Professional Documents
Culture Documents
MOUNIER 1
Abstract An algebraic theory for linear partial differential the boundary control of the wave equation. The final sections
systems controlled at the boundary is presented. Some structural are devoted to the heat and the Euler-Bernoulli equations.
properties, analogous to the differential flatness of nonlinear Acknowledgments 1: Several parts of the work presented
systems, are exposed yielding an easy solution to tracking
problems. Examples of the wave and Euler Bernoulli equations here were done in collaboration with our colleagues and
are treated in some detail. friends P. Rouchon (Ecole des Mines de Paris, France) and
J. Rudolph (Technische Universitat Dresden, Germany).
Index Terms Distributed parameter systems, Delay systems,
-freeness, Flatness, controllability, Trajectory tracking, Wave
equation, Euler-Bernoulli equation, Module theory. I. A BSTRACT LINEAR SYSTEM THEORY
A. Basic definitions
I NTRODUCTION
Any ring R is commutative, with 1 and without zero
T HE present document reports on recent works in the area
of infinite dimensional systems modelled through delay
differential or partial differential equations [1], [21], [23],
divisors.
Notation The submodule spanned by a subset S of an R-
[24], [25], [26], [37], [41], [42], [48], [49], [50], [52]. The module M is written [P ].
adopted framework emphasizes on equation structure (rather An R-system , or a system over R, is an R-module. Two
than solutions) in order to study a given system. When dealing R-systems 1 and 2 are said to be R-equivalent, or equiva-
with linear equations, a system is associated with a module lent over R, if the R-modules 1 and 2 are isomorphic.
over a differential ring, this notion playing for differential An R-dynamics, or a dynamics over R, is an R-system
equations the role played by vector spaces for linear algebraic equipped with an input, i. e., a subset u of which may be
equations. empty, such that the quotient R-module /[u] is torsion. The
Keeping constructive notions in mind, a natural general- input u is independent if the R-module [u] is free, with basis
ization of Kalmans controllability is -freeness [12], [13], u.
[21], [48], which allows the tracking of a reference trajectory An output y is a subset, which may be empty, of .
in a way bearing some analogy with flat finite dimensional An input-output R-system, or an input-output system over
nonlinear systems (see [17], [18] and the references therein). R, is an R-dynamics equipped with an output.
This property leads to a complete parametrization of the Remarks 1: 1) Kalmans module-theoretic setting [32] is
system in terms of the so-called basic output. The obtained related to the state variable description, whereas our
formulae yield every variable of the system without any module description encompasses all system variables
integration of a differential equation. without any distinction.
These ingredients are complemented by the use of Mikusin- 2) In informal terms, a system with equation
skis operational calculus [45], [46] which avoids many of
the severe analytical difficulties encountered with the classic a(s) x = b(s) u
Laplace transform (see, e.g., [8]). Another important tool is a where s stands for dt d
, the derivation operator, will be
special type of C -functions, called Gevrey functions [27], represented by a linear structure consisting in all the
yielding exact solutions to rest to rest problems for distributed linear combinations of x and u and their derivatives such
parameter systems. that the above relation is satisfied
The paper is organized as follows. We first introduce ab-
stract linear systems over arbitrary commutative rings. Finite = {p(s) x + q(s) u| p, q R[s], a(s) x = b(s) u}
dimensional linear systems are then presented within this
framework. We proceed to delay systems and relate them to 3) In the same informal terms, an input u = (u1 , . . . , um )
is such that every variable z of the system satisfies a
M. Fliess is with the Centre de Mathematiques et de Leurs Applications, relation of the form:
Ecole Normale Superieure de Cachan, 61 avenue du President Wilson, 94235
Cachan, France, AND with the Laboratoire GAGE, Ecole polytechnique, m
X
91128 Palaiseau Cedex, France. E-mail: fliesscmla.ens-cachan.fr p(s) z = qi (s) ui
H. Mounier is with the Robotics Center (CAOR) of the Ecole Nationale i=1
Superieure des Mines de Paris, 6062, Boulevard Saint-Michel, 75272 Paris
Cedex 06, France, E-mail: Hugues.Mounier@caor.ensmp.fr where p, qi R[s], p 6= 0.
