A TUTORIAL INTRODUCTION TO NONLINEAR DYNAMICS AND CHAOS,

PART 11: MODELlNG AND CONTROL

Luis Antonio Aguirre

Centro de Pesquisa e Desenvolvimento em Engenharia Elétrica
Universidade Federal de Minas Gerais
Av. Antônio Carlos 6627
:31270-901 Belo Horizonte, M.G., Brazil
Fax: El.5 31 499-5480, Phone: 55 :31 499-548:3
E-mail: aguirre(Q.icpc1ee.ufmg.br

ABSTRACT - This is the second and final parI. of a In a subsequent phase, it was necessary both to develop
series of two papers on nonlinear dynamics and chaos. In criteria to detect chaotic dynamics and to establish C!y-
the first parI some tools. developed for analysing nonlinear namical invariants to quantify chaos (Guckenheimer, 1982;
systems, were c!esniiJec! in conjunction with a seI. of moc!els Eckmann and Ruelle, 1985; Denton and Diamond, 1991).
commonly llsed as benchmarks in the literature. This papel' Having succeeded in diagnosing chaos, the next step was to
investigates a llllIllber of isslles concerning the modeling, build mo deIs which would learn the dynamics f1'Om data
signal processillg anel control of nonlinear e1ynamics. This on the strange attractor. In this respect a number of
is carrieel ou! llsillg tbe tools and 1D0elels described in the model structures have been investigated such as local linear
first papel'. This inwstigation has th1'Own some new light mappings (Farmer and Sidorowich, 1987; Crutchfield and
on relevant p1'Oblems such as modeI parametrization, modeI McNamara, 1987), radial basis functions (B1'Oomhead and
validation. data smoothing anel control of nonliear systems. Lowe, 1988; Casdagli, 1989), neural networks (EIsner, 1992)
These issues are inwstigated using NARMAX polynomial and global nonlinear polynomials (Aguirre and Billings,
modeIs but it is believed that the conclusions are relevant 199.5c; Kadtke et alii, 1993). This phase is being currently
to nonlinear represelltaJiolls in general. Some numerical investigated with other equa.lly important issues concerning
examples are includecl. nonlinear dynamics such as noise reduction and control. It
is the objective of this papel' to p1'Ovide a brief int1'Oduction
to these issues.
1 INTRODUCTION
The outline of the papel' is as follows. Section 2 discusses a
Chaotic systems have attractec! a great deal of attention in numbel' of issues concerning the modeling of nonlinea dy-
the last three decades. Systems and mo deIs which undergo namics. Embedding techniques in general and NARMAX
chaotic regimes for a rather wide range of operating condi- mo deIs in particular are brief!y presented in that section.
tions have been found in virtually every branch of science Section 3 is concerned with the noise reduction problem,
and engineering. which is a major limitation in the modeling of nonlinear
dynamics. Section 4 provides a very superficial int1'Oduc-
In the evolution of the study of chaotic systems, several tion to the subject of control and synchronization of chaos,
distinct but sometimes co-existent phases can be distin- nonetheless several references are p1'Ovided for further read-
guished. In the first phase, chaos was recognised as a de- ing. Some final remarks are made in section 5.
terministic dynamical regime which could be responsible for
f!uctuations that hitherto had been regarded as noise and 2 MODELlNG NONLlNEAR DYNAMICS
therefore modeled as stochastic processes (Lorenz, 1963).

This section gives a quick view at nonlinear dynamics 1D0d-

oArtigo submetido em 04/07/95;
Revisão em 27/11/95 eling. This vital issue is discussed under several related
Aceito por recomendação do Ed. Consultor Prof.Dr. Liu Hsu headings. A helpful account of some of the main points on

50 SBA Controle & Automação jVol.7 no. IjJan., Fev., Mar. e Abril 1996
 which operates on the entire state OI' phase space but which

y'L\/\fÇJ1 yields just a scala.r which is caIled the measured variable.

The question which naturaIIy arises at this stage is the fol-
lowing: given .f : IR" --;. IR" and h(y) : IR" ~ IR, is it
possible to reconstruct a trajectory OI' sulution of .f from
to the scalar measurement h( y)?
y(t) E IRn
Fortunately, it turns out that this question has an a:ffirma-
tive answer if certain requirements are met (Takens, 1980;
Packard et alii, 1980; Sauer et alii, 1991). Thus embedol-
ogy is concerned with how to reconstruet the phase space
of a dynamical system of order n from a limited set of mea-
surements q where q < n, and more often than not q = 1.
In other words, the objective is to reconstruct the phase
Yl space of a system from a single time series. The resulting
phase space is usuaIIy referred to as embedded phase space,
y(to)
embedding space or just em.bedding.

Another question which should be addressed is: why should

Y2 we be concerned in reconstrueting the trajectories of a dy-
namical system? In the companion papel' it was shown
that in state (01' phase) space the steady state dynamics of
Y3 a system are represented by geometrical figures which are
caIIed attractors. A stable autonomous linear system only
has one kind of attractor, a point attraetor. However, non-
Figure 1 - The n time series defined by the state variables linear systems may have more complicated a.ttractors such
of an ntn-order dynamical system can be used to compose as limit cycles, tori and the so-caIIed strange attraetors.
the trajeetory in state space.
Therefore if time series are used to reconstruct the phase
space of dynamicaI systems via embedding techniques, it is
this subject can be found in the literature (Casdagli et alii, possible to use results from differentiaI geometry and topol-
1992). ogy to analyse the resulting attractors which are geomet-
rical objeets in the reconstructed space. Moreover, if the
embedding is successful, both the reconstructed and the
2.1 Embedding Techniques original attraetors are equivalent from a topological point
of view, OI' in other words, they are said to be diffeomorphic.
An n th-order dynamical system iJ = f(y) can be repre-
sented as a set of n first-order ordinary differential equa- The practical consequences of this are obvious. No matter
tions each one governed by a state variable. The global sys- how complex a dynamical system might be, even if only one
tem would therefore have n time variahles {Yl, Y2, ... , y,,} variahle of such a system is measured, it is possible to recon-
and the solution of such a system could be thought of as 11 struct the originaI phase space via embedding techniques.
time series. It is also possible to estimate qalitative and quantitative
invariants of the original attraetor, such as Poincaré maps,
In a sense, the n time series mentioned above are obtained fractal dimension and Lyapunov exponents, directly from
from the original n th-order system by decomposition. AIso, the reconstructed attractor which is topologically equiva-
given th 11 times series it is possible to recover the original lent to the originaI one. These ideas are iIIustrated in figure
n-dimensional solution by taking each state variable to be 2.
a coordinate of a 'reconstruction space' and to represent
each time series in such a space. Thus n time series can A convenient but by no means unique way of reconstructing
be used to compose ar reconstruct the system solution OI' phase spaces from scalar measurements is achieved by us-
trajectory. This is iIIustrated in figure 1. ing delay coordinates (Packard et alii, 1980; Takens, 1980;
Sauer et alii, 1991). Other coordinates include the singular
A di:fficulty encountered in praetice with this approach is value (Broomhead and King, 1986; Albano et alii, 1988)
that the order of the system n is seidom known and even and derivatives (Baake et alii, 1992; Gouesbet and Ma-
when an accurate estimate of this variable exists the num- quet, 1992). A framework for the comparison of several re-
bel' of measurements wiII not be as large as n. Take for constructions has been developed in (Casdagli et alii, 1991)
instance the atmosphere which is usuaIIy thought of as and three of the most conU11on methods have been studied
a high-order system, nonetheless monitoring and weather in (Gibson ei alii, 1992).
forecasting stations only measure a very limited numbel' of
variables of this system in order to make predictions. A delay veetor has the folIowing form

This can be described in a more mathematical way by con-

sidering the aetion of a measuring function h(y) : IR" --;. IR

SBA Controle &. Automação /Vol.7 no. I/Jan., Fev., Mar. e Abril 1996 51
 window defined as (de - l)T (Albano et alú, 1988; Buzug
and Pfister, 1992; Martinerie et alii, 1992). Some of these
methods have recently been compareel in (Rosenstein et alii,
1994). There is some evielence that the 'optimum' vaIue
of T in system ielentification problems is shorter than for
phas-space reconstructions (Aguirre, 1994). Dynamical re-
constructions frol11 nonuniformly sampleel data has been
aelelressed in (Breeelon and Packarel, 1992) and phase space
reconstruction of symmetric a.ttractors has been consielereel
in (King anel Stewart, 1992).

