics of the bicycle model depicted in Fig. however. can also be incorporated for the real-time shaping of the trajectories. Secondly. δ) = δ − arctan v cos β αh (v. β. with γ ∈ [0. THE BICYCLE MODEL To start with.r. Informations gathered by scanning the environment. v= ˙ ˙ X 2 + Y 2. β. Flv = (1 − γ) Fl . αv (v. δ) as arbitrary smooth functions1 . The trajectory shaping task. r) = − arctan (3) However. regarding the conditions of the road. and lv . δ) cos δ + Flv sin δ) − J −lh Fsh (v. r) cos β − Flh sin β) 1 (lv (Fsv (v. r. (8) with r denoting the yaw rate. which allow for changes of the vehicle’s orientation. β. are regarded as functions of the respective side-slip angles of the wheels. (1) we will regard the functions Fsh (v. Finally. To this end. The torque Md denotes a disturbance acting w.

except for the restrictions (9) and (10). Then. the property of exact linearizability for (11). T (14) (15) x L2 c2 (¯. The requirement (15) amounts to excluding the family (9) of functions for the lateral rear tire force Fsh . mlv (11) (12) Proof: (sketch) The exact linearizability is straightforwardly veriﬁed with (11). additionally to (9). regarded as γ (t) or γ (v. A. as sketched in Section II. v sin β − r .. Moreover. 6494 . r). β. Q vX Q vY = R−1 Q −vy Q vx Q vx Q vy . y). β. The point Ξ is interesting also from another point of view. r) − v 2 cos β/ (lv + lh ) = 0 with Jv (∂v Fsh ) sin β + J (∂β Fsh ) cos β + mlv v (∂r Fsh ) 2 m2 l v has to be fulﬁlled due to the implicit function theorem. whereas the subscripts X. this system ¯ is exact input/state linearizable via static state feedback. mlv to qualify as a diffeomorphism. uT = [δ. R (ψ) = cos ψ − sin ψ sin ψ cos ψ . y are related to the vehicle coordinate system. Remark 5 provides a comment on the condition (9) to show that it does not imply a practically relevant restriction on the choice of Fsh . r) = (mlv ) v 2 sin β cos β+ J (lv + lh ) J + κh v cos β. the requirement Lf c2 = B (v. 1. the bicycle dynamics (4)-(8). and the acceleration reads Q d vX aQ x =R Q Q ay dt vY = vx ˙Q vy ˙Q + ˙ ψ. For the (local) coordinates T transformation z = ϕ (¯) = c1 (¯) . r) − vr cos β. To this end. The velocity of Q is given as Q vx Q vy (16) with respect to u = [δ. r. Fl ]. Particularly. is given as y 1 = c1 (¯) = v cos β x and y 2 = c2 (¯) = v sin β − x J r. note that the function κh involved in the restriction (9) for Fsh is a function of the Ξ Ξ velocity components vx . the group of rotations in the plane. with the new input w T = w1 . let us calculate the (x. see (13) and (14). i. Y ) to the vehicle coordinates (x. Thus. the conditions (9) and (10) do not impose practically relevant restrictions. Remark 5: Typically. δ) and Fsh (v. Let R = SO (2). Variables augmented with the subscripts x. c2 (¯) . Fsv (v. x x x x ⎤ ⎡ 1 ⎤ ⎡ z v cos β ⎣ z 2 ⎦ = ⎣ v sin β − Jr/ (mlv ) ⎦ . Lf c2 (¯) . e. is obtained (locally) as the solution of B= x Lf c1 (¯. (12) does not depend on the particular choice of the functions Fsv (v. β. β. r). u) = w 1 . w2 . y 2 entailing the exact linearization with relative degree (1. c2 (¯) as given in (11). Some Key Observations Associated with the Flatness Property of the Bicycle Dynamics Remark 1: Note that. β. (12). y)-decomposition of the acceleration of a point Q located at the x-axis at the distance x = ζ from the center of gravity C. see Fig.e. (12).g. Thus. note that the restrictions (9). smooth) function κv of the condition (10) on Fsv must not depend on the steering angle δ. 1] regarded either as a function of the time or the state x. β. m and lv which are known (rather) accurately in applications. Additionally. αv . In view of this. (local) solvability is guaranteed iff the condition γFsv sin δ − (1 − γ (1 − cos δ)) (∂δ Fsv + (1 − γ) Fl ) = 0. the output c2 of (12) is attached with a clear physical meaning.. r. Remark 6: Note that the transmission ratio γ. 0 ≤ ψ < 2π (9) and Fsv = κv (v. κv as arbitrary (smooth) functions. Remark 3: The function c1 (¯) of (11) depicts the xx component of the velocity of points located on the vehicle’s longitudinal axis. y) = (−J/ (mlv ) . vy of the point Ξ. cf. β. Remark 2: The function v sin β − J/ (mlv ) r occurring in Ξ (12) depicts the y-component vy of the velocity of the point Ξ located on the vehicle axis with the (vehicle) coordinates (x.e. Md = 0. we have to impose the requirement (10). (17) = v cos β v sin β + ζr . The static state feedback entailing the linear and timeinvariant dynamics outlined in the coordinates z. β. Remark 5 provides a comment on this restriction. to this end. 0). (13) z3 x Lf c2 (¯) with lv + l h Fsh (v. Following the implicit function theorem. the lateral tire forces Fsh and Fsv are considered as functions of the respective side-slip angles αh . δ) be arbitrary (smooth) functions up to the requirements Fsh (v. mlv 2 as well as (15) is met. Remark 4: The location of the point Ξ as introduced above only depends on the parameters J. An output y = y 1 . Additionally. does not explicitly appear in y 2 = z 3 = ˙ Lf c2 . a ﬂat output. with γ ∈ [0. Again. u) = w 2 f denote the mapping from the inertial frame (X. are differentially ﬂat (for v = 0). r). (10) are associated with the particular choice x x for c1 (¯). (10) 1 − γ (1 − cos δ) with κh . note that the (arbitrary. Y indicate the respective inertial frame representation. r) − (1 − γ) (γ sin δ + (1 − γ) δ) Fl .Proposition 1: Let the lateral tire forces Fsh (v. [2] for a very well-established approach. Fl ]. 2). i. β.

Fsv (v. r. the associated trajectories of the system variables xT = [v. In order to illustrate the proposed vehicle dynamics control approach. which is a consequence of the control objective involving functions of x instead of conﬁguration coordinates. k ∈ {v. the restriction (17) yields Fsv tan δ − ∂δ Fsv = 0. Fl ] are obtained as the solution of F = c1 (¯) = y 1 . ˙ ξ 2 = e2 (21) ¯ and µ > 0. see. u) = y 2 . y) = (J/ (mlh ) . respectively.(21) (16) x ¯ z = ϕ (¯) x (13) By obeying the restrictions regarding the lateral tire forces Fsv . The lateral forces Fsh . The asymptotic stabilization of the tracking error dynamics e2 = y 2 − yd = z 3 − yd = w2 − yd ¨ ¨ ¨2 ˙ ¨2 ¨2 can be obtained by means of the control law w2 = yd − ν1 e2 − ν2 e2 − ν ξ 2 = ¨2 ˙ ¯ = yd ¨2 − ν1 z − 2 2 yd 1 w1 = yd − µe1 − µξ 1 = yd − µ z 1 − yd − µξ 1 ¯ ˙1 ¯ ˙1 for the lateral acceleration. on conﬁguration ﬂatness of Lagrangian systems underactuated by one control [8]. w). the lateral acceleration does not explicitly depend on the lateral force Fsh (v. ν such that the characteristic ¯ polynomials are Hurwitz. . Thus. the lateral forces are given as functions yd . the requirement Fsv (v. E. z 2 = z 3 . with (·) = (v. β. β. 2. The nonlinear state feedback ui = χi (¯. u) = y 1 . β. δ) = κv (v. while the bodywork is considered