See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281070127

Perturbation Theory for Non-smooth Systems

Article · January 2009

DOI: 10.1007/978-0-387-30440-3_400

CITATIONS READS
82 587

1 author:

Marco Teixeira
University of Campinas
171 PUBLICATIONS 2,861 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ODE's and Geometry View project

Estabilidade assintótica e estrutural em sistemas dinâmicos descontínuos View project

All content following this page was uploaded by Marco Teixeira on 10 October 2018.

The user has requested enhancement of the downloaded file.

Perturbation Theory for Non-smooth
Systems
Marco Antonio Teixeiraa
a
Universidade Estadual de Campinas, Campinas, Brazil.

Article Outline
I. Glossary 1
II. Definition of the subject and its importance 1
III. Introduction 2
IV. Preliminaries 3
V. Vector fields near the boundary 6
VI. Generic Bifurcation in Ωr (n + 1) 11
VII. Singular Perturbation Problem in 2D 13
VIII. Future directions 15
IX. Bibliography 17
Primary Literature 17
Secondary Literature, further readings 20

I. Glossary
Non-smooth dynamical system: systems derived from ordinary differential
equations when the non-uniqueness of solutions is allowed. In this article we deal
with discontinuous vector fields in Rn where the discontinuities are concentrated in
a codimension-one surface.
Bifurcation: in a k − parameter families of systems, a bifurcation is a parameter
value at which the phase portrait is not structurally stable.
Typical singularity: are points on the discontinuity set where the orbits of the
system through them must be distinguished.

II. Definition of the subject and its importance

In this article we survey some qualitative and geometric aspects of non-smooth
dynamical systems theory. Our goal is to provide an overview of the state of the
art on the theory of contact between a vector field and a manifold, and on discon-
tinuous vector fields and their perturbations. We also establish a bridge between
2-dimensional non-smooth systems and the geometric singular perturbation theory.
Non-smooth dynamical systems is a subject that has been developing at a very
fast pace in recent years due to various factors: its mathematical beauty, its strong
relationship with others branches of science and the challenge in establishing rea-
sonable and consistent definitions and conventions. It has become certainly one
of the common frontiers between Mathematics and Physics/Engineering. We men-
tion that certain phenomena in control systems, impact in mechanical systems and
nonlinear oscillations are the main sources of motivation of our study concerning
1
2

the dynamics of those systems that emerge from differential equations with discon-
tinuous right-hand sides. We understand that non-smooth systems are driven by
applications and they play an intrinsic role in a wide range of technological areas.

III. Introduction
The purpose of this article is to present some aspects of the geometric theory of
a class of non-smooth systems. Our main concern is to bring the theory into the
domain of geometry and topology in a comprehensive mathematical manner.
Since this is an impossible task, we do not attempt to touch upon all sides of
this subject in one article. We focus in exploring the local behavior of systems
around typical singularities. The first task is to describe a generic persistence of a
local theory (structural stability and bifurcation) for discontinuous systems mainly
in the 2− and 3−dimensional cases. Afterwards we present some striking features
and results of the regularization process of two-dimensional discontinuous systems
in the framework developed by Sotomayor and Teixeira in [44] and establish a
bridge between those systems and the fundamental role played by the Geometric
Singular Perturbation Theory (GSPT). This transition was introduced in [10] and
we reproduce here its main features in the 2-dimensional case. For an introductory
reading on the methods of geometric singular perturbation theory we refer to [15],
[18] and [30]. In Section 2 we introduce the setting of this article. In Section 3 we
survey the state of the art of the contact between a vector field and a manifold. The
results contained in this section are crucial for the development of our approach. In
Section 4 we discuss the classification of typical singularities of non-smooth vector
fields. The study of non-smooth systems, via GSPT, is presented in Section 5. In
Section 6 some theoretical open problems are presented.
One aspect of the qualitative point of view is the problem of structural stability,
the most comprehensive of many different notions of stability. This theme was
studied in 1937 by Andronov-Pontryagin (see [4]). This problem is of obvious
importance, since in practice one obtains a lot of qualitative information not only
on a fixed system but also on its nearby systems.
We deal with non-smooth vector fields in Rn+1 having a codimension-one sub-
manifold M as its discontinuity set. The scheme in this work toward a systematic
classification of typical singularities of non-smooth systems follows the ideas devel-
oped by Sotomayor-Teixeira in [43] where the problem of contact between a vector
field and the boundary of a manifold was discussed. Our approach intends to be
self-contained and is accompanied by an extensive bibliography. We will try to
focus here on areas that are complimentary to some recent reviews made elsewhere.
The concept of structural stability in the space of non-smooth vector fields is
based on the following definition:
Definition 1. Two vector fields Z and Z̃ are C 0 equivalent if there is an M −invariant
homeomorphism h : Rn+1 → Rn+1 that sends orbits of Z to orbits of Z̃.
A general discussion is presented to study certain unstable non-smooth vector
fields within a generic context. The framework in which we shall pursue these
unstable systems is sometimes called generic bifurcation theory. In [4] the concept of
kth-order structural stability is also presented; in a local approach such setting gives
rise to the notion of a codimension-k singularity. In studies of classical dynamical
systems, normal form theory has been well accepted as a powerful tool in studying
 3

the local theory (see [6]). Observe that, so far, bifurcation and normal form theories
for non-smooth vector fields have not been extensively studied in a systematic way
Control Theory is a natural source of mathematical models of these systems (see,
for instance, [5], [8], [20], [41] and [45]). Interesting problems concerning discontin-
uous systems can be formulated in systems with hysteresis ([41]), economics ([25],
[23]) and biology ([7]). It is worth mentioning that in [3] a class of relay systems in
Rn is discussed. They have the form:
X = A x + sgn(x1 ) k
where x = (x1 , x2 , . . . , xn ), A ∈ MR (n, n) and k = (k1 , k2 , ...kn ) is a constant vector
in Rn . In [27] and [28] the generic singularities of reversible relay systems in 4D
were classified. In [54] some properties of non-smooth dynamics are discussed in
order to understand some phenomena that arise in chattering control. We mention
the presence of chaotic behavior in some non-smooth systems (see for example [12]).
It is worthwhile to cite [17], where the main problem in the classical calculus of vari-
ations was carried out to study discontinuous Hamiltonian vector fields. We refer
to [14] for a comprehensive text involving non-smooth systems which includes many
models and applications. In particular motivating models of several non-smooth
dynamical systems arising in the occurrence of impacting motion in mechanical
systems, switchings in electronic systems and hybrid dynamics in control systems
are presented together with an extensive literature on impact oscillators which we
do not attempt to survey here. For further reading on some mathematical aspects
of this subject we recommend [11] and references therein. A setting of general as-
pects of non-smooth systems can be found also in [35] and references therein. Our
discussion does not focus on continuous but rather on non-smooth dynamical sys-
tems and we are aware that the interest in this subject goes beyond the approach
adopted here.
The author wishes to thank R. Garcia, T.M. Seara and J. Sotomayor for many
helpful conversations.

IV. Preliminaries
Now we introduce some of the terminology, basic concepts and some results that
will be used in the sequel.

Definition 2. Two vector fields Z and Z̃ on Rn with Z(0) = Z̃(0) are germ-
equivalent if they coincide on some neighborhood V of 0.