This work was partially supported by the European Commissions Training
and Mobility of Researchers (TMR) Contract # ERBFMRX-CT970137, by the Let A be an R-algebra et and be an R-system. The A-
G.D.R. Medicis and by the G.D.R.-P.R.C. Automatique. module A R is an A-system, which extends .
2 M. FLIESS AND H. MOUNIER
Proposition 2: An R-system with finite and independent II. F INITE DIMENSIONAL LINEAR SYSTEMS
input and finite output is left invertible (resp. right invertible) A. Modules over principal ideal rings
if, and only if, its transfer matrix is left invertible (resp. right
invertible). In this section R is theP principal ideal ring k[s], whose
The above system is said to be square if m = p. elements are of the form f inite a s , a k, where k
d
Proposition 3: Consider a square R-system with an inde- is a field, and s stands for dt . All k[s]-modules are finitely
pendent input. Then, the two following properties are equiva- generated and, therefore, finitely presented. The following two
lent: theorems are classic (see, e.g., [36])
the system is left invertible;
Theorem 2: For a k[s]-module M , the next two conditions
the system is right invertible.
are equivalent:
An R-system with an independent input, which is both left 1) M is torsion,
and right invertible, is necessarily square. 2) the dimension of M as a k-vector space is finite, i. e.,
Such a left and right invertible system is called invertible. dimk M < .
Theorem 3: Any k[s]-module M may be written
Examples 1: 1) The system y + y = u is k[s, ehs ]-free sentation (see [12], [19], [20])
controllable, with basis y. fut fut
x1 x1 u1
2) The system y = (1 + )u is k[s, ehs ]-torsion free .. .. .
s . = A . + B .. (7)
controllable, but not k[s, ehs ]-free controllable.
3) The system y = u is free (e.g., k[s, ehs , ehs ]-free xfut
n xfut um
controllable), but not k[s, ehs ]-free controllable. L sn
y1fut
fut
e y1
1
x1 u1
2) Spectral controllability: The following definition of .. =
..
= C
..
+ D
..
spectral controllability extends previous ones (see, e.g., [4],
. . . .
[57]) in our context. ypfut eL p s yp xfut
n um
(8)
Definition 1: Let be a delay system, with presentation
matrix P of full generic rank . It is called spectrally where
controllable if
A k nn , B k nm , C k pn , D k[s]pm ;
s C, rkC P (s, ehs ) =
n = dimk spec /spank[s] (u)
Set Sr = k(s)[ehs , ehs ] E, where E denotes the ring The definition of quasi-finite input-output systems leads at
of entire functions. We have the following interpretation of once to the following property
spectral controllability [49] (see also [28] and [5]): Proposition 11: The matrix T k(s, eLs )pm is the
Proposition 9: Let be a torsion free controllable delay transfer matrix of a nominal (i.e. perturbation-free) quasi-finite
system. It is spectrally controllable if, and only if, it is Sr input-output system if, and only if, it is of the form
torsion free controllable.
0
The following result [21] gives implication relationships be- eL1 s
tween the notion of k[s, ehs , ehs ]-freeness, k[s, ehs , ehs ]-
eL2 s
torsion freeness and spectral controllability. T = .. R
Proposition 10: Let be a delay system. The following
.
chain of implications is true 0 eLp s
free = spectrally controllable = torsion free.
Proof: The proof follows directly from the inclusion where R k(s)pm . In particular, T k(s, eLs ) is the
nominal transfer function of a quasi-finite SISO system if,
k[s, ehs , ehs ] Sr . and only if, T = ReLs , R k(s), L 0.
Such a transfer matrix, or function, will also be called quasi-
finite.