as a rigid body. ν2 .ADAMS [7]. in particular. The total number of degrees of freedom amounts to 96. often referred to as “Pacejka’s magic formula” in the literature. β. occurs in the analysis of Ackermann leading to the “robustly decoupling control law of DLR” [1]. [10].g. the subscript d is used to indicate the desired values. Remark 8: For ζ = J/ (mlh ). Fig. see also (13). r. c2 (¯) = y 2 . the proposed vehicle dynamics control involves static state feedback only. derived from (16) entails the exact linearization of the bicycle ˙ ˙ dynamics. β. x x x ˙ x ˙ x ¨ Lf c1 (¯. e2 = −ν1 e2 − ν2 e2 − ν ξ 2 with (21). The point Ξ.e. h}. r) cos−1 δ. r) of the rear wheel. δ) = κv (v. y d . Lf c2 (¯) = y 2 . 2 depicts the scheme of a ﬂatness based control. y) = (−0... β. Remark 9: There is an interesting relation to prior work on ﬂatness. ˙ This implies the linearity and time-invariance of the tracking error dynamics. Thus. is located at (x. Remark 7: The lateral acceleration ay at the point Ξ does not explicitly depend on the contact forces of the front wheel. In the case of the front-wheel driven bicycle. and ν1 . The parameters of the corresponding bicycle model. δ). z 1 = w1 .ADAMS vehicle. J = 1344 kgm2 . z 3 = w2 . the error dynamics read e1 = ˙ ¯ ¨ ˙ ¯ −µe1 − µξ 1 . we have 1 aQ = (Flv cos δ + Flh − Fsv (v. f (18) e1 = y 1 − yd = z 1 − yd = w1 − yd ˙ ˙ ˙1 ˙ ˙1 ˙1 (19) (20) yd ˙2 − νξ ¯ 2 − ν2 z − 3 with the integral part ˙ ξ 1 = e1 . 0). r. IV. Fsh given in Proposition 1 and Section III-B. Fsk (αk ) = 2Ds sin (Cs arctan (Bs αk − − Es (Bs αk − arctan (Bs αk )))) . L2 c2 (¯. and. The Front. with coordinates (x.and the Rear-Wheel Driven Bicycle The ﬂatness property of the rear-wheel driven vehicle (γ = 1) and the front-wheel driven vehicle (γ = 0) follow as special cases from Proposition 1. µ > 0. [6]. ˙ z tracking w nonlinear u bicycle y state feedback dynamics controller (4)-(8) (20). β. i.594 m. [11]. r] and the ¯ control inputs uT = [δ.g. This point. δ) sin δ) − ζr 2 x m for the longitudinal component and aQ = y + 1 m 1−ζ mlv J mlh J Fsh (v. the restriction Fsv (v. 2. However. spring-mounted on the chassis. which is applied to the bicycle x model. Scheme of a ﬂatness based vehicle dynamics control. the planar rigid body actuated by two body-ﬁxed forces allows for a conﬁguration-ﬂat output (Huyghens oscillation center). with the PVTOL being a well-known representative of this class. are given by m = 1529 kg. thus. 6495 . SCHEME OF A FLATNESS BASED VEHICLE DYNAMICS CONTROL Given the trajectory of the ﬂat output y. Fig. r. β.Thus. with the tracking error e = y − yd . β. r) − Fl δ has to be obeyed. r) + ζ Md J 1+ζ (Flv sin δ + Fsv (·) cos δ) + Following [4]. see (10). see Remark 2. [3]. a close-to-reality modelled sports car which is provided by the multi-body simulation program MSC.481 m. 0). i = 1. δ) and Flv . which have been extracted from the MSC. β. This demo vehicle comprises fullydetailed models of the suspension. the powertrain and the steering system. ¯ B. see (10). lh = 1. in contrast to the ﬂatness based PVTOL (position) control. is used as a testrig. e. For the rear-wheel driven car. r. Fsv acting on the tires are modelled following the widely-used model of Pacejka [2]. which is attached with particular meaning regarding the ﬂatness property.08 m and lv = 1. Here. the implicit function theorem guarantees the local solvabiltiy of (18).