Taken 's theorem gives sufficient conelitions for equation (2

(De) Composition to holel, that is, in order to be able to infer elynamical in-

) variants of the original system from the time series of a

single variable, however no indication is given as to how to
estimate the map fT' A number of papers have been de-
voteel to this goal and snch methoels cano be separated into
two major groups, namely local anel global a.pproximation
techniqnes.

The local approaches usually begin by partitioning the em-

bedeling space into neighbourhoods {U;} ~\ within which
the elynamics can be appropriately elescribed by a linear
Figure 2 - In many practical situations the number of mea- map gT : IR~ -+ IR such that
sured variables is limited. Embedding techniques enable
the reconstruction of the state space even from a single
measurement. The reconstructed (OI' embedded) and the
original state spaces are equivaIent. y(k + T) ~ gT;(y(k)) for y(k) E Ui, i = 1, ... , N n . (3)

Several choices for gT have been suggesteel in the litera-

ture such as linear polynomiaIs (Farmer anel Sielorowich,
y(k) = [y(k) y(k - T) ... y(k - (de - I)TW, (1)
1987; Casdagli, 1991) which can be interpolateel to obtain
an approximation of the map h (Abarbanel et alii, 1990).
where de is the embedding dimension and T is the delay Simpler choices incluele zeroth-order approximations, aIso
time. Clearly, y( k) can be represented as a point in the known as local constant predictors (Farmer and Sielorowich,
de-dimensionaI embedding space. Takens (1980) has shown 19B7; Kennel anel Isabelle, 1992; Wayland et alii, 1993) anel
that embeddings with de > 2n will be faithful generically a weighted predictor (Linsay, 1991).
so that there is a smooth map h : IRde -+ IR such that
Global a.pproximators overcome some of the elifficulties
faced by local maps. Although global moelels have prob-
y(k + T) = h(y(h~)) (2) lems of their own, some attention has been elevoted to the
investigation of such models (Cremers anel Hübler, 1987;
Crutchfield anel McN amara, 1987; Kaeltke et alii, 1993;
for ali integers k, and where the forecasting time T and T Aguirre and Billings, 1995e).
are also assumed to be integers. A consequence of Taken's
theorem is that the attractor reconstructed in IR~ is diffeo-
2.2 Representation of Nonlinear Systems
morphic to the original attractor in state space and there-
fore the fonner retains dynamical and topological charac-
teristics of the latter. The Volterra series anel other related functional represen-
tations were among the first modeIs to be useel in nonlinear
In the case of delay reconstructions, the choice of the re- approximation. A well known elifficulty with such represen-
construction parameters, that is, the embedding dimension tations is the enormous amount of parameters requireel in
de and the delay time T is of the greatest importance since order to approximate simpIe nonlinearities (Billings, 1980).
such parameters strongly affect the quality of the embed- Related techniques seem to suffer from the same problem
ded space. The selection of de has been investigated in and, in addition, tend to require very large data sets (Giona
(Cenys and Pyragas, 1988; Aleksié, 1991; Cheng and Tang, et alii, 1991).
1992; Kennel et alii, 1992). The choice of the delay time
has been discussed in (Albano et alii, 1991; Buzug et alii, One of the most popular representations of dynamical mod-
1990; Fraser, 1989; Kember and Fowler, 1993; Liebert and eIs is the polynomial formo Apart from being easy to inter-
Schuster, 1989; Billings and Aguirre, 1995; Aguirre, 1995a). pret, simulate and operate, algorithms for the estimation of
Many authors have suggested that in some applications it the parameters of polynomial models are currently widely
is more meaningful to estimate these parameters simulta- available. One of the disadvantages of global polynomials,
neously, this is tantamount to estimating the embedding however, is that even for polynomial models of moderate

52 SBA Controle & Automação jVol.7 no. 1jJan., Fev., Mar. e Abril 1996
oreler, the number of terms can become impractically large a.pproximation depeneis on the choice of several operating
(Farmer anel Sielorowich, 198830; Caselagli, 1989). Using regimes where the system dynamics are approximately lin-
polynomials to forecast chaotic time series, Caselagli (1989) ear. This information has to be available a priori anel is
has reporteel that such preelictors blow up in the iterative somewhat criticaI. The problem of selecting the operat-
proceelure anel suggests that this is because polynomial pre- ing points is similar to the choice of neighborhooels anel of
elictors give bael approximants to the true elynamics except centres in other approaches.
very elose to the attractor. On the other hanel, some of
the problems relateel to global polynomials are believeel Other representations for modelling nonlinear systems in-
to be connecteel to the structure of the moelels (Aguirre elude Legenelre polynomials (Cremers anel H ü bler, 1987),
anel Billings, 1995b) anel promising results have been re- neural networks (EIsner, 1992; Principe ei alii, 1992) and
porteel for some systems using nonlinear global polynomials weighteelmaps (Stokbro anel Umberger. 1992).
with simplifieel structure (Kaeltke ei aliz, 1993; Aguirre anel
Billings, 1994c). At present no particular representation can be regareleel as
the best for any application anel "fineling a good represen-
Rational moelels share with polynomials the advantage of tation is largely a matter of trial anel error" (Farmer anel
being linear in the parameters. This feature males it pos- Sielorowich, 198830). On the other hanel, it seems that global
sible to use well known anel numerically robust algorithms polynomial models are in many respects simpler and there-
to estimate the parameters of such models. Moreover, 1'30- fore more convenient (Kadtke ei alii, 1993).
tional moelels seem to extrapolate better than polynomials
(Farmer and Sidorowich, 198830). The remainder of this section investigates the use of global
polynomials with simplifieel structure to estimate elynami-
The radial basis funetion (RBF) approach is a global inter- cal invariants of strange attractors.
polation technique with good localization properties anel it
is easy to implement as the algorithm is essentially inelepen-
2.3 NARMAX Models
elent of the elimension (Broomheael and Lowe, 1988; Cas-
elagli, 1989; Whaha, 1992). However performance of raelial
basis functions depeneis critically upon the centres (Chen ei Consider the nonlinear -ªutoregressive moving -ªverage
alii, 1990). For a few hunelreel elata points the choice of the model with e2:;ogenous inputs (NARMAX) (Leontaritis anel
centres is a elifficult task and the solution of the problem Billings, 198530; Leontaritis anel Billings, 1985b)
could become infeasible (Casdagli, 1989; Billings anel Chen,
1992).
y(k) = F( [y(k - 1), y(k - n y ),
Local approximants are concerned with the mapping of a
set of neighbouring points in a reconstructed state space u(k - d), u(k-d-n,,+l),
into their future values. A major problem here is to se- e(k), ,e(k-n e )] , (4)
lect the neighbourhooels because such a choice is criticaI
anel there coulel be hundreels OI' even thousands of these
(Farmer anel Sielorowich, 1988ab). The size of the neigh- where n y , 71" anel n e are the maxÍlnUll1 lags consielereel for
bourhooels elepenels on the noise leveI and the complexity the output, input and noise terms, respectively anel d is the
of the dynamics (Farmer anel Sielorowich, 1991). elelay measured in sampling intervals, T s . Moreover, u( k)
and y(k) are respectively input and output time series ob-
A simpIe alternative to nonlinear modeling is the use of tained by sampling the continuous data ll(k) anel y(ld at
piecewise-linear representations (Billings anel Voon, 1987). T s . Furthermore, e( k) accounts for uncertainties, possible
The eliscontinuities among the several linear mo deIs which noise, unmodelleel dynamics, etc. anel p([.] is some nonlin-
compose a piecewise-linear moelel, can provide effects sim- ear function ofy(k), 1l.(k) anel e(k) with nonlinearity degree
ilar to those observeel in nonlinear mo deIs such as elmos f.E Z+. In this papel', the map F([.] is taken to be a poly-
(Mahfouz and Badrakhan, 199030; MahfollZ and Baelrakhan, nomial of elegreee. In oreler to estimate the parameters of
1990b ). However, as other local representations, the fi- this map, equation (4) has to be expresseel in preeliction
nal model is piecewise-linear and therefore eliscontinuous. errar form as
Piecewise-linear moelels have been founel to be unreliahle in-
dicators of the underlying dynamics in some cases (Billings y(k) = wT(k - 1)<3 + ç(k) (5)
and Voon, 1987), anel a possible explanation for this is that
such modeIs violate the physically motivated hypothesis of
smooth elynamical systems (Crutchfield anel McN amara, where
1987). Thus local predictors may not always be suitable
for predicting invariant measures (Brown ei alii, 1991).