Example 2: The SISO unperturbed system sx = (1
3) Quasi-finite systems: These systems, quite frequent in
eLs )u, y = x, with a 1-dimensional state, is not quasi-finite,
practice, are typically ones with a transfer matrix of the form Ls
H(s)ehs with H(s) a transfer matrix without delays. since its transfer function 1es is not quasi-finite.
A linear delay system is said to be special if =
k[s, ehs , ehs ] k[s] spec , where spec is a finitely generated B. The wave equation
k[s]-module. It is equivalent saying there exists a presentation
matrix with entries in k[s]. Note that spec is not unique: it 1) The torsional behavior of a flexible rod: Consider [50]
may be replaced by ehs spec , spec spec , h spanN (h). the torsional behavior of a flexible rod with a torque applied
The k[s]-linear mapping spec , spec 7 = 1 spec to one end. A mass is attached to the other end. The system
is injective. We will therefore consider, with a slight abuse of is described by the one dimensional wave equation.
notations, spec as a subset of . A dynamics is said to be
quasi-finite if it is special and there exists spec such that the
2q 2q
control variables u belong to it. This spec is of course unique. 2 (, z) = (, z) (9)
2 z 2
It is called the fundamental dynamics, or module, and will be
q q 2q
designated by spec0 . (, 0) = u( ), (, L) = J 2 (, L) (10)
z z
A special input-output system is said to be quasi-finite if q
the next two conditions are satisfied: q (0, z) = q0 (z), (0, z) = q1 (z) (11)
1) its corresponding dynamics is quasi-finite,
Here q(, z) denotes the angular displacement from the unex-
2) there exists p non-negative real numbers L =
cited position at a point z [0, L] at time > 0, as shown
(L1 , . . . , Lp ) such that
in Figure 1; L is the length of the rod, the inverse of the
spec wave propagation speed, J the inertial momentum of the mass,
y1fut = eL1 s y1 , . . . , ypfut = eLp s yp 0 u( ) the control torque and q0 , q1 describe the initial angular
displacement and velocity, respectively.
The definition of quasi-finite input-output systems leads to
the following nominal (i.e. unperturbed) state-variable repre-
6 M. FLIESS AND H. MOUNIER
h
w(x, s) = (cosh( s) + cosh( s))(i sinh(x s)+ [14] M. F LIESS, Une interpretation algebrique de la transformation
de Laplace et des matrices de transfert, Linear Algebra Appl.,
sinh(x s)) + (i sinh( s) sinh( s)) 203-204, 1994, pp. 429-442.
i2 s [15] M. F LIESS and H. B OURL ES, Discussing some examples of
( cosh(x s) + cosh(x s) y linear system interconnections, System Control Lett., 27, 1996,
pp. 1-7.
where [16] M. F LIESS and R. H OTZEL , Sur les systemes lineaires a
derivation non entiere, C.R. Acad. Sci. Paris, IIb-324, 1997,
(c + d) i sinh( s) sinh( s) pp. 99-105.
y = 1 and =
2 s s [17] M. F LIESS, J. L EVINE , P. M ARTIN and P. ROUCHON, Flat-
ness and defect of non-linear systems: introductory theory and
Set u = y and w = y. The next result [24] is obtained in applications, Internat. J. Control, 61, 1995, pp. 1327-1361.
the same manner as in Subsection IV-C. [18] M. F LIESS, J. L EVINE , P. M ARTIN and P. ROUCHON, A Lie-
Theorem 9: The C[, ]module spanned by u and Backlund approach to equivalence and flatness of nonlinear
systems, IEEE Trans. Automat. Contr., 44, 1999, pp. 928937.
w is torsion free, of rank 1, but not free. The localized
[19] F LIESS , M. and M ARQUEZ , R., Continuous-time linear pre-
C[, , ()1 ]module , is free, with basis y. dictive control and flatness: a module-theoretic setting with
examples, International Journal of Control, 73, 2000, pp. 606
623.