The equivalent classes for this equivalence are called germs of vector fields. In
the same way as defined above, we may define germs of functions. For simplicity we
are considering the germ notation and we will not distinguish a germ of a function
and any one of its representatives. So, for example, the notation h : Rn , 0 → R
means that the h is a germ of a function defined in a neighborhood of 0 in Rn .
Refer to [16] for a brief and nice introduction of the concepts of germ and k-jet of
functions.

IV.1. Discontinuous Systems. Let M = h−1 (0), where h is (a germ of) a smooth
function h : Rn+1 , 0 −→ R having 0 ∈ R as its regular value. We assume that
0 ∈ M.
4

Designate by χ(n + 1) the space of all germs of C r vector fields on Rn+1 at 0

endowed with the C r –topology with r > 1 and large enough for our purposes. Call
Ω(n + 1) the space of all germs of vector fields Z in Rn+1 , 0 such that
½
X(q), for h(q) > 0,
(1) Z(q) =
Y (q), for h(q) < 0,
The above field is denoted by Z = (X, Y ). So we are considering Ω(n + 1) =
χ(n + 1) × χ(n + 1) endowed with the product topology.
Definition 3. We say that Z ∈ Ω(n + 1) is structurally stable if there exists a
neighborhood U of Z in Ω(n + 1) such that every Z̃ ∈ U is C 0 − equivalent with Z.
To define the orbit solutions of Z on the switching surface M we take a pragmatic
approach. In a well characterized open set O of M (described below) the solution
of Z through a point p ∈ O obeys the Filippov rules and on M − O we accept it to
be multivalued. Roughly speaking, as we are interested in studying the structural
stability in Ω(n + 1) it is convenient to take into account all the leaves of the
foliation in Rn+1 generated by the orbits of Z (and also the orbits of X and Y )
passing through p ∈ M. (see Figure 1)
The trajectories of Z are the solutions of the autonomous differential system
q̇ = Z(q).
In what follows we illustrate our terminology by presenting a simplified model
that is found in the classical electromagnetism theory (see for instance [26]):
...
ẍ − x + αsignx = 0.
with α > 0.
So this system can be expressed by the following objects: h(x, y, z) = x and
Z = (X, Y ) with X(x, y, z) = (y, z, z + α) and Y (x, y, z) = (y, z, z − α).
For each X ∈ χ(n + 1) we define the smooth function Xh : Rn+1 → R given by
Xh = X · ∇h where · is the canonical scalar product in Rn+1 .
We distinguish the following regions on the discontinuity set M :
(i) M1 is the sewing region that is represented by h = 0 and (Xh)(Y h) > 0;
(ii) M2 is the escaping region that is represented by h = 0, (Xh) > 0 and (Y h) < 0;
(iii) M3 is the sliding
S region that is represented by h = 0, (Xh) < 0 and (Y h) > 0.
We set O = i=1,2,3 Mi .
Consider Z = (X, Y ) ∈ Ω(n + 1) and p ∈ M3 . In this case, following Filippov’s
convention, the solution γ(t) of Z through p follows, for t ≥ 0, the orbit of a vector
field tangent to M. Such system is called sliding vector field associated to Z and it
will be defined below.
Definition 4. The sliding vector field associated to Z = (X, Y ) is the smooth vector
field Z s tangent to M and defined at q ∈ M3 by Z s (q) = m − q with m being the
point where the segment joining q + X(q) and q + Y (q) is tangent to M .
It is clear that if q ∈ M3 then q ∈ M2 for −Z and then we define the escaping
vector field on M2 associated to Z by Z e = −(−Z)s . In what follows we use the
notation Z M for both cases.
We recall that sometimes Z M is defined in an open region U with boundary. In
this case it can be C r extended to a full neighborhood of p ∈ ∂U in M.
 5

sewing sliding

sliding

Figure 1. A discontinuous system and its regularization.

S
When the vectors X(p) and Y (p), with p ∈ M2 M3 are linearly dependent
then Z M (p) = 0. In this case we say that p is a simple singularity of Z. The others
singularities of Z are concentrated outside the set O.
We finish this subsection with a 3 − dimensional example:
Let Z = (X, Y ) ∈ Ω(3) with h(x, y, z) = z, X = (1, 0, x) and Y = (0, 1, y).
The system determines around 0 four quadrants, bounded by τX = {x = 0} and
τY = {y = 0}. They are: Q+ −
1 = {x > 0, y > 0}, Q1 = {x < 0, y < 0}, Q2 = {x <
0, y > 0} (sliding
S − region) and Q3 = {x > 0, y < 0} (escaping region). Observe that
M1 = Q+ 1 Q1 .
The sliding vector field defined in Q2 is expressed by:

y+x
Z s (x, y, z) = (y − x)−1 (x + y, , 0))
8
Such system is (in Q2 ) equivalent to G(x, y, z) = (x+y, y+x 8 , 0)). In our terminol-
ogy we consider G a smooth extension of Z s , that is defined in a whole neighborhood
of 0. It is worth to say that G is in fact a system which is equivalent to the original
system in Q2 .
In [50] a generic classification of one-parameter families of sliding vector fields is
presented.

IV.2. Singular perturbation problem. A singular perturbation problem is ex-

pressed by a differential equation z ′ = α (z, ε) (refer to [18], [15] and [30]) where
z ∈ Rn+m , ε is a small non-negative real number and α is a C ∞ mapping.
Let z = (x, y) ∈ Rn+m and f : Rm+n → Rm , g : Rm+n → Rn be smooth
mappings. We deal with equations that may be written in the form
½ ′
x = f (x, y, ε)
(2) x = x (τ ) , y = y (τ ) .
y ′ = εg (x, y, ε)
An interesting model of such systems can be obtained from the singular van der
Pol’s equation
(3) εx′′ + (x2 + x)x′ + x − a = 0.
6

The main trick in the geometric singular perturbation (GSP) is to consider the
family (2) in addition to the family
½
εẋ = f (x, y, ε)
(4) x = x (t) , y = y (t)
ẏ = g (x, y, ε)
obtained after the time rescaling t = ετ .
Equation (2) is called the fast system and (4) the slow system. Observe that for
ε > 0 the phase portrait of fast and slow systems coincide.
For ε = 0, let S be the set of all singular points of (2). We call S the slow
manifold of the singular perturbation problem and it is important to notice that
equation (4) defines a dynamical system on S called the reduced problem.
Combining results on the dynamics of these two limiting problems (2) and (4),
with ε = 0, one obtains information on the dynamics for small values of ε. In fact,
such techniques can be exploited to formally construct approximated solutions on
pieces of curves that satisfy some limiting version of the original equation as ε goes
to zero.
Definition 5. Let A, B ⊂ Rn+m be compact sets. The Hausdorff distance between
A and B is D (A, B) = maxz1 ∈A,z2 ∈B {d (z1 , B) , d (z2 , A)}.
The main question in GSP-theory is to exhibit conditions under which a singular
orbit can be approximated by regular orbits for ε ↓ 0, with respect to the Hausdorff
distance.
IV.3. Regularization Process. An approximation of the discontinuous vector
field Z = (X, Y ) by a one-parameter family of continuous vector fields will be called
a regularization of Z. In [44], Sotomayor and Teixeira introduced the regularization
procedure of a discontinuous vector field. A transition function is used to average
X and Y in order to get a family of continuous vector fields that approximates the
discontinuous one. Figure 1 gives a clear illustration of the regularization process.
Let Z = (X, Y ) ∈ Ω(n + 1).
Definition 6. A C ∞ function ϕ : R −→ R is a transition function if ϕ(x) = −1
for x ≤ −1, ϕ(x) = 1 for x ≥ 1 and ϕ′ (x) > 0 if x ∈ (−1, 1). The ϕ–regularization
of Z = (X, Y ) is the 1–parameter family Xε ∈ C r given by
µ ¶ µ ¶
1 ϕε (h(q)) 1 ϕε (h(q))
(5) Zε (q) = + X(q) + − Y (q).
2 2 2 2
with h given in subsection 2.1 and ϕε (x) = ϕ(x/ε), for ε > 0.
As already said before, a point in the phase space which moves on an orbit of Z
crosses M when it reaches the region M1 . Solutions of Z through points of M3 , will
remain in M in forward time. Analogously, solutions of Z through points of M2
will remain in M in backward time. In [34] and [44] such conventions are justified
by the regularization method in dimensions 2 and 3 respectively.

