External Stabilization 119
G2 for each i ∈ {1, . . . , m} only one of the following mutually exclusive conditions holds:
– for all p ∈ ∂V t, x p · 1, g
i
t, x 0,
– for all p ∈ ∂V t, x p · 1, g
i
t, x 0,
– for all p
∈ ∂V t, x p · 1, g
i
t, x = 0;
G3 there exists ı ∈ {1, . . . , m} such that for each i ∈ {1, . . . , m}\{ı} only one of the following
mutually exclusive conditions holds:
– for all p
∈ ∂V t, x p · 1, g
i
t, x 0,
– for all p
∈ ∂V t, x p · 1, g
i
t, x 0,
– for all p
∈ ∂V t, x p · 1, g
i
t, x = 0;
Let us remark that f3 ⇒ f2 ⇒ f1 and G1 ⇒ G2 ⇒ G3.
T
HEOREM
2. Let V :
n+1
→ be such that there exists L 0 such that V0 and V1
hold. If for all x
∈
n
with kxk L one of the following couples of conditions holds for a.e.
t ≥ 0:
i f1 and G1, ii f2 and G2,
iii f3 and G3, then system 1 is UBIBS stabilizable.
Let us make some remarks. If for all x
∈
n
with kxk L assumption f1 or f2 or f3 holds for a.e. t ≥ 0, then, by
Theorem 1 in Section 2, system 2 is uniformly Lagrange stable. Actually in [4] the authors introduce the concept of robust uniform Lagrange stability and prove that it is equivalent to the
existence of a locally Lipschitz continuous Lyapunov-like function. Then assumption f1 or f2 or f3 implies more than uniform Lagrange stability of system 2. In [9], the author has
also proved that, under mild additional assumptions on f , robust Lagrange stability implies the existence of a C
∞
Lyapunov-like function, but the proof of this result is not actually constructive. Then we could still have to deal with nonsmooth Lyapunov-like functions even if we know that
there exist smooth ones. Moreover Theorem 2 can be restated for autonomous systems with the function V not de-
pending on time. In this case the feedback law is autonomous and it is possible to deal with a situation in which the results in [9] don’t help.
Finally let us remark that if f is locally Lipschitz continuous, then, by [14] page 105, the Lagrange stability of system 2 implies the existence of a time-dependent Lyapunov-like
function of class C
∞
. In this case, in order to get UBIBS stabilizability of system 1, the regularity assumption on G can be weakened to G
∈ L
∞ loc
n+1
;
m
as in [2].
4. Proof of Theorem 2
We first state and prove a lemma. L
EMMA
2. Let V :
n+1
→ be such that there exists L 0 such that V0 and V1
hold. If t , x , with kxk L, is such that, for all p ∈ ∂V t, x p · 1, g
i
t, x 0, then there exists δ
x
0 such that, for all x ∈ Bx, δ
x
, for all p ∈ ∂V t, x, p · 1, g
i
t, x 0.
120 F. Ceragioli
Analagously if t , x , with kxk L, is such that for all p ∈ ∂V t, x, p · 1, g
i
t, x 0, then there exists δ
x
0 such that, for all x ∈ Bx, δ
x
, for all p ∈ ∂V t, x, p·1, g
i
t, x 0. Proof. Let γ 0 be such that
kxk L + γ , and let L
x
0 be the Lipschitz constant of V in the set
{t} × Bx, γ . For all t, x ∈ {t} × Bx, γ and for all p ∈ ∂V t, x kpk ≤ L
x
see [5], page 27.
Since g
i
is continuous there exist η and M such that k1, g
i
t, x k ≤ M in {t} × Bx, η.
Let d = min{ p · 1, g
i
t, x , p ∈ ∂V t, x}. By assumption d 0.
Let us consider ǫ
d 2L
x
+M
. By the continuity of g
i
, there exists δ
i
such that, if kx − xk δ
i
, then k1, g
i
t, x −
1, g
i
t, x k ǫ.