C ONCLUSION
[20] F LIESS , M., M ARQUEZ , R., D ELALEAU , E., and S IRA -
The comparison with other approaches [3], [7], [11], [34], R AM IREZ H., Correcteurs proportionnels-integraux generalises,
[38], [39] on the control of partial differential equations will ESAIM COCV, 7, 2002, pp. 2342.
URL: http://www.edpsciences.org/cocv.
be made in forthcoming papers. [21] M. F LIESS and H. M OUNIER, Controllability and observability
Related methods have been recently employed for linear of linear delay systems: an algebraic approach, ESAIM COCV,
systems with fractional derivatives [16], [31]. Other works in 3, 1998, pp. 301314.
progress [51], [53] are also exploring the same circle of ideas URL: http://www.edpsciences.org/cocv.
for some types of infinite dimensional nonlinear systems. [22] M. F LIESS and H. M OUNIER, Quasi-finite linear delay systems:
theory and applications, Proc. IFAC Workshop Linear Time
Delay Systems, Grenoble, 1998, pp. 211215.
R EFERENCES [23] M. F LIESS and H. M OUNIER, Tracking control and -freeness
of inifinite dimensional linear systems, in Dynamical Sys-
[1] Y. AOUSTIN, M. F LIESS , H. M OUNIER , P. ROUCHON and J. tems, Control, Coding and Computer Vision, G. P ICCI and D.
RUDOLPH, Theory and practice in the motion planning control G ILLIAM Eds., Birkhauser, 1999, pp. 4568.
of a flexible robot arm using Mikusinski operators, Proc. 4th [24] M. F LIESS , H. M OUNIER , P. ROUCHON and J. RUDOLPH,
Symp. Robotics Control, Nantes, 1997, pp. 287293. Systemes lineaires sur les operateurs de Mikusinski et com-
[2] Z. A RSTEIN, Linear systems with delayed control: a reduction, mande dune poutre flexible, ESAIM Proc., 2, 1997, pp. 183
IEEE Trans. Automat. Contr., 27, 1982, pp. 869879. 193.
[3] A. B ENSOUSSAN, G. DA P RATO, M.C. D ELFOUR and S.K. URL: http://www.edpsciences.com/cocv.html.
M ITTER, Representation and Control of Infinite Dimensional [25] M. F LIESS , H. M OUNIER , P. ROUCHON and J. RUDOLPH,
Systems, vol. 1 & 2, Birkhauser, Boston, 1992 & 1993. Controlling the transient of a chemical reactor: a distributed
[4] K. B HAT and H. KOIVO, Modal characterizations of control- parameter approach, Proc. CESA 98, Hammamet, Tunisia, 1998.
lability and observability for time-delay systems, IEEE Trans. [26] M. F LIESS , H. M OUNIER , P. ROUCHON and J. RUDOLPH,
Automat. Contr., 21, 1976, pp. 292293. A distributed parameter approach to the control of a tubular
[5] B RETH E , D. and L OISEAU , J.J., Stabilization of time-delay reactor: a multi-variable case, Proc. CDC, 1998.
systems, Journal europeen des systemes automatises, 31, 1997, [27] M. G EVREY, La nature analytique des solutions des equations
pp. 10251042. aux derivees partielles, Ann. Sci. Ec. Norm. Sup., 25, 1918,
[6] C.I. B YRNES , M. S PONG and T.J. TARN, A several complex pp. 125190.
variables approach to feedback stabilization of linear neutral [28] H. G LUESING -L UERSSEN, Linear Delay-Differential Systems
delay-differential systems, Math. Systems Theory, 17, 1984, with Commensurate Delays: An Algebraic Approach, Springer,
pp. 97133. Berlin, 2002.
[7] R.F. C URTAIN and H.J. Z WART , An Introduction to Infinite [29] I.M. G UELFAND and G.E. C HILOV, Les distributions, vols. 2
Dimensional Linear Systems Theory, Springer, New York, 1995. & 3, Dunod, Paris, 1964 & 1965.