V. Vector fields near the boundary

In this section we discuss the behavior of smooth vector fields in Rn+1 relative to
a codimension one submanifold (say, the above defined M ). We base our approach
on the concepts and results contained in [43] and [53]. The principal advantage
of this setting is that the generic contact between a smooth vector field and M
 7

can often be easily recognized. As an application the typical singularities of a

discontinuous system can be further classified in a straightforward way.
We say that X, Y ∈ χ(n+1) are M −equivalent if there exists an M −preserving
homeomorphism h : Rn+1 , 0 −→ Rn+1 , 0 that sends orbits of X into orbits of Y. In
this way we get the concept of M −structural stability in χ(n + 1).
We call Γ0 (n + 1) the set of elements X in χ(n + 1) satisfying one of the following
conditions:
0) Xh(0) 6= 0 (0 is a regular point of X). In this case X is transversal to M at 0.
1) Xh(0) = 0 and X 2 h(0) 6= 0 (0 is a 2-fold point of X;)
2) Xh(0) = X 2 h(0) = 0, X 3 h(0) 6= 0 and the set {Dh(0), DXh(0), DX 2 h(0)} is
linearly independent (0 is a cusp point of X;)
......................................................
n) Xh(0) = X 2 h(0) = ... = X n h(0) = 0 and X n+1 h(0) 6= 0. Moreover the set
{Dh(0), DXh(0), DX 2 h(0), ...., DX n h(0)} is linearly independent, and 0 is a regu-
lar point of the mapping Xh|M .
We say that 0 is an M −singularity of X if h(0) = Xh(0) = 0. It is a codimension−
zero M −singularity provided that X ∈ Γ0 (n + 1).
We know that Γ0 (n + 1) is an open and dense set in χ(n + 1) and it coincides
with the M −structurally stable vector fields in χ(n + 1) (see [53]).
Denote by τX ⊂ M the M −singular set of X ∈ χ(n + 1); this set is represented
by the equations h = Xh = 0. It is worth to point out that, generically, all 2−f olds
constitute an open and dense subset of τX . Observe that if X(0) = 0 then X 6∈
Γ0 (n + 1).
The M −bifurcation set is represented by χ1 (n + 1) = χ(n + 1) − Γ0 (n + 1)
Vishik in [53] exhibited the normal forms of a codimension−zero M −singularity.
They are:
I) Straightened vector field

X = (1, 0, ...., 0)
and
h(x) = xk+1
1 + x2 xk−1
1 + x3 xk−2
1 + ... + xk+1 k = 0, 1...n
or
II) Straightened boundary

h(x) = x1
and
X(x) = (x2 , x3 , ...., xk , 1, 0, 0...0)
We now discuss an important interaction between vector fields near M and
singularities of mapping theory. We discuss how singularity-theoretic techniques
help the understanding of the dynamics of our systems.
We outline this setting, which will be very useful in the sequel. The starting
point is the following construction.
V.1. A construction. Let X ∈ χ(n + 1). Consider a coordinate system x =
(x1 , x2 , ...., xn+1 ) in Rn+1 , 0 such that

M = {x1 = 0}
and
8

M M

cusp
singularity

Figure 2. The cusp singularity and its unfolding.

X = (X 1 , X 2 , ....., X n+1 )
Assume that X(0) = 6 0 and X 1 (0) = 0. Let N0 be any transversal section to X
at 0.
By the implicit function theorem, we derive that:
for each p ∈ M, 0 there exists a unique t = t(p) in R, 0 such that the orbit-solution
t 7→ γ(p, t) of X through p meets N0 at a point p̃ = γ(p, t(p)).
We define the smooth mapping ρX : Rn , 0 −→ Rn , 0 by ρX (p) = p̃. This
mapping is a powerful tool in the study of vector fields around the boundary of a
manifold (refer to [21], [42] [43], [53] and [47]). We observe that τX coincides with
the singular set of ρX .
The late construction implements the following method. If we are interested
in finding an equivalence between two vector fields which preserve M , then the
problem can be sometimes reduced to finding an equivalence between ρX and ρY
in the sense of singularities of mappings.
We recall that when 0 is a fold M −singularity of X then associated to the fold
mapping ρX there is the symmetric diffeomorphism βX that satisfies ρX ◦βX = ρX .
Given Z = (X, Y ) ∈ Ω(n + 1) such that ρX and ρY are fold mappings with
X 2 h(0) < 0 and Y 2 h(0) > 0 then the composition of the associated symmetric
mappings βX and βY provides a first return mapping βZ associated to Z and M .
This situation is usually called a distinguished f old − f old singularity, and the
mapping βZ plays a fundamental role in the study of the dynamics of Z.
V.2. Codimension-one M-singularity in dimensions 2 and 3.
V.2.1. Case n=1. In this case the unique codimension-zero M −singularity is a fold
point in R2 , 0. The codimension-one M −singularities are represented by the subset
Γ1 (2) of χ1 (2) and it is defined as follows.
Definition 7. A codimension-one M −singularity of X ∈ Γ1 (2) is either a cusp
singularity or an M −hyperbolic critical point p in M of the vector field X. A
cusp singularity (illustrated in Figure 2) is characterized by Xh(p) = X 2 h(p) = 0,
X 3 h(p) 6= 0. In the second case this means that p is a hyperbolic critical point
(illustrated in Figure 3) of X with distinct eigenvalues and with invariant manifolds
(stable, unstable and strong stable and strong unstable) transversal to M .
In this subsubsection we consider a coordinate system in R2 , 0 such that h(x, y) =
y.
 9

saddle point
X

M M

Figure 3. The saddle point in the boundary and its unfolding.

The next result was proved in [47]. It presents the normal forms of the codimen-
sion one singularities defined above.