Smooth interpolation functions have been suggested as a

way of alleviating the problem causeel by eliscontinuities in (6)
piecewise-linear moelels (Johansen anel Foss, 1993). Such
functions have localiseel properties which confer to the final
moelels composeel in this way some similarities with raelial anel where w~" (k -1) is a matrix which contains linear anel
basis functions. As woulel be expecteel, the quality of the nonlinear combinations of output anel input terms up to and

5BA Controle & Automação IVo!.? no. 1/Jan., Fev., Mar. e Abril 1996 53
including time k-l. The matrices \Ii;"ç(k:-l) and \Ii[(k-l) ana.logous to the monovariable case (Bi11ings cf ahi, 1989).
are defined similarly. The parameters corresponding to ~ach
term in such matrices are the elements of the vectors 8 y ", The quantity ERR provides an indication of which terms to
êy"ç anel ê ç , respectively. Fina11y, é,( k:) are the residuaIs include in the model. In other words, the ERR test provides
which are defined as the difference between the measureel a means of ordering all the candidate terms according to a
data y(k) and the one-step-ahead prediction \liT(k - 1)8. hierarchy which depends on the l'elative importance of each
The parameter vector 8 can be estimated by minimizing termo It should be noted that no trial-and-error is necessary
the fo11owing cost function (Chen ci alú, 1989). for this. However, the fo11owing question arises: how many
terms should be included in the model? A practica.l way of
addressing this question is by means of informaiion criieria
(7) such as the final prediction error (FPE) (Akaike, 1974),
Akaikc 's information criterion (AIC) (Akaike, 19(4), the
Bayes'ian infon11aiion criterion (ElC) (Kashyap, 1977), the
where II . II is the Euclidean norm. Moreover, least model eniropy (Crutchfield and McN amara, 1987) and the
squares minimization is performed using orthogonal tech- Schwar:: informaiion criierion (Mees, 199:3). See (Gooijer
niques in order to effectively overcome two major elifficul- ei alá, 198:3) for a survey of such techniques.
ties in nonlinear model identification, namely i) numeri-
cal i11-conditioning and ii) structure selection. In order
to circumvent such problems orthogonal techniques may 2.5 Model Validation
be used (Bi11ings ei alá, 1988; Korenberg and Paarmann,
1991). The last step in any identification problem is the valida-
tion of the estimated models which is not a trivial problem.
Most 'conventional' approaches to mo dei validation are not
2.4 Structure Selection particularly attractive when the mo deis are chaotic and
therefore alternative invariants should be used to quantify
The number of terms in a polynomial grows very rapidly the quality and adequacy of the estimated models. When
even for relativdy lo\\' values ofe, n y , n" and n e . This is va.lidating nonlinear models it is desirable that the crite-
too difficult a problem to be solved by trial and error. How- ria used should be sensitive to the 'fundamental' features
ever, effective anel elegant solutions to handle this prob- of the models. In this respect it has been shown that bi-
lem are availahk. see (Bi11ings ci alá, 1988; Aguirre and furcation diagrams are far more sensitive to variations in
Bi11ings, 199!)d) anel the survey paper by Haber anel Un- model structure (Aguirre and Billings, 1994c) than many
behauen (Haber anel Unbehauen, 1990). One solution is other nonlineal' invariants used in model validation such as
the crror ndl/clion ratio (ERR) test (Bi11ings ei alii, 1988; Poincaré maps (Casdagli, 1989; Gottwa.ld ei alii, 1992) cor-
Bi11ings ei alii. H)S9: h:orenberg ei alii, 1988). Two advan- relation dimension (Grassberger ei alli, 1991), Lyapunov
tages of this approach are i) it does not require the estima- exponents (Abarbanel et alii, 1989; Principe ei alii, 1992),
tion of a complete model to determine the significance of a reconstructed phase-space plots (Adomaitis ei alli, 1990).
candidate term and its contribution to the output, and ii) Recently, the concept of synchl'onization, which is rather
the ERR test is eleriveel as a by-product of the orthogonal we11 known in the field of control of chaos, has been sug-
estimatioll algori t hm. gested as a nontrivia.l test for validating estimated mo deis
(Brown ei alli, 1994).
Because the final models will be composed of a reduced
number of terms. which is a sma11 fraction of the total num- Finally, it should be pointed out that most of the results
ber of candidate terms, the models in this paper can be described so far in this section have been obtained using
viewed as 'simplified' or 'concise' global polynomials. It is the tools described in the first paper of the series. Many of
believed that these moelels overcome some of the practica.l the commonly used statistica.l toois are of very limited use
difficulties usua11y reported for non-simplified polynomials. in assessing dynamical properties of estimated models.