[8] G. D OETSCH, Theorie und Anwendung der Laplace- [30] L. H ORMANDER, The Analysis of Partial Differential Opera-
Transformation, Springer, Berlin, 1937. tors, vol. 1, Springer, Berlin, 1983.
[9] S. D RAKUNOV and U. O ZG UNER, Generalized sliding modes [31] R. H OTZEL , Contributions a la theorie structurelle et la com-
for manifold control of distributed parameter systems, in mande des systemes lineaires fractionnaires, These, Universite
Variable Structure and Lyapounov Control, A. S. Zinober, Ed., Paris-Sud, Orsay, 1998.
Lect. Notes Control Inform. Sci. 193, pp. 109129. Springer, [32] R.E. K ALMAN , L. FALB and M.A. A RBIB, Topics in Mathe-
London, 1994. matical System Theory, McGraw-Hill, New York, 1969.
[10] F. D UBOIS, N. P ETIT and P. ROUCHON, Motion planning and [33] H. KOMATSU, Microlocal Analysis in Gevrey classes and
nonlinear simulations for a tank containing a fluid, Proc.5th in complex domains, in Microlocal Analysis and Applications,
European Control Conference, Karlsruhe, Germany, 1999. L. Cattabriga and L. Rodino, Eds., Lect. Notes Math. 1495,
[11] A. E L JAI and A.J. P RITCHARD, Capteurs et actionneurs dans pp. 161236. Springer, Berlin, 1991.
lanalyse des systemes distribues, Masson, Paris, 1986. [34] V. KOMORNIK, Exact Controllability and Stabilization, Wiley,
[12] M. F LIESS, Some basic structural properties of generalized Chichester & Masson, Paris, 1994.
linear systems, Systems Control. Lett., 15, 1990, pp. 391-396. [35] T.Y. L AM, Serres Conjecture. Springer, Berlin, 1978.
[13] M. F LIESS, A remark on Willems trajectory characterization of [36] S. L ANG, Algebra, 3rd ed., Addison-Wesley, Reading, MA,
linear controllability, Systems Control Lett., 19, 1992, pp. 43-45. 1993.
ALGEBRAIC FRAMEWORK FOR INFINITE DIMENSIONAL SYSTEMS 9
[37] B. L AROCHE , P. M ARTIN and P. ROUCHON, Motion planning [64] A.A. S USLIN, Projectives modules over a polynomial ring
for the heat equation, Int. J. Robust and Nonlinear Control, 10, are free (in russian), Dokl. Akad. Nauk. SSSR, 229, 1976,
2000, pp. 629643. pp. 1063-1066 (english translation: Soviet. Math. Dokl., 17,
[38] J.L. L IONS, Controle optimal des systemes gouvernes par des 1976, pp. 1160-1164).
equations aux derivees partielles, Dunod, Paris, 1968. [65] J.C. W ILLEMS, Paradigms and puzzles in the theory of dynam-
[39] J.L. L IONS, Controlabilite exacte, perturbations et stabilisa- ical systems, IEEE Trans. Automat. Contr., 36, 1991, pp. 259-
tion des systemes distribues, vol. 1 & 2, Masson, Paris, 1988. 294.
[40] J.L. L IONS and E. M AGENES, Problemes aux limites non [66] K. YOSIDA, Operational Calculus, Springer, New York, 1984.
homogenes et applications, vol. 3, Dunod, Paris, 1970. [67] D.C. YOULA and G. G NAVI, Notes on n-dimensional system
[41] A.F. LYNCH and J. RUDOLPH, Flachheitsbasierte Rands- theory, IEEE Trans. Circuits Syst., 26, 1979, pp. 105111.
teuerung parabolischer Systeme mit verteilen Parametern, Au- [68] D.C. YOULA and P.F. P ICKEL , The Quillen-Suslin theorem and
tomatisierungstechnik, 48, 2000, pp. 478486. the structure of n-dimensional elementary polynomial matrices,
[42] A.F. LYNCH and J. RUDOLPH, Flatness based boundary control IEEE Trans. Circuits Syst., 31, 1984, pp. 513518.
of a nonlinear parabolic equation modelling a tubular reactor, [69] S. Z AMPIERI, Modellizzazione di Sequenze di Dati Mutlidi-
Nonlinear Control in the Year 2000, A. I SIDORI, F. L AMNABHI - mensionali, Tesi, Universita di Padova, 1993.