By the upper semi-continuity of ∂ V see [5], page 29, there exists δ
V
0 such that, if kx − xk δ
V
, then ∂ V t , x ⊆ ∂V t, x + ǫ B0, 1, i.e. for all p ∈ ∂V t, x there exists
p ∈ ∂V t, x such that kp − pk ǫ.
Let δ
x
= min{γ, η, δ
i
, δ
V
}, x be such that kx − xk δ
x
and p ∈ ∂V t, x, p ∈ ∂V t, x
be such that kp − pk ǫ.
It is easy to see that |p · 1, g
i
t, x − p · 1, g
i
t, x |
d 2
, hence p · 1, g
i
t, x p
· 1, g
i
t , x −
d 2
=
d 2
0. The second part of the lemma can be proved in a perfectly analogous way.
Proof of Theorem 2. For each x ∈
n
, let N
x
be the zero-measure subset of
+
in which no one of the couples of conditions i , ii and iii holds. Let k :
n+1
→
m
, kt, x
= k
1
t, x, . . . , k
m
t, x, be defined by k
i
t, x =
−kxk if ∀ p ∈ ∂V t, x p · 1, g
i
t, x 0 if
∀ p ∈ ∂V t, x p · g
i
t, x = 0 ,
or f3 and G3 hold and i = ı, or t ∈ N
x
kxk if
∀ p ∈ ∂V t, x p · 1, g
i
t, x 0 . It is clear that k
∈ L
∞ loc
n+1
,
m
. By Theorem 1 it is sufficient to prove that for all R 0 there exists ρ L , R such that for
all x ∈
n
and v ∈
m
the following holds: kxk ρ, kvk R ⇒ max ˙V
3
t, x ≤ 0 for all t ∈
+
\N
x
where ˙ V
3
t, x = {a ∈
: ∃w ∈ K f t, x + Gt, xkt, x + Gt, xv such that ∀ p ∈
∂ V t, x p
· 1, w = a}. Let x be fixed and t
∈
+
\N
x
. Let a ∈ ˙V
3
t, x, w ∈ K f t, x + Gt, xkt, x +
Gt, xv be such that for all p ∈ ∂V t, x p · w = a.
By Theorem 1 in [8] we have that K f t, x
+ Gt, xkt, x + vx ⊆ K f t, x + P
m i=1
g
i
t, xK k
i
t, x + v
i
, then there exists z
∈ K f t, x, z
i
∈ K k
i
t, x + v
i
, i ∈ {1, . . . , m}, such that w = z +
P
m i=1
g
i
t, xz
i
. Let us show that a
≤ 0. We distinguish the three cases i, ii, iii. i b
= p · 1, z = a − P
m i=1
c
i t,x
z
i
does not depend on p, then b ∈ ˙V
2
t, x and, by f1, b
≤ 0.
External Stabilization 121
Let us now show that for each i ∈ {1, . . . , m} c
i t,x
z
i
≤ 0. If i is such that c
i t,x
= 0, obviously c
i t,x
z
i
≤ 0. If i is such that c
i t,x
0 then, by Lemma 1, there exists δ
x
such that k
i
t, y = −kyk in {t} × Bx, δ
x
, then k
i
is continuous at x with respect to y. This implies that K k
i
t, x + v
i
= −kxk + v
i
, i.e. z
i
= −kxk + v
i
and c
i t,x
z
i
≤ 0, provided that
kvk ρ ≥ max{L, R}. The case in which i is such that c
i t,x
0 can be treated analogously. We finally get that a
= b + P
m i=1
c
i t,x
z
i
≤ 0. ii By f2 there exists p
∈ ∂V t, x such that p · 1, z ≤ 0. a = p · 1, z + P
m i=1
p ·
1, g
i
t, xz
i
. The fact that for each i ∈ {1, . . . , m} we have p · 1, g
i
t, xz
i
≤ 0 can be proved as in i we have proved that for each i
∈ {1, . . . , m} c
i t,x
z
i
≤ 0. We finally get that a
≤ 0. iii Let us remark that if G2 is not verified, i.e. we are not in the case ii , there exists
p ∈ ∂V t, x corresponding to ı such that p · 1, g
ı
t, x = 0. Indeed, because of the
convexity of ∂ V t, x, for all v ∈
n
, if there exist p
1
, p
2
∈ ∂V t, x such that p
1
·v 0 and p
2
· v 0, then there also exists p
3
∈ ∂V t, x such that p
3
· v = 0. Let p
∈ ∂V t, x be such that p·1, g
ı
t, x = 0. For all p ∈ ∂V t, x a = p ·1, w. In
particular we have a = p·1, w = p·1, z+
P
i6=ı
p ·1, g
i
t, xz
i
+ p·1, g
ı
t, xz
ı
. By f3, p
· 1, z ≤ 0. If i 6= ı the proof that p · 1, g
i
t, xz
i
≤ 0 is the same as in ii. If i
= ı, because of the choice of p, p · 1, g
ı
t, x = 0. Also in this case we can then
conclude that a ≤ 0.