In a recent paper it has been shown that the ERR criterion

gives qua.litatively similar results as higher order spectrum
2.6 NARMAX and Other Approaches
techniques in eletecting nonlinear interactions within the
under1ying dynamics (Aguirre and Bi11ings, 1994a). A cri- Unlike many localized techniques, the NARMAX approach
terion similar to ERR has been used to generate radial basis does not involve finding neighborhoods, thus N n = 1 and
functions with a small number of parameters (Mees, 1993). a11 the data belong to a unique 'neighbourhood', that is,
Other techniques can be used in connection with ERR such y(k) EU 1 k = de, ... , N. This reduces the number of data
as the zeroing-and-refitting approach (Kadtke ct alii, 1993) required to estimate the dynamics. Moreover, the delay
and the concept of term clustering (Aguirre and Bi11ings, time is taken to be equal to the sampling period, thus T =T s .
1995d).
There are a number of important differences between NAR-
In a multivariable modei with l' inputs and m outputs, MAX polynomial identification and other methods. First, a
the entries in equation (4) are vectors, that is tt(k) = NARMAX modei includes input terms. This enables fitting
[uttk) 'Ur(kW, y(k) = [ydk) ... Ym(kW and , e(k) = data from non-autonomous systems and therefore estimat-
[e1 (k) em (k)]T, and both structure detection and pa- ing input/output maps, see (Casdagli, 1992; Hunter, 1992)
rameter estimation can be performed 111 a way which is for related ideas on this subject. An immediate consequence

54 SBA Controle & Automação IVo!.? no. l/Jan., Fev., Mar. e Abril 1996
of this is that for inputjoutput systems, it is not required (a)
that the output be on any particular attractor. Once an
inputjoutput modeI has been estimated, a particular input 1.Sr--~--~--~-~--~----,
can be used to generate data on a specific attractor.

Another important difference is the presence of noise terms,

that is, the moving average part of the model. It should be o.s
noted that equation (2) will only hold in the unlikely case
when noise is ahsent. Any noise in the data OI' any im-
perfection in the estimate of the map h will result in an
-o.s
extra term in the right hand side of equation (2). Such
a term would be responsible for modelling the mismatch -1
introduced by the noise and unmodelled dynamics. It is
a well known result in the theory of system identification -':l.Ls--_~,---o~.-=-s --~o--~o.-=-s --~----",.s
that if such a term is omitted from the model structure, the x(k)

estimat.e of the map IT will become biased during param-

(b)
eter estimation (Soderstrom and Stoica, 1989) and llonlin-
ear models are no exception to this mIe (Billings and Voon,
1984). 1.Sr--~--~--~-~--~----'

It seems that when the noise is white and enters the sys-
tem as a purely additive component, the division of the
o.s
data into neighbourhoods and subsequent estimation re-
duces the bias. This will not be the case however if the
modeI is global 01' if the noise is correlated. Thus in 01'-
der to avoid bias a mo deI for the noise and uncertainties, -o.s
w;lIç(k - 1)0 ylIç + wl(J..~ - 1)Ê>ç, is included in the model
-, .-.- .....
structure before proceeding to parameter estimation. Once
parameters have been estimated, only the deterministic
part of the model is used, namely W;lI (J..~ - 1)Ê>Yll' This -':l'::-.s---~,---o-::'.-=-s---=-o---::'o.-=-s--c------:",.s
x(k)
procedure can handle moderate amounts of white and cor-
related noise.
Figure :3 - First return map for (a) the Hénon map con-
taminated with noise. Only the encircled data were used in
Summarising, equation (5) is a hybrid mo deI since it is com-
the estimation, (b) the identified map of equation (8).
posed of a deterministic part and a stochastic component.
The latter is only used during parameter estimation in 01'-
der to avoid bias on the formeI'. Therefore, in this papel'
the modeIs used to generate the surrogate data are purely marked with circles
deterministic a.!though the stochastic part of the mo deIs is
also represented for clarity. Thus, the deterministic com-
ponent of the identified mo deIs is an approximation to the
dynamics, that is, fT :::::: W;lI(J..~ - 1)Ê>y" where T=Ts · x(k) = 0.99433 - 1.3818 x(k - 1)2 + 0.29302 x( J..~ - 2) . (8)

2.7 Modeling of Dynamical Systems

The first return map for this equation is shown in figure 3b
and has a correlation dimension of De = 1.11 ± 0.22 which
In this section some of examples are given to illustrate the shows good agreement with the origina.! map for which
performance of NARMAX polynomia.!s in the moelelling of De = 1.21 ± 0.01. The correlation elimension estimated eli-
nonliner systems. It is worth stressing that the stochastic rectly from 20000 data points with the same SNR as above
component of some modeIs is represented for greater clarity was De = 1.76 ± 0.06 revealing that the estimated value
since such component is needeel eluring parameter estima- is quite sensitive to such leveIs of noise. Further improve-
tion to avoid bias. However, only the deterministic part is ment can be achieved by using more than 50 points, but
aetually used to iterate the modeIs in arder to generate the the objective in this example was to show that the map
figures. The emphasis is on the reconstruction of dynamical can be estimated fairly accurately from a short and noisy
properties. time series.

In the case of driven oscillators, the practica.! reconstruction

2.7.1 Poincaré Sections of Poincaré sections is restricteel to controlled experiments
because of the large amount of data required since only one
point in such sections is obtained for each forcing period.
The first return map for 1000 points taken from data ob- Moreover, sma.ll amounts of noise often blur the delicate
tained from the Hénon map (Hénon, 1976) is shown in figure fracta.! structure of the attractor anel the Poincaré sections
3a. The following model was estimated from the 50 points tend to become fuzzy.

SBA Controle & Automação jVol.? no. 1jJan., Fev., Mar. e Abril 1996 55
The fuzziness in the Poincaré sections introduced by the (a)
noise is a direct consequence of superimposing a stochas-
tic component on the top of a purely deterministic tra-
jectory. The Poincaré sections reconstructed using NARX
modeIs do not suffer from such fuzziness because although
the stochastic component was present during the model es-
0.5
timation, the effects of such a component were 'absorbec!'
by the moving average (MA) part of the moc!el enahling ã: O
unbiased estimation. The NARX part of the moc!el, which ~
>;
is purely deterministic. will 110i inc!uce any fuzziness in the -0.5
Poincaré sections. The fact that the moc!els are not per-
fect, however, will be revealed by possible distortions in the -1
shape 01' the reconstructec! attractors.
-1.5
The following model was obtainec! from 1500 data points -1.5 -1 -0.5 O 0.5 1.5
on the Duffing- Holmes attractor shown in figure 4a y(t)
(b)
y(k) 0.84725 y(l.:-l)+ 0.35713 y(I.:-3)
+ 0.12780 x 10- 1 y( k-4)
0.694:31 x 10- 1 y( k _1)3
+ 0.61:319 x 10- 1 tl( k-l) + 0.40:325 y( k - 2)
0.24:349 x 10- 3 y( k-1)y(I.:-2)y(k-5) - 0.46215 y(k-5)
+ 0.096:n9u(k-:3) - 0.15316 y(k-2)y(k-3)y(k-4)
0.7:3618 x 10- 2 y(k-l )2 y ( k-5)
+ 0.0711:.'22 1/(/;-I)y(k-3)y(k-4)
+ IJiT(k - 1 )Êle+~(k) . (9)

-1.5
An estimate of lhe original Poincaré section is shown in fig-
ure 4b which \"as obtained by iterating equation (9). This -1.5 -1 -0.5 O 0.5 1.5
reconstruct.eel Poincaré section is very similar to the original y(k)
one.
Figure 4 - Poincaré sections (a) obtained f1'Om a noisy 01'-
bit of the Duffing- Holmes oscillator, (b) of the identified
moc!el ofequation (9). A=0.3 anel w=lrad/s,~) = 5.
2.7.2 Bifurcation Diagrams