J. vehicle dynamics control is required to work seamlessly under various road conditions. on Automotive Technology and Automation. equivalence and trajectory generation. 101-113. for an application. [2] E. Baumgarten.5 3 2. K¨ hn. pp. However. 1998. Pacejka and L. Besides the objective of real-time trajectory shaping for the ﬂat output.e. 3. Control. P. a2 .3. else 1 yd (t) = v0 + for the lateral component. Odenthal. 6496 . so. µ = 10 s−2 . this input (supplied at the current time Ξ t) might be used to deﬁne the desired trajectory for vx for a future time horizon. [5] P. 2003. 263-270. (23) implies a single lane change maneuver. Martin and P. v0 = 27. in general. Clearly. The ﬂat output revealed in this contribution could be identiﬁed as the longitudinal and the lateral component of the velocity of a certain point located on the vehicle’s longitudinal axis. a1 . [7] MSC.M.mscsoftware. G. which can be regarded as well-known parameters in practical applications. R EFERENCES 3 of the respective side-slip angles. Single lane change maneuver due to (22). The authors wish to thank the reviewers for the useful comments. a. V. Remark 11: The deviation of the yaw rate and the steering input results from unequal loading of the inner and the outer track during cornering. p (t. (23): Trajectory of the vehicle’s center of gravity C. Vienna. respectively. the angle of the steering wheel may be regarded as the driver’s demand on the lateral dynamics. then this trajectory yd (t) of (22). CDS 2003-008. associated with an acceleration of the vehicle. Es = 0. t2 ) 2 p (t − t2 . Fliess. the shaping of the trajectories is to be done in real-time. B¨ nte and D. To this end. ¯ ν2 = 60 s−1 . this input might be used to adjust a desired Ξ trajectory for vy . The coefﬁcients of the tracking controller (20) are ¯ chosen as µ = 10 s−1 . 1999. 1995. H. Martin. ”Die Aktivlenkung – das neue fahrdynamiso che Lenksystem von BMW”. in 11. Additionally.5 0 0 20 40 60 80 100 120 140 160 X (m) Fig. and ⎧ ⎨ p (t − t1 . Bakker. Ackermann. Int. 4 ﬁnally depicts the simulation results of the proposed ﬂatness based control. with t3 (τ − t) . t2 = 2. vT = 33. Murray. Aachener Kolloquium Fahrzeugund Motorentechnik. Austria. 890087. L´ vine. T. The location of this distinguished point is determined in terms of the mass. i. PhD thesis. T ] → R2 of the ﬂat output is chosen as 3t2 T − 2t3 (vT − v0 ) (22) T3 for the longitudinal component of the velocity at Ξ. τ ) = −a τ6 Here. front.ADAMS model involves aerodynamic forces and the road resistance. Finally.5 s. see Fig. Fig. Comparably. ν = 8000 s−3 .5 Y (m) 2 1. 2002.5 s. Rouchon.68. the moment of inertia and the distance between the front wheel and the center of gravity. the parameters of the rear tire model are chosen to be Bs = 13 rad−1 . This unequal loading clearly affects the lateral and longitudinal tire forces via the normal force. CONCLUSIONS AND FUTURE WORKS Individually controlled braking of all four wheels (ESP) to maintain the vehicle’s stability and steering response has proven as very valuable to increase safety. Beitr¨ ge zur robusten Lenkregelung von Personenkraftwau a gen. SAE Paper No. T = 5 s. CDS Technical Report. Cs = 1. R. Let t1 = 1. t2 − t1 ) . in 32nd Int. as the position of the throttle/brake pedal is thought of as to reﬂect the driver’s demands on the longitudinal dynamics. it holds for the rear-. which have to be overcome to follow the desired trajectory.7 m/s. 61:1327-1361.3 m/s.and all-wheel driven car equivalently. based on the inputs supplied by the driver. [6] P. [3] T. t3 − t2 ) . t3 ) yd (t) = (23) ⎩ 0. Rouchon. a1 = 50 m/s. 1989. is strongly reluctant to implement side-slip angle sensors. Technische Hochschule Aachen. B¨ nte. t ∈ [t1 . Accordingly. see also Remark 5. To this end. pp. Remark 10: The difference between the actual rear tire longitudinal force Flh and the predicted value (dashed line) is due to the fact that the MSC. While the ﬁrst issue deals with driver’s experience during demanding maneuvers. ”Advantages of active steeru ing for vehicle dynamics control”. our future research will be concerned with observer design and parameter identiﬁcation of the tire characteristics. [4] M. Ds = 4789 N. t ∈ [t2 . Lidner. the second point focuses on the fact that the automotive industry. the desired trajectory yd (t) : [0. t3 = 3.5 1 0. For instance. v0 and vT denote the x-component of the velocity at the time t = 0 and t = T . Symp.ADAMS. A new tire model with an application in vehicle dynamics studies.5 s.com. www. the differential ﬂatness property of the bicycle model might offer new perspectives to cope with the vehicle dynamics control problem involving the longitudinal forces of the tires and the steering angle as control inputs. Flatness and defect of e non-linear systems: Introductory theory and examples. P. making available the steering system to the vehicle dynamics control offers additional potential to support the driver in emergency situations. Flat systems. J. [1] J. this ﬂat output does not depend on the particular actuation of the vehicle.65. and ν1 = 1200 s−2 . for the current sake of illustration.. a2 = −57 m/s. 3.