References
[1] B
ACCIOTTI
A., External Stabilizability of Nonlinear Systems with some applications, In-
ternational Journal of Robust and Nonlinear Control 8 1998, 1–10.
[2] B
ACCIOTTI
A., B
ECCARI
G., External Stabilizabilty by Discontinuous Feedback, Pro- ceedings of the second Portuguese Conference on Automatic Control 1996.
[3] B
ACCIOTTI
A., C
ERAGIOLI
F., Stability and Stabilization of Discontinuous Systems and Nonsmooth Lyapunov Functions, Preprint Dipartimento di Matematica del Politecnico di
Torino. [4] B
ACCIOTTI
A., R
OSIER
L., Liapunov and Lagrange Stability: Inverse Theorems for Dis-
continuous Systems, Mathematics of Control, Signals and Systems 11 1998, 101–128.
[5] C
LARKE
F. H., Optimization and Nonsmooth Analysis, Wiley and Sons 1983. [6] F
ILIPPOV
A. F., Differential Equations with Discontinuous Right Handsides, Kluwer Aca- demic Publishers 1988.
[7] M
C
S
HANE
E. J., Integration, Princeton University Press 1947. [8] P
ADEN
B., S
ASTRY
S., A Calculus for Computing Filippov’s Differential Inclusion with Application to the Variable Structure Control of Robot Manipulators, IEEE Transaction on
Circuits and Systems, Vol. Cas-34, No. 1, January 1997, 73–81. [9] R
OSIER
L., Smooth Lyapunov Functions for Lagrange Stable Systems, Preprint. [10] S
HEVITZ
D., P
ADEN
B., Lyapunov Stability Theory of Nonsmooth Systems, IEEE Trans-
action on Automatic Control 39 9, September 1994, 1910–1914.
122 F. Ceragioli
[11] S
ONTAG
E. D., A Lyapunov-like Characterization of Asymptotic Controllability, SIAM J.
Control and Opt. 21 1983, 462–471.
[12] S
ONTAG
E. D., S
USSMANN
H., Nonsmooth Control Lyapunov Functions, Proc. IEEE Conf. Decision and Control, New Orleans, Dec. 1995, IEEE Publications 1995, 2799–
2805. [13] V
ARAIYA
P. P., L
IU
R., Bounded-input Bounded-output Stability of Nonlinear Time-
varying Differential Systems, SIAM Journal Control 4 1966, 698–704.
[14] Y
OSHIZAWA
T., Stability Theory by Liapunov’s Second Method, The Mathematical Soci- ety of Japan 1966.
AMS Subject Classification: ???.
Francesca CERAGIOLI Dipartimento di Matematica “U. Dini”
Universit`a di Firenze viale Morgagni 67A
50134 Firenze, ITALY e-mail:
✄ ✂
✆ ✁
✞ ✟
✠ ✆
✝ ✆
✡ ☛
✄ ☞
✌ ✡
✟ ✝
✆ ✍
✆ ✡
✆ ☞
Rend. Sem. Mat. Univ. Pol. Torino Vol. 56, 4 1998
L. Pandolfi