Using 1500 data poilllS generatec! by simulation of the mod-

ified van der Pol eqnat.ion (see first paper), with this SNR, See figure 14a of the first papel' in the series for the original
sampled at T" =;r /SO. the foJlowing NARMAX model was bifurcation diagramo This is usually more demanding than,
estimated for instance, requiring that a model should reproduce in-
variants associated with particular attractors (Aguirre and
Billings, 1994c).
y( k) + 0.87488 x 10- 1 y( k -4)
0.8359~1 y( k -1)
+ 0.68539 x 1O- 1 u(k-2) + 0.46776 x 10- 2 y(k_l)3
0.47330 y(k-6) + 0.12786 y(k-2)
+ 0.:37341 y( k - 3) - 0.22840 X 10- 2u.( k -1)
+ 0.49.504 x 10- 1 y(k-5) - 0.014841 y(k-l)2 y (k-2) 2.7.3 Original and Embedded Trajectories
0.081389u.(k-3) + 0.038305 tl(k-5)
0.13.5.54 x 10- 1 u( k -4) + 0.20404 x 10- 2 y(k - 2)2 y( k - 3) This section reports some results Cüncerning the use of
0.34234 X 10- 2 y( k -1)y( k - 6)2 NARMAX polynomials in rep1'Oducing embedded and orig-
+ 0.35999 x 10- 2 y(k-2)y(k-4)y( k-6) inal trajectories of strange attractors. To investigate this
+ IJiT(1.: - I)Êle+~(k) . (lO) the well known Chua's double scroll attractor is used. If
ali variables are measureel, multivariable NARMAX mod-
eIs can be fitted to the data and the iterated discrete-time
outputs can be used to reconstruct the original attractor
This model also has a self-sustained oscillation with w = geometry in state-space.
1.56rad/s.
The data in figure 6a were used to identify the following
Figure 5 shows that the ic!entified model (10) does repro- multivariable model
duce the major bifurcation patterns of the original system.

56 SBA Controle 8< Automação jVol.7 no. ljJan., Fev., Mar. e Abril 1996
 3,--------.----.--------.-----,.-------, (a)

20
N

-2

-4
0.5
_4'------'--.----~-----"----L-------.J

5 10 15 20 O 2
-0.5 -2 O
A
y(t) x(t)
Figure 5 - Bifurcation diagram of the identified model of
(b)
equation (10). w=4rad/s.

2
:r(k) 0.11282 x 10x(k -1) + 0.55867 y(k - 1)
0.47190 X 10- 1 :r( k - 1)3
go
N
+ 0.39895 X 10-1 y( k - 1 )z( k - 1)2
0.31229 X 10- 2 ;;( k - 1 )2. + 0.18363 x 10-1 z(k - 1)
-2
+ 'lilrçyç, (k - 1 )Ê>çrçyç, +~x( k)

-4
y(k) 0.91948y(k -1) - 0.10392x10- 3 z(k _1)2' 0.5
+ 0.70843 x 10-1 x( k - 1) + 0.67800 x 10- 1 ;;( k - 1) O
O 2
-0.5 -2
0.13424 X 10- 2 x( k - 1)3 y(k) x(k)
+ 0.44206 x 10- 3 x(k - l/y(k - 1)
Figure 6 - (a) Noisy trajectory used for identification,
+ 'liL M , (k - l)Ê>çrM' +~y(k)
SNR=72.9, 39.9 and75.5 dB for x, y and z components, re-
spectively. (b) Double scroll Chua's attractor reconstructed
;;(k) 0.96628 z(k - 1) - 0.95854 y(k - 1) from the identified modei in equation (11).
0.36719 X 10- 1 x( k - 1) - 0.55765 X 10- 1 y(k - 1)3
+ 0.10333x10- 2 x(k _1)3
+ 0.0020536 x(k-1)y(k-l)z(k-1)
+ 'lilrçyç, (k - l)Ê>çrçyç, +Uk) . (11) systems. This comprises one of the current phases in the
investigation of chaos (Mitschke, 1990; Chennaoui et alii,
1990; Schreiber and Grassberger, 1991; Broomhead et alii,
1992; Davies, 1992; Grassberger et alii, 1993; Holzfuss and
This estimated model settles to a strange attractor which Kadtke, 1993).
closely resembles the original double scroll Chua's attractor,
see figure 6b. Some authors assume that some kind of a priori knowl-
edge concerning the original system is available such as a
Similar models for the Lorenz and Rossler attractors have piece of noise-free data (Marteau and Abarbanel, 1991),
been reported in (Aguirre and Billings, 1995e). the structure of the maps describing the underlying dy-
namics (Davies, 1992), ar even the complete maps, that
is, structure and parameters are known (Hammel, 1990;
3 NOl5E REDUCTION Ozaki, 1993). However, a clear limitation in any real noise
reduction problem is that the underlying dynamics are not
A difficulty which a.ppears to be common to most ap- usually known a priori and the map has to be estimated
proaches for modelling nonlinear dynamical systems and (learned) from the noisy data as an integral part of the
chaotic attractors is that realistically noise will be present noise reduction processo Consequently, the noise will pose
in the data. In particular, it has been conjectured that the limitations on the amount of noise which can effectively be
local divergence of nearby orbits in a chaotic system seems eliminated. In the field of nonlinear dynamics, the main
to impose a natural limit on the accuracy of prediction- objective of filtering a chaotic time series is to enable the
based identification algorithms when the data are noisy reconstruction and estimation of dynamical invariants such
(Aguirre and Billings, 1995c). Consequently, there has been as Poincarê sections, Lyapunov exponents and fractal di-
great motivation to develop filtering techniques for chaotic 111enSlOns.

SBA Controle & Automação jVo!.7 no. ljJan., Fev., Mar. e Abril 1996 57
Another objective of filtering nonlinear data is to enable
the identification of dynamically valid mo deis which would
reproduce the aforementioned invariants from sequences of II [.] =11 f}(k) - gdf}(k - 1)) li" + 11 f}(k + 1) - gJ.;(f}(k)) li"
filtered data. In a first attempt to solve this identifica- (13)
tion problem, globalnonlinear predictors were used to filter
the data (Aguirre and Billings, 1995c). In such a proce- where 11 . II is the Euclielean norm, anel
dure the noise was separated from the signal by means of
(very) short-term predictions. Because a chaotic predictor
actl1 ally am plifies uncertainties in some' direction' in state h['] =11 f}(k) - y(k) li" (14)
space, the aforementioned approach cannot be used to fil-
ter the noise by successive passes through the data and this
Another option is to choose h['] as above and
therefore limits the achievable noise reduction.

This problem can be alleviated by using global smoothers

because, unlike prediction-based techniql1es, smoothers are
lI['] = 211 gJ.;(f}(k)) - f}(J..~ + 1) II T
Pk (15)
well suited for filtering chaotic data via successive noise-
reduction iterations (Aguirre et ali!, 1995). where Jik are Lagrange multipliers (Fanner and Sielorowich,
1991).