34.ADAMS.1 0.M.4 -0. T.2 0.6 -0.0 1. stability and control. analysis. VDI Verlag. New York.0 -0. Nonlinear systems. R. 11:210-224. SIAM J.5 -2. Shaker.0 1.5 0.5 0.0 27.3 1. Riekert.2 0. 2003. Schunck.5 4.0 0.5 -1.0 αh (◦ ) 0. [12] S.5 1.5 3.5 -1. Anwendung der ﬂachheitsbasierten Analyse und Regelung nichtlinearer Mehrgr¨ ßensysteme. 1998.0 2. [9] P.5 1. The desired trajectories are indicated with dashed lines.8 0.2 -0. Simulation results of the proposed ﬂatness based vehicle dynamics control approach. Control Optim. Murray. Zur Fahrmechanik des gummibereiften Kraftfahrzeugs.0 8.0 -4.0 -0.0 -0.8 2. 6497 . Ingenieur Archiv.2 -0.0 1. Rathinam. Rudolph..0 y 1 (m/s) β (◦ ) r (rad/s) δ (◦ ) Flh (kN) 31.0 33.5 0.1 -0.0 -2.0 3.0 -0.0 4. 4. Springer.0 0 1 2 t (s) 3 4 5 1.0 -0.0 28.5 -1.0 0.5 1. o [11] J.0 -1. Sastry. [10] R.0 0. Conﬁguration ﬂatness of Lagrangian systems underactuated by one control. Aachen.5 1.0 -1.0 0.5 2. using a close-to-reality modelled sports car provided by the multi-body simulation program MSC.0 30. 1997.0 -0.5 0 1 2 t (s) 3 4 5 Fig.5 0.4 0. 36(1):164179.0 -8.0 29. ay (m/s2 ) y 2 (m/s) [8] M. 1940.0 -6. 1999. Flatness based control of distributed parameter systems.5 0.0 2.6 0.0 32.3 0.0 -1. Rothfuß.0 6.0 0.

Sign up to vote on this title

UsefulNot useful- Consumer Reports-How to Spot a Lemonby MRT10
- VANET- Based Vehicle Tacking on Message Authentication and Secure Navigation Routeby arun_shakar511
- Technical Note Edition Anglaiseby pera1988
- HOUSE HEARING, 112TH CONGRESS - REVIEW OF DOE VEHICLE TECHNOLOGIES PROGRAM MANAGEMENT AND ACTIVITIES: ASSURING APPROPRIATE AND EFFECTIVE USE OF TAXPAYER FUNDINGby Scribd Government Docs

- Automotive Repair
- Parts Access
- 1996 E320 Owners Manual Pt1
- Heavy Combination Vehicles
- 2010 12 Steering Wheel Pull.pdf
- WAUZZZ8EX5A006226
- article486
- Hamm HD110
- 162_A Review of Yaw Rate and Sideslip Controllers for Passenger Vehicles
- TM 5-3805-254-14P-2 PART 7 IHC F-5070 DUMP TRUCK
- AC140-14247-Part2_en
- 938K_AEHQ6423-00
- Steering Gear Power
- CorvetteAssyProcess.pdf
- Alignment
- Auto Finance Manage,Ment System
- BRC
- advisor1
- 28026881-93ZJ-Secc-19-Steering
- Volvo Driver Information Display Manual
- Consumer Reports-How to Spot a Lemon
- VANET- Based Vehicle Tacking on Message Authentication and Secure Navigation Route
- Technical Note Edition Anglaise
- HOUSE HEARING, 112TH CONGRESS - REVIEW OF DOE VEHICLE TECHNOLOGIES PROGRAM MANAGEMENT AND ACTIVITIES
- Platinum Gazette 01 April 2016
- Transportation for Hire Regulations
- Purchasing a Used Vehicle Can Be Risky
- 10_12
- Cost Accounting/Series-4-2011(Code3017)
- Technical Data Book PDF Nl
- Vehicle Trajectory

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.