Theorem 8. Let X ∈ χ1 (2). The vector field X is M −structurally stable relative

to χ1 (2) if and only if X ∈ Γ1 (2). Moreover, Γ1 (2) is an embedded codimension-one
sub manifold and dense in χ1 (2). We still require that any one-parameter family
Xλ , (λ ∈ (−ε, ε)) in χ(1) transverse to Γ1 (2) at X0 , has one of the following normal
forms:
0.1 : Xλ (x, y) = (1, 0) (regular point);
0.2 : Xλ (x, y) = (1, x) (fold singularity);
1.1 Xλ (x, y) = (1, λ + x2 ) (cusp singularity);
1.2 Xλ (x, y) = (ax, x + by + λ), a = ±1, b = ±2;
1.3 Xλ (x, y) = (x, x − y + λ);
1.4 Xλ (x, y) = (x + y, −x + y + λ).
V.2.2. Case n=2.
Definition 9. A vector field X ∈ χ(3) belongs to the set Γ1 (a) if the following
conditions hold:

(i) X(0) = 0 and 0 is a hyperbolic critical point of X;

(ii) the eigenvalues of DX(0) are pairwise distinct and the corresponding eigenspaces
are transversal to M at 0;

(iii) each pair of non complex conjugate eigenvalues of DX(0) has distinct real
parts.
Definition 10. A vector field X ∈ χ(3) belongs to the set Γ1 (b) if X(0) 6= 0,
Xh(0) = 0, X 2 h(0) = 0 and one of the following conditions hold:
(1) X 3 h(0) 6= 0, rank{Dh(0), DXh(0), DX 2 h(0)} = 2 and 0 is a non-degenerate
critical point of Xh|M .
(2) X 3 h(0) = 0, X 4 h(0) 6= 0 and 0 is a regular point of Xh|M .
The next results can be found in [43].
Theorem 11. The following statements hold:
(i) Γ1 (3) = Γ1 (a) ∪ Γ2 (b) is a codimension one submanifold of χ(3).
10

(ii) Γ1 (3) is open and dense set in χ1 (3) in the topology induced from χ1 (3).
(iv) For a residual set of smooth curves γ : R, 0 → χ(3), γ meets Γ1 (3) transversally.

Throughout this subsubsection we fix the function h(x, y, z) = z.

Lemma 12. (Classification Lemma) (1) The elements of Γ1 (3) are classified as
follows:
(a11 ) Nodal M-Singularity:: X(0) = 0, the eigenvalues of DX(0), λ1 , λ2 , and λ3 ,
are real, distinct, λ1 λj > 0, j = 2, 3 and the eigenspaces are transverse to
M at 0;
(a12 ) Saddle M-Singularity:: X(0) = 0, the eigenvalues of DX(0), λ1 , λ2 and λ3 ,
are real, distinct, λ1 λj < 0, j = 2 or 3 and the eigenspaces are transverse
to M at 0;
(a13 ) Focal M-Singularity:: 0 is a hyperbolic critical point of X , the eigen-
values of DX(0) are λ12 = a ± ib, λ3 = c, with a, b, c distinct from zero and
c 6= a, and the eigenspaces are transverse to M at 0.
(b11 ) Lips M-Singularity:: presented in Definition 8, item 1, when Hess(F h/S (0)) >
0:
(b12 ) Bec to Bec M-Singularity:: presented in Definition 8, item 1, when
Hess(F h/S (0)) < 0;
(b13 ) Dove’s Tail M-Singularity: presented in Definition 8, item 2.

The next result is proved in [38]. It deals with the normal forms of a codimension
one singularity.

Theorem 13. i) (Generic Bifurcation and normal forms) Let X ∈ χ(3). The
vector field X is M −structurally stable relative to χ1 (3) if and only if X ∈ Γ1 (3).
ii) (Versal unfolding) In the space of one-parameter families of vector fields Xα in
χ(3), α ∈ (−ε, ε) an everywhere dense set is formed by generic families such that
their normal forms are:
• Xα ∈ Γ0 (3)
0.1 : Xα (x, y, z) = (0, 0, 1)
0.2 : Xα (x, y, z) = (z, 0, ±x)
0.3 : Xα (x, y, z) = (z, 0, x2 + y)
• X0 ∈ Γ1 (3)
2 2
1.1 : Xα (x, y, z) = (z, 0, −3x +y
2
+α
)
2
−y 2 +α
1.2 : Xα (x, y, z) = (z, 0, −3x 2 )
4δx3 +y+αx
1.3 : Xα (x, y, z) = (z, 0, 2 ), with δ = ±1
ax+by+cz 2 +α
1.4 : Xα (x, y, z) = (axz, byz, 2 ), with
(a, b, c) = δ(3, 2, 1), δ =
±1
2
1.5 : Xα (x, y, z) = (axz, byz, ax+by+cz
2
+α
), with (a, b, c) = δ(1, 3, 2), δ =
±1
2
1.6 : Xα (x, y, z) = (axz, byz, ax+by+cz
2
+α
), with (a, b, c) = δ(1, 2, 3), δ =
±1
2
1.7 : Xα (x, y, z) = (xz, 2yz, x+2y−cz
2
+α
)
2
1.8 : Xα (x, y, z) = ((−x + y)z, (−x − y)z, −3x−y+z
2
+α
)
 11

VI. Generic Bifurcation in Ωr (n + 1)

Let Z = (X, Y ) ∈ Ωr (n + 1). Call by Σ0 (n + 1) (resp. Σ1 (n + 1)) the set of all ele-
ments that are structurally stable in Ωr (n+1) (resp. Ωr1 (n+1) = Ωr (n+1)\Σ0 (n+1) )
in Ωr (n + 1). It is clear that a pre-classification of the generic singularities is im-
mediately reached by:

If Z = (X, Y ) ∈ Σ0 (n + 1) (resp. Z = (X, Y ) ∈ Σ1 (n + 1)) then X and Y are in

Γ0 (n + 1) (resp. X ∈ Γ0 (n + 1) and Y ∈ Γ1 (n + 1) or vice versa). Of course, the
case when both X and Y are in Γ1 (n + 1) is a codimension two phenomenon.
VI.1. 2-dimensional case. The following result characterizes the structural sta-
bility in Ωr (2).
Theorem A: (see [31] and [44]): Σ0 (2) is an open and dense set of Ωr (2). The
vector field Z = (X, Y ) is in Σ0 (2) if and only one of the following conditions is
satisfied:
i) Both elements X and Y are regular. When 0 ∈ M is a simple singularity of Z
then we assume that it is a hyperbolic critical point of Z M .
ii) X is a fold singularity and Y is regular (and vice-versa).
The following result still deserves a systematic proof. Following the same strategy
stipulated in the generic classification of an M −singularity, Theorem 11 could be
very useful. It is worth to mention [33] where the problem of generic bifurcation in
2D was also addressed.
Theorem B (Generic Bifurcation): (see [37] and [43]) Σ1 (2) is an open and
dense set of Ωr1 (2). The vector field Z = (X, Y ) is in Σ1 (2) provided that one of the
following conditions is satisfied:
i) Both elements X and Y are M −regular. When 0 ∈ M is a simple singularity
of Z then we assume that it is a codimension one critical point (saddle-node or a
Bogdanov-Takens singularity) of Z M .
ii) 0 is a codimension-one M −singularity of X and Y is M −regular. This case in-
cludes when 0 is either a cusp M −singularity or a critical point. Figure 4 illustrates
the case when 0 is a saddle critical point in the boundary.
ii) both X and Y are fold M −singularities at 0. In this case we have to impose that
0 is a hyperbolic critical point of the C r − extension of Z M provided that it is in
the boundary of M2 ∪ M3 (see example below). Moreover when 0 is a distinguished
f old − f old singularity of Z then 0 is a hyperbolic fixed point of the first return
mapping βZ .
Consider in a small neighborhood of 0 in R2 , the system Z = (X, Y ) with
X(x, y) = (1 − x3 + y 2 , x), Y (x, y) = (1 + x + y, −x + x2 ) and h(x, y) = y. The
point 0 is a f old − f old-singularity of Z with M2 = {x < 0} and Z s (x, 0) =
(2x − x2 )−1 (2x − x4 + x5 ). Observe that 0 is a hyperbolic critical point of the
extended system G(x, y) = 2x − x4 + x5 .
The classification of the codimension two singularities in Ωr (2) is still an open
problem. In this direction [51] contains information about the classification of
codimension two M −singularities.
VI.2. 3-dimensional case. Let Z = (X, Y ) ∈ Ωr (3).
The most interesting case to be analyzed is when both vector fields, X and Y are
fold singularities at 0 and the tangency sets τX and τY in M are in general position
at 0. In fact they determine (in M ) four quadrants, two of them are M1 −regions,
12