3.1 Filtering Techniques In the field of system identification, improving the signol to
noise ratio (SNR) is also of interest because this facilitates
both the unbiaseel estimation of the parameter vector anel
It is usua.lly assumed that the noise is purely additive, 01' in
other words the noise is entirely observational (Crutchfield the correct determination of the moelel structure. The chief
and McNamara, 1987; Casdagli ef alii, 1991; Grassberger et idea is to estimate the noise-fl'ee e1ata anel then use this
alii, 1991). Thus the noise reduction problem can be stated estimate to perform parameter estimation. A way of doing
as follows: given a chaotic time series x(t). it is desired to this is to use the following preelictor which can be e1erived
from equation (5)
filter the measured data y(t)=x(t)+e(t), where e(t) is the
addi tive noise, in order to recover x (i). This is useful in
'deaning' Poincaré sections and embedded attractors which
have been blurreel by noise. (16)

Another aspect of this problem is to find a 'noise-reduced' It should be realiseel that in the last equation the parame-
orbit f}(i) from which invariants such as '\1, De and the ter vector 0 yu was estimateel from the original noisy data
attractor geometry can be more accurately estimated than as is indicateel by the absence of the hat on the subscript
if the noisy data y(t) were used. This is sometimes referred y. On the other hand, the matrix lliJu
(t - 1) was formed
to as statistical noise reduction as opposed to recovering using predicted values of the data, that is y(t.) up to and
x(t) from y(t) which has been called detailed noise reduetion induding time t - 1. Because y( t) is an estimate of x(t),
(Farmer anel Sidorowich, 1991). In this paper, the objective equation (16) can be used in suboptimal parameter esti-
is to be able to identify e1ynamically valid models from f}(t). mation schemes (Billings and Voon, 1984). However, ifthe
data were chaotic aftel' a few iterations iJ(t) would not be
Filtering based on model predicteel outputs, whilst reducing an accurate estimate of x(t) because of the sensitive de-
the noise content, in the data, willnot guarantee that f}(i) pendence on initial conditions. Therefore the use of iJ(t) in
remains dose to y(t) (and ultimately dose to xCi)) if the suboptimal schemes seems somewhat restricted for chaotic
latter is chaotic. Thus, to ensure that f}(t) remains close to systems. The next two sections describe approaches which
y(t), the following cost function can be useel overcome some of these problems.

3.2 The Resetting Filter

N
JNR =L {lI [y(k) - gJ.;(f}(k - 1))] + h [y(l.~) - y(k)]} The following predictor has been suggesteel to overcome
k=l
some of the e1ifficulties associated with the filtering of
(12)
chaotic data (Aguirre and Billings, 1995c)
where N is the number of points in the data, II [.] and
h['] indicate functions which are usually metric norms and
gk (-) are linear maps which describe the dynamics in a
neighbourhood of a point on the tme orbit. Clearly II [.]
penalizes deviations from the true deterministic dynamics It should be noted that in this case iJ( t) is predicted based
described by gk(-) while h['] guarantees that the deaned on previous values of the measured data y( s), s ::; t-l, and
orbit remains dose to the measured orbit. not based on previously prec!ictec! values such as in equation
(16). Moreover, since this prec!ictor is usec! to prec!ict only
In particular, the following cost functions have been used one step into the future, the preelicted value iJ(t) is, in most
(Kostelich and Yorke, 1988; Kostelich and Yorke, 1990) cases, guaranteed to remain dose to the data y(t). This can

58 SBA Controle & Automação jVol.? no. 1jJan., Fev., Mar. e Abril 1996
be interpreted as being a consequence of the resetting effect
achieved by using measured data to initialise the predictor r ~ r r
1.2
r J
~
at each step. The predictor in equation (17) wiH be referred
to as the resetting filter (RF) and it is adequate for filtering
chaotic signals. N

~
A
The qualitative effect attained by the resetting filter is, in 0.8
V
\~
some respects, analogous to other methods. The resetting
V
effect of the RF guarantees that h['] (see equation (12))
is kept smal!. Moreover. the parameter vector of the RF is
obtained by millimising JLS in equation (7), which is clearly 0.4
analogous to -h['] in equation (12). The main difference ~ V \
~
is that whilst gd') usually represents local linear maps,
\JíT(t - 1) e is a global nonlinear map which may include
o 200 500 700
N
inputs and residual in addition to output terms. Figure I - Noise-free original time senes for the
Mackey-Glass modei
Predictor-basecl filtering for chaotic systems will not work
in general because of the inability of making long-term
~
#~
~
accurate predictions along the unstable manifold. There-
fore in such directions, the filter would actually amplify 1.2
M
~
the noise (Schreiber and Grassberger, 1991). The same is
valid for the RF, but to a much lesser extent because of the
J ~ ~ ~
resetting effect which will guarantee that any noise ampli-
fication along the unstable manifold is kept to a minimum. 0.8 r ~
However, if several passes through the data are required to
attain the desired levei of noise reduction, it is inevitable

,
that the effect of positive Lyapunov exponents be mani-
fest. Consequently, the filtered data may not resemble the 0.4
original sequence and, in fact, might have a greater noise
content than the raw data. \
o 200 500 700
N
3.3 Global Nonlinear Smoothers Figure 8 Noisy (raw) time series for the Mackey-Glass
mo dei
The difficulty with the resetting filter in equation (17) is
that it only uses past information to predict the future.
However, the dynamics can only be predicted with any cer-
tainty as t --;. ex) along the stable manifold. Conversely, the
dynamics can only be 'predicted' along the unstable man- unstable directions because it contains terms which relate
ifold in reverse time, that is as t --;. -<Xi (Schreiber and to the future. Such terms will enable 'predicting' in reverse
Grassberger, 1991). In other words, in order to estimate time since orbits converge along the unstable manifold as
y(t), future information as weH as past information is re- t --;. -'x'. An iterative procedure for smoothing data us-
quired (Schreiber and Grassberger, 1991). ing nonlinear smoothers has been given in (Aguirre et alii,
1995).
This motivates the search for NARMAX smoothers of the
form

y(t) F s' [y(t - n y ), ... , y(t - I), y(t + 1), ... , y(t + n y ),
'U(t - d - nu + 1), ... , 'U(t - d), 'U(t + d), ...
+ d + nu -
... , 'll(t I),
ç(t-1), ... ,Ç(t-n e )]+ç(t). (18)
3.3.1 An Example

Zero-mean Gaussian noise was added to a set of data ob-

It should be noted that equations analogous to (5)-(7) can tained from the Mackey-Glass model (Mackey and Glass,
be derived for the smoother in equation (18). Moreover, 1977) (see first papel' in the series). The resulting records
the ERR criterion, used to select the most important terms with SNR=40.2 dB were smoothed with global nonlinear
to compose a NARMAX model, can also be used to select smoothers. Figures 7-9 respectively show the noise-free,
the structure of the smoother and the same least squares noisy and smoothed data for the Mackey-Glass model and
algorithm can be used to estimate the parameters. figure 10 shows the resulting data filtered following the
RF approach. These figures make it plainly clear that
From a dynamical point of view, the smoother in equation prediction-type approaches for noise reduction are not suit-
(18) will also succeed in predicting the dynamics along the able for chaotic data.

SBA Controle & Automação /Vol.7 no. l/Jan., Fev., Mar. e Abril 1996 59
 enhances heat transfer (Chang, 1992), improves 111lXlllg
in chemical reactions (Ottino, 1992), reduces ielle-channel
tones in modulators (Schreier, 1994) anel seems to have a
promising future in secure communication systems (Cuomo
ei alá, 199;3; "'"U anel CIma, 1993; Parlitz ef alii, 1992).
In aclelition, some authors have suggested that chaotic cly-
0.8 namics inelicate a hea.lthy state as opposed to the eliseases
which manifest as physiological periodic signa.ls (Glass et
alii, 1987; Golelberger ei alii, 1990). The matter of how
healthy chaos is, however. is far from settleel (Pool, 1989).
0.4 Consequently, techniques for controlling nonlinear elynam-
v V ics are requireel in oreler to provoke or suppress chaos or any
other elynamical regime accoreling to the particular appli-
o 200 700 cation at hand (Chen anel Dong, 1993a; Ditto anel Pecora,
1993; Hunt anel Johnson, 1993).
Figure 9 - Smoothed time series for the Mackey-Glass
model. Terms with both negative and positive lags were
used to compose the smoother, N p = 10. Most of the works concerned with the control of chaos are
clevoted to stabilising a chaotic system to regular clynam-
ics, that is, fixeel points, perioelic orbits or quasiperioelic
regimes. The relateel pl'Oblem of elriving a system from a
1.2 regular to a chaotic regime has received less attention. This
type of control coulel be important in situations where chaos
is not only welcome but a.lso desirable (Golelberger ei alá,
1990; Chang, 1992; Ottino, 1992; Cuomo ei alii, 199;3; Wu
0.8 and CIma, 1993).