L AGARRIGUE and W. R ESPONDEK Eds., pp. 4554, Springer-
Verlag, 2000.
[43] J. M ARSHALL , H. G ORECKI, A. KORYTOWSKI and K. WAL -
TON Time delay systems stability and performance criteria with
applications, Ellis Horwood, New York, 1992.
[44] P. M ARTIN, R.M. M URRAY and P. ROUCHON, Flat Systems,
in Plenary Lectures and Mini-Courses, ECC 97, G. Bastin and
M. Gevers, Eds., Brussels, 1997, pp. 211-264.
[45] J. M IKUSI NSKI, Operational Calculus, vol. 1, Pergamon, Ox-
ford & PWN, Warsaw, 1983.
[46] J. M IKUSI NSKI and T.K. B OEHME , Operational Calculus,
vol. 2, Pergamon, Oxford & PWN, Warsaw, 1987.
[47] A.S. M ORSE , Ring models for delay-differential systems, Au-
tomatica, 12, 1976, pp. 529531.
[48] H. M OUNIER, Proprietes structurelles des systemes lineaires a
retards : aspects theoriques et pratiques, These, Universite Paris-
Sud, Orsay, 1995.
[49] H. M OUNIER, Algebraic interpretations of the spectral con-
trollability of a linear delay system, Forum Mathematicum, 10,
1998, pp. 3958.
[50] H. M OUNIER , P. ROUCHON and J. RUDOLPH, Some examples
of linear systems with delays, J. Europ. Syst. Autom., 31, 1997,
pp. 911925.
[51] H. M OUNIER and J. RUDOLPH, Flatness based control of non-
linear delay systems: Example of a class of chemical reactors,
Internat. J. Contr., 71, 1998, pp. 838-871.
[52] H. M OUNIER , J. RUDOLPH , M. F LIESS and P. ROUCHON,
Tracking control of a vibrating string with an interior mass
viewed as delay system, ESAIM COCV, 3, 1998, pp. 315321.
URL: http://www.epdsciences.org/cocv
[53] N. P ETIT, Y. C REFF and P. ROUCHON, Motion planning for
two classes of nonlinear systems with delays depending on the
control, Proc. CDC, 1998.
[54] D. Q UILLEN, Projective modules over polynomial rings, Inven-
tiones Math., 36, 1976, pp. 167-171.
[55] J.P. R AMIS, Series divergentes et theories asymptotiques, Soc.
Math. France, Marseille, 1993.
[56] B. R AO, Uniform stabilization of a hybrid system of elasticity,
SIAM J. Contr. Optim., 33, 1995, pp. 440454.
[57] P. ROCHA and J.C. W ILLEMS, Behavioral controllability of
D-D systems. SIAM J. Contr. Opt., 35, 1987, pp. 254264.
[58] L. RODINO, Linear Partial Differential Operators in Gevrey
Spaces, World Scientific, Singapore, 1993.
[59] J. ROTMAN, An Introduction to Homological Algebra, Aca-
demic Press, Orlando, 1979.
[60] L.H. ROWEN, Ring Theory, Academic Press, Boston, 1991.
[61] J.P. S ERRE , Faisceaux algebriques coherents. Annals. of
Math., 61, 1955, pp. 197278.
[62] E. D. S ONTAG, Linear systems over commutative rings: a
survey, Richerche di Automatica, 7, 1976, pp. 134.
[63] M.W. S PONG and T.J. TARN, On the spectral controllability of
delay-differential equations, IEEE Trans. Automat. Contr., 26,
1981, pp. 527528.