saddle point
X

M M
Y

Figure 4. M −critical point for X, M −regular for Y and its unfolding.

escaping region sewing region

$Y−$tangency line

sewing region

$X$ $M$

$X−$tangency line

$YY sliding region

Figure 5. The distinguished fold-fold singularity.

one is an M3 −region and the other is an M2 −region (see Figure 5). We emphasize
that the sliding vector field Z M can be C r −extended to a full neighborhood of 0 in
M . Moreover, Z M (0) = 0. Inside this class the distinguished f old − f old singularity
(as defined in Subsection 3.1) must be taken into account. Denote by A the set of
all distinguished f old√− f old singularities Z ∈ Ωr (3). Moreover, the eigenvalues of
DβZ (0) are λ = a ± (a2 − 1). If λ ∈ R we say that Z belongs to As . Otherwise
Z is in Ae . Recall that βZ is the first return mapping associated to Z and M at 0
as defined in Subsection 3.1.
It is evident that the elements in the open set Ae are structurally unstable
in Ωr (3). It is worthwhile to mention that in Ae we detect elements which are
asymptotically stable at the origin [48]. Concerning As few things are known.
We have the following result:
Theorem C: The vector field Z = (X, Y ) belongs to Σ0 (3) provided that one of
the following conditions occurs:
i) Both elements X and Y are regular. When 0 ∈ M is a simple singularity of Z
then we assume that it is a hyperbolic critical point of Z M .
ii) X is a fold singularity at 0 and Y is regular.
iii) X is a cusp singularity at 0 and Y is regular.
 13

iv) Both systems X and Y are of fold type at 0. Moreover: a) the tangency sets τX
and τY are in general position at 0 in M ; b) The eigenspaces associated to Z M are
transverse to τX and τY at 0 ∈ M and c) Z is not in A. Moreover the real parts
of non conjugate eigenvalues are distinct.
We recall that bifurcation diagrams of sliding vector fields are presented in [50]
and [52].

VII. Singular Perturbation Problem in 2D

Geometric singular perturbation theory is an important tool in the field of con-
tinuous dynamical systems. Needless to say that in this area very good surveys are
available (refer to [15], [18] and [30]). Here we highligh some results (see [10]) that
bridge the space between discontinuous systems in Ωr (2) and singularly perturbed
smooth systems.
Definition 14. Let U ⊂ R2 be an open subset and ε ≥ 0. A singular perturbation
problem in U ( SP–Problem) is a differential system which can be written as

(6) x′ = dx/dτ = f (x, y, ε), y ′ = dy/dτ = εg(x, y, ε)

or equivalently, after the time re-scaling t = ετ

(7) εẋ = εdx/dt = f (x, y, ε), ẏ = dy/dt = g(x, y, ε),

with (x, y) ∈ U and f, g smooth in all variables.
Our first result is concerned with the transition between non-smooth systems
and GSPT.
Theorem D: Consider Z ∈ Ωr (2), Zε its ϕ–regularization, and p ∈ M . Sup-
pose that ϕ is a polynomial of degree k in ©a small interval I ⊆ ª(−1, 1) with
0 ∈ I. Then the trajectories of Zε in Vε = q ∈ R2 , 0 : h(q)/ε ∈ I are in cor-
respondence with the solutions of an ordinary differential equation z ′ = α(z, ε),
satisfying that α is smooth in both variables and α(z, 0) = 0 for any z ∈ M. More-
over, if ((X − Y )hk )(p) 6= 0 then we can take a C r−1 –local coordinate system
{(∂/∂x) (p), (∂/∂y) (p)} such that this smooth ordinary differential equation is a
SP–problem.

The understanding of the phase portrait of the vector field associated to a

SP–problem is the main goal of the geometric singular perturbation–theory (GSP–
theory). The techniques of GSP–theory can be used to obtain information on the
dynamics of (6) for small values of ε > 0, mainly in searching minimal sets.
System (6) is called the fast system, and (7) the slow system of SP–problem.
Observe that for ε > 0 the phase portraits of the fast and the slow systems coincide.
Theorem D says that we can transform a discontinuous vector field in a SP–
problem. In general this transition can not be done explicitly. Theorem E provides
an explicit formula of the SP–problem for a suitable class of vector fields. Before
the statement of such a result we need to present some preliminaries.
Consider C = {ξ : R2 , 0 → R} with ξ ∈ C r and L(ξ) = 0 where L(ξ) denotes
the linear part of ξ at (0, 0).
Let Ωd ⊂ Ωr (2) be the set of vector fields Z = (X, Y ) in Ωr (2) such that there
exists ξ ∈ C that is a solution of
14

(8) ∇ξ(X − Y ) = Πi (X − Y ),
where ∇ξ is the gradient of the function and Πi denote the canonical projections,
for i = 1 or i = 2.
Theorem E: Consider Z ∈ Ωd and Zε its ϕ–regularization. Suppose that ϕ is a
polynomial of degree
© k in a small interval
ª I ⊂ R with 0 ∈ I. Then the trajectories
of Zε on Vε = q ∈ R2 , 0 : h(q)/ε ∈ I are solutions of a SP–problem.
We remark that the singular problems discussed in the previous theorems, when
ε ց 0, defines a dynamical system on the discontinuous set of the original problem.
This fact can be very useful for problems in Control Theory.
Our third theorem says how the fast and the slow systems approximate the
discontinuous vector field. Moreover, we can deduce from the proof that whereas the
fast system approximates the discontinuous vector field, the slow system approaches
the corresponding sliding vector field.
Consider Z ∈ Ωr (2) and ρ : R2 , 0 −→ R with ρ(x, y) being the distance between
(x, y) and M. We denote by Z b the vector field given by Z(x,
b y) = ρ(x, y)Z(x, y).
b
In what follows we identify Zε and the vector field on {{R2 , 0} \ M × R} ⊂ R3
b y, ε) = (Z
given by Z(x, bε (x, y), 0).

Theorem F: Consider p = 0 ∈ M . Then there exist an open set U ⊂ R2 , p ∈ U ,

a 3–dimensional manifold M , a smooth function Φ : M −→ R3 and a SP–problem
b (U ×(0,+∞)) .
W on M such that Φ sends orbits of W |Φ−1 (U ×(0,+∞)) in orbits of Z|

VII.1. Examples. 1- Take X(x, y) = (1, x), Y (x, y) = (−1, −3x), and h(x, y) = y.
The discontinuity set is {(x, 0) | x ∈ R}. We have Xh = x, Y h = −3x, and then
the unique non–regular point is (0, 0). In this case we may apply Theorem E.