Clearly, chaos is per se neither beneficiaI nor harmful as

describeel by James Gleick "In some applications, turbu-
0.4 lence is desirahle - insiele a jet engine, for example, where
efficient burning elepenels on rapiel mixing. But in most,
turbulence means disaster. Turbulent airfiow over a wing
o 200 500 700 elestl'OYs lift. Turbulent fiow in an oil pipe creates stupefy-
N
ing elrag" (Gleick, 1987, p. 122). Therefore it seems appro-
Figure 10 Filtewd time series for the Mackey-Glass priate to search for control schemes which would perform
model. Only l('rl11;'; \"it-h negative lags were used to compose well in both situations.
the resetting filIeI'. X p = 10.
If on the one hand sensitivity to initia.l conditions hampers
prediction-based contl'Ol schemes, on the other hand such
a property might turn out to be greatly advantageous from
4 CONTROL AND SYNCHRONIZATIDN a control point of view. To see this it should be recalleel
Df CHAOS that if a system is sensitive to initial conditions, a sma.ll
perturbation at time ia can provoke relatively large effects
Is chaos a beneficiaI dynamical steaely state? This is a cen- at time i > ia. This means that to achieve a certain contraI
tral question in tlw control of chaotic systems. Of course, objective may require a much smaller control action if the
if the answer to the above question is yes, a.pplieel scien- system were chaotic. The prablem of COUl'se is to determine
tists and contl'Ol engineers would be investigating ways of how and when shoulel the control adion be applied. Some
provoking chaos rat-her than suppressing it. A negative an- works in this direction have appeared in the literature (Ott
swer, on the other hand, woulcl pl'Ompt researchers in the ei alii, 1990; Ditto ei alii, 1990; Garfinkel ei alii, 1992;
opposite elirection. Nitsche anel Dressler, 1992; Romeiras ei alii, 1992; Shinbrot
ei alii, 1990; Spano ei alá, 1991).
Because of the sensitive depenclence on initial conelitions,
displayed by chaotic systems, it is impossible to make ac- Many different techniques have been investigated in the
curate long-term preclictions of such systems. In many sit- context of contl'Olling chaos. Most met.hods can be groupeel
uations, however, it is clesirable that the system uneler in- into two categories. \iVhen it. is desireel that. chaos be sup-
vestigation be predictable. Furthermore, the appearance of pressed the appraaches are labelled under conirol of chaos
chaotic dynamics is not always welcome because in some and when t.he main objective is to make a syst.em follow
situations it has been associatecl with abnonnal behaviour a chaot.ic trajectory t.he problem at hand is referred t.o as
(Glass and Mackey, 1988, pages 177, 179). synchronizaiion of chaos.

In other applications the onset of chaos seems to have sev- Chaos can be suppressed by the addition of small ampli-
eral aclvantages. For instance, it has been argued that "a tude perturbations (Braiman and Goldhirsch, 1991), ran-
cognitive system musi be chaotic in order to perform ef- dom perturbations (Kapitaniak, 1991), by paramet.ric driv-
fective signal processing" (Nicolis, 1984). Further, chaos ing (Dorning ei alii, 1992; Fl'Onzoni ei alii, 1991; Lima and

60 SBA Controle & Automação jVol.? no. 1jJan., Fev., Mar. e Abril 1996
Pettini, 1990), by means of feedback (Liu et alii, 1994). revieweel in a rather general framework anel two algorithms,
the resetting filter anel nonlinear smoothers, have been de-
The problem of synchronization has been investigated in scribed in some eletail. It has been pointeel out that if the
(CIma et alii, 1993; Kocarev et alii, 1993; Ogorzalek, 199:3; elata are chaotic special algorithms are usually required to
Pecora, 1990; Wu anel CIma, 1993). achieve effeetive noise reduction.

The stahilization of chaotic systems has been achieveel by Finally, a major issue in nonlinear dynamics nowaelays is
applying feedback (Chen and Dong, 1993b; Deelieu anel the control of chaotic systems. An enormous amount of
Ogorzalek, 1994; Hunt, 1991; Pyragas, 1992; Roy et alii, papers have been publisheel on this subjeet in the last years
1992), frequency harmonic balance techniques (Genesio and and a thorough review would be impossible. Nonetheless,
Tesi, 199:3; Genesio anel Tesi, 1992), conventional control several relevant references have been provieled in oreler to
techniques (Hartley anel Mossayebi, 1993), open plus closeel enable the reaeler to further investigate this topic.
loop control (.lackson anel Grosu . 1994), dynamical vibra-
tion absorbers (Kapitaniak et alii, 1993), adaptive control Throughout the papel' it has been shown how NARMAX
(Sinha et alii, 1990; Vassiliaelis, 1993) and quantitative moelels can be useel in the various problems concerning the
feeelback elesign (QFD) (Yau et alii, 1993). The control of moeleling, noise reduction and control of nonlinear systems
multipie attraetor systems has been investigateel in (.J ack- and chaos.
son, 1990).
Acknowledgements
Most of the references above are concerneel with systems
which are chaotic before control is applied. However, chaos
has been eletected in control systems in which the plant Financiai support from CNPq under grant 522538/95-9 is
was not chaotic. Conelitions for the occurrence of chaos in gratefully acknowledged.
feeelback systems (Genesio anel Tesi, 1991), adaptive control
(Mareels anel Bitmead, 1986; Mareels and Bitmeael, 1988;
Golden anel Yelstie, 1992) anel in eligital systems (Ushio and REFERENCE5
Hsu, 1987) have been reported in the literature.
Abarbanel, H. D. 1., Brown, R., and Kadtke, .l. B. (1989).
The use of ielentifieel moelels in the design of control schemes Preeliction anel system identification in chaotic non-
has been addressed in (Aguirre anel Billings, 1994b; Aguirre linear systems: time series with broaelband speetra.
anel Billings, 1995a; Aguirre, 1995b). It turns out that as Phys. Leit., 138(8):401-408.
long as an ielentified model reproduces some of the major
Abarbanel, H. D. 1., Brown, R., anel Kaeltke, .l. B. (1990).
elynamical features of the system, such a moelel can be used
Prediction in chaotic nonlinear systems: Methoels for
effectively in control problems. Many control schemes do
time series with broaelband Fourier speetra. Phys.
not require a moelel 01' may even work with a moelel which is
Rev. A, 41(4):1782-1807.
not dynamically valid but in such cases the control effort is
usually greater and the control quality significantly poorer. Adomaitis, R. A., Farber, R. M., Huelson, .l. L.,
Kevrekielis, I. G., Kube, M., anel Lapedes, A. S. (1990).
Application of neuralnets to system identification anel
5 FINAL REMARK5
bifurcation ana.\ysis of real world experimental elata.
In Neural N etworks: Biological cornputers or e!ectronic
The subject of nonlinear dynamical systems has attraeted brains, pages 87-97. Springer Verlag, Paris.
great attention in recent years. It is therefore natural that
various techniques for moeleling and reconstructing such Aguirre, L. A. (1994). Some remarks on strueture se-
systems shoulel be investigated. In this respeet a landmark leetion for nonlinear moelels. Int. 1. Bifurcation and
has been Taken's theorem and a number of subsequent re- Chaos, 4(6):1707-1714.
sults which today form the fielel of embedology. Parallel to
these results, other techniques were developed by the en- Aguirre, L. A. (1995a). A nonlinear correlation function
gineering community. Such methoels for the identification for seleeting the delay time in dynamical reconstruc-
of nonlinear systems useel other model structures such as tions. Phys. Leit. A, 203(2,3):88-94.
Volterra and Wiener models, NARMAX modeIs and neu-
Aguirre, L. A. (1995b). The use of identified mo deis in
ral networks. In the first part of this papel', the basic idea
the control of chaotic systems. In Preprints of the
of embeelding techniques has been reviewd. Similarly, the
IEEE International Symposium 011. Circu-its and Sys-
estimation on NARMAX polynomial moelels has been dis-
tems, pages 1528-1531, Seattle, USA.
cusseel and some differences between such approaches have
been pointed out. The moeleling of some nonlinear systems Aguirre, L. A. and Billings, S. A. (1994a.). Discrete recon-
has been illustrateel by numerical examples. struetion of strange attractors in Chua's circuito 111.1.
J. Bifurcatio1/. and Chaos, 4( 4) :853-864.
A major limitation in obtaining a good model for a nonlin-
ear system is the noise present in the data. Noise can be, in Aguirre, L. A. anel Billings, S. A. (l994b). Moelel ref-
a few cases, kept to a minimum but cannot be totally elim- erence control of regular and chaotic dynamics in the
inated. Consequently there has been great interest in noise Duffing-Ueda oscillator. IEEE Trans. Circuits Syst. I,
reduction algorithms. Some of such algorithms have been 41(7):477-480.