2- Let Zε (x, y) = (y/ε, 2xy/ε − x) . The associated partial differential equation

(refer to Theorem E) with i = 2 given above becomes 2 (∂ξ/∂x) + 4x (∂ξ/∂y) = 4x.
We take the coordinate change x = x, y = y − x2 . The trajectories of Xε in these
coordinates are the solutions of the singular system
εẋ = y + x2 , ẏ = −x.

3- In what follows we try, by means an example, to present a rough idea on the

transition from non-smooth systems to GSPT. Consider X(x, y) = (3y 2 − y − 2, 1),
Y (x, y) = (−3y 2 − y + 2, −1) and h(x, y) = x. The regularized vector field is
µ ¶ µ ¶
1 1 x 1 1 x
Zε (x, y) = + ϕ( ) (3y 2 − y − 2, 1) + − ϕ( ) (−3y 2 − y + 2, −1).
2 2 ε 2 2 ε
After performing the polar blow up coordinates α : [0, +∞) × [0, π] × R → R3
given by x = r cos θ and ε = r sin θ the last system is expressed by:

rθ̇ = − sin θ(−y + ϕ(cot θ)(3y 2 − 2)), ẏ = ϕ(cot θ).

So the slow manifold is given implicitly by ϕ(cot θ) = 3y2y−2 which defines two
√ √
1+ 1+24ϕ2 (cot θ) 1− 1+24ϕ2 (cot θ)
functions y1 (θ) = 6ϕ(cot θ) and y 2 (θ) = 6ϕ(cot θ) . The function y1 (θ)
 15

π 0

Figure 6. Example of fast and slow dynamics of the SP–Problem

is increasing, y1 (0) = 1, lim y1 (θ) = +∞, lim y1 (θ) = −∞ and y1 (π) = −1.
θ→ 2 −
π π
θ→ 2 +
2
The function y2 (θ) is increasing, y2 (0) = − 3 , limπ y2 (θ) = 0 and y2 (π) = 23 . We can
θ→ 2
extend y2 to (0, π) as a differential function with y2 ( π2 ) = 0.
The fast vector field is (θ′ , 0) with θ′ > 0 if (θ, y) belongs to
h π [ π [ π i
(0, ) × (y2 (θ), y1 (θ)) ( , π) × (y2 (θ), +∞) ( , π) × (−∞, y1 (θ))
2 2 2
′
and with θ < 0 if (θ, y) belongs to
h π [ π [ π i
(0, ) × (y1 (θ), +∞) (0, ) × (−∞, y2 (θ)) ( , π) × (y1 (θ), y2 (θ)) .
2 2 2
The reduced flow has one singular point at (0, 0) and it takes the positive direc-
tion of the y-axis if y ∈ (− 32 , 0) ∪ (1, ∞) and the negative direction of the y-axis if
y ∈ (−∞, −1) ∪ (0, 23 ).
One can see that the singularities (θ, y, r) = (0, 1, 0) and (θ, y, r) = (0, −1, 0) are
not normally hyperbolic points. In this way, as usual, we perform additional blow
ups. In Figure 6 we illustrate the fast and the slow dynamics of the SP–problem.
We present a phase portrait on the blowing up locus where double arrow over a
trajectory means that the trajectory belongs to the fast dynamical system, and
simple arrow means that the trajectory belongs to the slow dynamical system.

VIII. Future directions

Our concluding section is devoted to an outlook. Firstly we present some open
problems linked with the setting that point out future directions of research. The
main task for the future seems to bring the theory of non-smooth dynamical systems
to a similar maturity as that of smooth systems. Finally we briefly discuss the main
results in this text.
VIII.1. Some problems. In connection to this present work, some theoretical
problems remain open:
1- The description of the bifurcation diagram of the codimension-two singularities in
Ω(2). In this last class we find some models (see [36]) where the following questions
can also be addressed: a) when is a typical singularity topological equivalent to a
16

regular center?; b) how about the isochronicity of such a center?; c) when does a
polynomial perturbation of such a system in Ω(2) produce limit cycles? The articles
[13], [9] , [22], [21] and [46] can be useful auxiliary references.
2- Let Ω(N ) be the set of all non-smooth vector fields on a two-dimensional compact
manifold N having a codimension one compact submanifold M as its discontinuity
set. The problem is to study the global generic bifurcation in Ω(N ). The articles
[31], [33], [47] and [40] can be useful auxiliary references.
3- Study of the bifurcation set in Ωr (3). The articles [38], [43], [50] and [40] can be
useful auxiliary references.
4- Study of the dynamics of the distinguished f old − f old singularity in Ωr (n + 1).
The article [48] can be an useful auxiliary reference.
5- In many applications examples of non-smooth systems where the discontinu-
ities are located on algebraic varieties are available. For instance, consider the
system ẍ + xsign(x) + sign(ẋ) = 0. Motivated by such models we present the
following problem. Let 0 be a non-degenerate critical point of a smooth mapping
h : Rn+1 , 0 → R, 0. Let Φ(n + 1) be the space of all vector fields Z on Rn+1 , 0
defined in the same way as Ω(n + 1). We propose: i) classify the typical singular-
ities in that space; ii) Analyze the elements of Φ(2) by means of ”regularization
processes”’ and the methods of GSPT, similarly to Section 5. The articles [1] and
[2] can be very useful auxiliary references.
6- In [28] and [29] classes of 4D−relay systems are considered. Conditions for
the existence of one-parameter families of periodic orbits terminating at typical
singularities are provided. We propose to find conditions for the existence of such
families for n − dimensional relay systems.

VIII.2. Conclusion. In this paper we have presented a compact survey of the

geometric/qualitative theoretical features of non-smooth dynamical systems. We
feel that our survey illustrates that this field is still in its early stages but enjoying
growing interest. Given the importance and the relevance of such theme, we have
pointed above some open questions and we remark that there is still a wide range
of bifurcation problems to be tackled. A brief summary of the main results in the
text is given below.
1- We firstly deal with 2 − dimensional non-smooth vector fields Z = (X, Y )
defined around the origin in R2 , where the discontinuity set is concentrated on the
line {y = 0}. The first task is to characterize those systems which are structurally
stable. This characterization is a starting point to establish a bifurcation theory as
indicated by the Thom-Smale program.
2- In higher dimension the problem becomes much more complicated. We have
presented here sufficient conditions for the 3−dimensional local structural stability.
Any further investigation on bifurcation in this context must pass through a deep
analysis of the so called f old − f old singularity.
3- We have established a bridge between discontinuous and singularly perturbed
smooth systems. Many similarities between such systems were observed and a
comparative study of the two categories is called for.
 17

IX. Bibliography
Primary Literature
[1] Alexander, James C.; Seidman, Thomas I. Sliding modes in intersecting
switching surfaces I: Blending, Houston J. Math. 24 (1998), no. 3, 545–569.
[2] Alexander J.C. and Seidman, T. Sliding modes in intersecting switching
surfaces II: hysteresis, Houston J. Math.V. 25, (1999), 185–211.
[3] Anosov D.V., Stability of the equilibrium positions in relay systems, Au-
tomation and remote control, V.XX, 2, (1959).
[4] Andronov A. and Pontryagin S., Structurally stable systems, Dokl. Akad.
Nauk SSSR, V. 14, (1937), 247-250.
[5] Andronov A. A., Vitt A. A. and Khaikin, S. E., Theory of ocillators,
Dover, New York, (1966).
[6] Arnold, V.I. Methods in the theory of ordinary differential equations ,
Springer-Verlag, New York, (1983).
[7] Bazykin A.D., Nonlinear dynamics of interacting populations, World Sc.
Publ. Co. Inc., River-Edge, NJ , (1998).
[8] Bonnard B. and Chyba M. Singular trajectories and their role in control
theory, Mathématiques and Applications V.40, Springer Verlag (2000).
[9] Broucke M.E., Pugh,C.C. and Simić,S.N., Structural stability of piecewise
smooth systems, Computational and Applied Mathematics V. 20, (2001),
no.1-2, pp. 51–89.
[10] Buzzi C., Silva P.R. and Teixeira M.A.. A singular approach to discon-
tinuous vector fields on the plane: , J. of Differential Eq., V. 23, (2006),
633-655.
[11] Chillingworth D. R. J.(4-SHMP) Discontinuity geometry for an impact
oscillator Dyn. Syst., V. 17, (2002), no. 4, 389–420.
[12] Chua L.O. , The Genesis of Chua’s Circuit, AEU V.46, 250 (1992).
[13] Coll B., Gasull A.and Prohens R., Degenerate Hopf bifurcations in dis-
continuous planar systems. J. Math. Anal. Appl., V.253 no. 2, (2001)
671690.
[14] di Bernardo, M., Budd, C., Champneys, A. R., Kowalczyk, P., Nord-
mark, A. B., Olivar, G. and Piiroinen, P. T.. Bifurcations in non-smooth
dynamical systems, Publications of the Bristol Centre for Applied Nonlinear
Mathematics, N. 2005-4.
[15] Dumortier F. and Roussarie R., Canard cycles and center manifolds,
Memoirs Amer. Mat. Soc. V. 121, (1996).
18

[16] Dumortier F. , Singularities of vector fields, Publications of IMPA, Rio

de Janeiro, (1977).
[17] Ekeland, I. Discontinuités de champs hamiltoniens et existence de so-
lutions optimales en calcul des variations, Inst. Hautes Études Sci. Publ.
Math. No. 47 (1977), 5–32.
[18] Fenichel N., Geometric singular perturbation theory for ordinary dif-
ferential equations, J. Diff. Equations V. 31 (1979), 53–98.
[19] Filippov A.F., Differential equations with discontinuous righthand
sides,Kluwer Academic Publishers, Dordrecht, 1988.
[20] Flügge-Lotz I., Discontinuous automatic control, Princeton University,
(1953).
[21] Garcia, R. and Teixeira, M.A., Vector fields in manifolds with boundary
and reversibility-an expository account , Qualitative Theory of Dynamical
Systems, V.4, (2004), 311-327.
[22] Gasull A. and Torregrosa J., Center-focus problem for discontinuous
planar differential equations, Internat. J. Bifur. Chaos Appl. Sci. Engrg.,
V.13 no. 7 , (2003), 17551766.
[23] Henry P., Diff. equations with discontinuous right-hand side for plan-
ning procedures, J. of Economy Theory, V.4 (1972), pp. 545–551.
[24] Hogan S., On the dynamics of rigid-block motion under harmonic forc-
ing, Proc. Roy. Soc. London A, V. 425 (1989), pp. 441–476.
[25] Ito T., A Filippov solution of a system of diff. eq. with discontinuous
right-hand sides, Economic Letters, V. 4 (1979), pp. 349–354.
[26] Jackson J.D., Classical eletrodynamics, Wiley, 3rd Ed. (1999).
[27] Jacquemard A. and Teixeira M.A., On singularities of discontinuous
vector fields, Bulletin des Sciences Mathématiques, V. 127, (2003), 611-
633.
[28] Jacquemard A. and Teixeira M.A. Invariant varieties of discontinuous
vector fields, Nonlinearity, V.18, (2005), 21-43.
[29] Jacquemard A. and Teixeira M.A. Computer analysis of periodic orbits
of discontinuous vector fields, Journal of Symbolic Computation, V. 35,
(2003), 617-636.
[30] Jones, C. Geometric Singular Perturbation Theory, C.I.M.E. Lec-
tures, Montecatini Terme, June 1994, Lecture Notes in Mathematics 1609,
Springer-Verlag, Heidelberg, (1995).
[31] Kozlova V.S., Roughness of a Discontinuous System, Vestinik
Moskovskogo Universiteta, Matematika V. 5 (1984), 16–20.
 19

[32] Kunze M. and Kupper T., Qualitative bifurcation analysis of a non-

smooth friction oscillator model, Z. Angew. Math. Phys., V. 48, (1997),
pp. 87-101.
[33] Kuznetsov Yu.A.et al, One-parameter bifurcations in planar Filippov
systems, Int. J. Bifuraction Chaos, V. 13, (2003), pp. 2157-2188.
[34] Llibre J. and Teixeira M.A., Regularization of discontinuous vector
fields in dimension 3, Discrete and Continuous Dynamical Systems, Vol.
3, N. 2, (1997), 235-241.
[35] Luo, C.J. Singularity and dynamics on discontinuous vector fields,
Monograph Series on Nonlinear Science and Complexity, Elsevier Sc.,
(2006), i+310 pp.
[36] Manosas F. and Torres P.J., Isochronicity of a class of piecewise con-
tinuous oscillators, Proc. of AMS, V.133, N. 10, (2005), 3027–3035
[37] Machado A.L. and Sotomayor J. , Structurally stable discontinuous
vector fields in the plane, Qual. Theory Dyn. Syst., V. 3 (2002), 227–250.
[38] Medrado J. and Teixeira M.A., Symmetric singularities of reversible
vector fields in dimension three, Physica D, V. 112 (1998) , 122-131.
[39] Medrado J. and Teixeira M.A., Codimension-two singularities of re-
versible vector fields in 3D , Qualitative theory of dynamical systems J. ,
V.2, N.2 (2001), 399-428 .
[40] Percell P.B., Structural stability on manifolds with boundary, Topology,
V. 12, (1973), 123-144.
[41] Seidman, T. Aspects of modeling with discontinuities in Advances in
Applied and Computational Mathematics Proc. Dover Conf. (2006), (G.
N’Guerekata, ed.), to appear.
[42] Sotomayor J., Structural stability in manifolds with boundary, in Global
analysis and its applications, V. III, IEAA, Vienna (1974,)167-176.
[43] Sotomayor J. and Teixeira M.A., Vector fields near the boundary of a
3-manifold, Lect. Notes in Math., 331, Springer Verlag, (1988), 169-195.
[44] Sotomayor J. and Teixeira M.A., Regularization of Discontinuous
Vector Fields, International Conference on Differential Equations, Lisboa
(1996), 207–223.
[45] Sussmann, H., Subanalytic sets and feedback control, J. of Differential
equations, V.31, (1979), 31-52
[46] Teixeira M.A., Generic bifurcation of certain singularities, Bolletino
della Unione Matematica Italiana, V.16-B, (1979), 238-254.
[47] Teixeira M.A., Generic bifurcations in manifolds with boundary, J. of
Diff. Eq., V. 25 (1977) pp. 65-89.
20

[48] Teixeira M.A., Stability conditions for discontinuous vector fields, J. of

Diff Eq., V. 88, (1990), 15-24.
[49] Teixeira M.A., Generic Singularities of Discontinuous Vector Fields,
An. Ac. Bras. Cienc. V. 53, n2, (1991), 257–260.
[50] Teixeira M.A., Generic bifurcation of sliding vector fields, J. of Math.
Analysis and Appications V. 176, 1993, 436-457.
[51] Teixeira M.A.,Singularities of Reversible Vector Fields, Physica D, V.
100, (1997), pp. 101-118.
[52] Teixeira M.A., Codimension two singularities of sliding vector fields;
Bull. of Belgium Math. Soc. V. 6, N.3 (1999), 369-381.
[53] Vishik S.M., Vector fields near the boundary of a manifold, Vestnik
Moskovskogo Universiteta, Matematika, V. 27, N.1, (1972), 13-19.
[54] Zelikin M.I. and Borisov V.F., Theory of chattering control with ap-
plications to Astronautics, Robotics, Economics, and Engineering, (1994),
Birkhäuser.

Secondary Literature, further readings

[AS] Agrachev, Andrei A.; Sachkov, Yuri L. Control theory from the geo-
metric viewpoint. Encyclopaedia of Mathematical Sciences, 87. Control
Theory and Optimization, II. Springer-Verlag, Berlin, (2004). xiv+412
pp.
[Ba] Barbashin, E. A.; Introduction to the theory of stability, Translated
from the Russian by Transcripta Service, London. Edited by T. Lukes
Wolters-Noordhoff Publishing, Groningen (1970), 223pp.
[Bro] Brogliato B., ed., Impacts in Mechanical Systems, Lect. Notes in Phys.
V.551, Springer-Verlag, Berlin (2000), 160.
[CFPT] Carmona, V., Freire E., Ponce E. and Torres F.: Bifurcation of
invariant cones In Piecewise Linear Homogeneous Systems , Interna-
tional Journal of Bifurcations and Chaos, V.15, N. 8, (2005), 2469-
2484.
[Da] Davydov, A. A. Qualitative theory of control systems (English sum-
mary) Translated from the Russian manuscript by V. M. Volosov.
Translations of Mathematical Monographs, 141. American Mathemat-
ical Society, Providence, RI, 1994. viii+147 pp.
[De] Dercole F., Gragnani F., Kuznetsov Yu.A., and Rinaldi S. Numerical
sliding bifurcation analysis: An application to a relay control system,
IEEE Trans. Circuit Systems - I: Fund. Theory Appl. V. 50 (2003),
pp. 1058-1063.
 21

[Gl] Glocker Ch., Set-Valued Force Laws: Dynamics of Non-Smooth Sys-

tems, Lecture Notes in Applied Mechanics 1, Springer Verlag, Berlin,
Heidelberg (2001).
[HC] Hogan, John (ed.); Champneys, Alan (ed.); Krauskopf, Bernd (ed.);
di Bernardo, Mario (ed.); Wilson, Eddie (ed.); Osinga, Hinke (ed.);
Homer, Martin (ed.) Nonlinear dynamics and chaos: Where do we go
from here? (English) Bristol: Institute of Physics Publishing (IOP).
xi, 358 pp.
[HCM] Huertas J.L, Chen W.K. and Madan R.N (editors) Visions of non-
linear science in the 21st century. Part I, Festschrift dedicated to Leon
O. Chua on the occasion of his 60th birthday. Papers from the work-
shop held in Sevilla, June 26, 1996. , Internat. J. Bifur. Chaos Appl.
Sci. Engrg. 7 (1997), no. 9. World Scientific Publishing Co. Pte. Ltd.,
Singapore, 1997. pp. i–iv and 1907–2173.
[Kom] Komuro M. Periodic bifurcation of continuous piece-wise vector fields,
Advanced Series in Dynamical systems, V. 9 (ed. K. Shiraiwa), (1990),
World Sc., Singapore.
[KK] Kunze, M. and Küpper, T. Non-smooth dynamical systems: an
overview, (Fiedler, B. (ed.)), Ergodic theory, analysis, and efficient
simulation of dynamical systems. Berlin: Springer. 431-452 (2001).
[Kun] Kunze M. Non-smooth dynamical systems, Lecture Notes in Mathe-
matics, V. 1744, Springer-Verlag, Berlin, (2000). x+228 pp.
[LST] Llibre J., Silva P.R. and Teixeira M.A., Regularization of discontinu-
ous vector fields on R3 via singular perturbation ,J. of Dynamics and
Differential Eq., V. 19, Number 2, (2007), 309-331.
[Ma] Martinet, J., Singularités des fonctions et applications différentiables,
(French) Pontifı́cia Universidade Catolica do Rio de Janeiro, Rio de
Janeiro, 1974. xiii+181 pp.
[Mi1] Minorsky, N. Nonlinear oscillations, D. Van Nostrand Co., Inc.,
Princeton, N.J.-Toronto-London-New York, (1962) xviii+714 pp.
[Mi2] Minorsky, N. Theory of nonlinear control systems, McGraw-Hill Book
Co., New York-London-Sydney, (1969) xx+331 pp.
[Mi3] Minorsky, N. Théorie des oscillations, (French) Mémorial des Sciences
Mathématiques, Fasc. 163 Gauthier-Villars Éditeur, Paris, (1967)
i+114 pp.
[OB] Ostrowski, J.P.; Burdick, J.W., Controllability for mechanical sys-
tems with symmetries and constraints, Recent developments in robot-
ics. Appl. Math. Comput. Sci. V.7 (1997), no. 2, 305–331.
 22

[PP] Peixoto M.C. and Peixoto M.M.,, Structural Stability in the plane with
enlarged conditions, Anais da Acad. Bras. Ciências V.31, (1959), 135–
160.
[RL] Rega G. and Lenci S., Nonsmooth dynamics, bifurcation and control in
an impact system, Systems Analysis Modelling Simulation archive V.
43 , Issue 3, Gordon and Breach Science Publishers, Inc. Newark, NJ,
USA, ( 2003).
[Se3] Seidman, T., Some limit problems for relays, in Proc.First World Con-
gress of Nonlinear Analysis (Lakshmikantham V., ed.), V.I, , Walter
de Gruyter, Berlin (1995), 787-796.
[S1] Sotomayor J., Generic one-parameter families of vector fields on 2-
manifolds, Publ. Math. IHES, V. 43, (1974), pp. 5-46.
[Sz] Szmolyan P., Transversal Heteroclinic and Homoclinic Orbits in Sin-
gular Perturbation Problems, J. Diff. Equations V. 92 (1991), 252–281.
[T2] Teixeira M.A., Topological Invariant for Discontinuous Vector Fields,
Nonlinear Analysis: TMA, V.9 (10)(1985), 1073 - 1080.
[Ut] Utkin V., Sliding Modes and their Application in Variable Structure
Systems, Mir, Moscow (1978); Sliding Modes in Control and Opti-
mization, Springer, Berlin, (1992).

View publication stats