SBA Controle &. Automação /Vol.7 no. l/Jan., Fev., Mar. e Abril 1996 61
 Aguirre, L. A. and Billings, S. A. (1994c). Validating iden- Billings, S. A., Korenberg, M . .1., anel Chen, S. (1988).
tified nonlinear models with chaotic dynamics. Int. J. Identification of nonlinear output afline systems using
Bifureation anel Chao,s, 4( 1):109-125. an orthogonalleast squares a.lgorithm. Int. J. Sy,ste111,s
Sei., 19(8):1559-1568.
Aguirre, L. A. and Billings, S. A. (1995a). Closed-Ioop
suppression of chaos in a nonlinear oscillator. Journa! Billings, S. A. anel Voon, W. S. F. (1984). Least squares
of Non!inear Seienee, 5(3):189-206. parameter estimation algorithms for nonlinear sys-
tems. Inl. J. 8y,sle1118 8ei., 15(6):601-615.
Aguirre, L. A. and Billings, S. A. (1995b). Dynami-
cal effects of overparametrization in nonlinear models. Billings, S. A. and Voon, W. S. F. (1987). Piecewise linear
Phy,sica D, 80(1,2):26-40. identification of non-linear systems. Inl. J. Contra!,
46( 1):215-235.
Aguirre, L. A. anel Billings, S. A. (1995c). lelentification
of moelels for chaotic systems from noisy elata: impli- Braiman, Y. and Goldhirsch, I. (1991). Taming chaotic
cation for performance anel nonlinear filtering. Phy,siea dynamics with weak periodic perturbations. Phy,s.
D, 85( 1,2 ):239-2 138. Rm LeU., 66(20):2545-2548.
Aguirre, L. A. anel Billings, S. A. (1995el). Improveel struc-
Breedon,.1. L. and Packard, N. H. (1992). Nonlinear anal-
ture selection for nonlinear moelels based on term C!us-
ysis of data sampleelnonuniformly in time. Phy,siea D,
tering. 1nt. J. Contro!,62(3):.569-587.
58:273-283.
Aguirre, L. A. and Billings, S. A. (1995e). Retrieving dy-
namica.l invariants from chaotic elata using NARMAX Broomhead, D. S., Huke, .1. P., and Muldoon, M. R.
models. In1. J. Bif1/,reation anel Chao,s, 5(2):449-474. (1992). Linear filters and nonlinear systems. J. R.
Stat. Soe. B, 54(2)::373-382.
Aguirre, L. A., Mendes, E. M., and Billings, S. A. (1995).
Smoothing data with local instahilities for the identi- Broomhead, D. S. anel King, G. P. (1986). Extracting
fication of chaotic systems. Int. J. Contro!, (in press). qualitative dynamics from experimental data. Phy,siea
D, 20:217-236.
Akaike, H. (1974). A new look at the statistical model
identification. IEEE Tran,s. Autom.a1. Contr., Broomhead, D. S. and Lowe, D. (1988). Multivariable
19( 6) :7l()-723. functional interpolation anel adaptive networks. Co 111-
pio; 8:I),st em.,s , 2:321-355.
Albano, A. M., Muench, .1., Schwartz, C., Mees, A. 1., and
Rapp, P. E. (1988). Singular-value decomposition anel Brown, R., Bryant, P., and Abarbanel, H. D. I. (1991).
the Grassberger-Procaccia algorithm. Phy,s. Rell. A, Computing the Lyapunov spectrum of a dynamical sys-
38(6 ):3017-3026. tem from an observeel time series. Phy,s. Rell. A,
4:3(6):2787-2806.
Albano, A. M., Passamante, A., and Farrell, Iv!. E. (1991).
Using higher-order correlations to define an embedding Brown, R., Rulkov, N. F., and Tracy, E. R. (1994). Model-
window. Phy,siea D, 54:85-97. ing and synchronizing chaotic systems from time-series
data. Phy,s. Rell. E, 49(5):3784-3800.
Aleksié, Z. (1991). Estimating the embedding dimensiono
Phy,siea D, 52:362-368. Buzug, T. and Pfister, G. (1992). Comparison of algo-
rithms calculating optimal embedding parameters for
Baake, E., Baake, M., Bock, H. G., and Briggs, K. M.
delay doordinates. Phy,sica D, 58:127-137.
(1992). Fitting ordinary differentia.l equations to
chaotic data. Phy,s. Rev. A, 45(8):5524-5529.
Buzug, T., Reimers, R., and Pfister, G. (1990). Optimal
Billings, S. A. (1980). Identification of nonlinear systems recollstruction of strange attractors from purely geo-
- a survey. IEE Proceeelig,s P1. D, 127(6):272-285. metricalarguments. Europhy,s. LeU., 13(7):605-610.

Billings, S. A. and Aguirre, L. A. (1995). Effects of the Casdagli, M. (1989). Nolllinear prediction of chaotic time
sampling time on the dynamics and ielentification of senes. Phy,sica D, 35:335-356.
nonlinear models. Int. J. Bifurcation anel Chao,s,
5(6):1541 - 1556. Casdagli, M. (1991). Chaos and deterministic ver,sl1,s
stochastic non-linear modelling. J. R. Stat. Soe. B,
Billings, S. A. and Chen, S. (1992). Neural networks and 54(2):303-328.
system ielentification. In VVarwick, K., Irwing, G. W.,
and Hunt, K . .1., editors, Neura! Netwok,s for Sy,stem,s Casdagli, M. (1992). A dynamical systems approach to
anel Contro!, chapter 9, pages 181-205. Peter Peregri- modeling input-output systems. In Casdagli, M. and
nus, London. Eubank, S., editors, Non!inear Moele!ing anel Foreea,st-
ing, pages 265-281. Aeldison Wesley, New York.
Billings, S. A., Chen, S., and Korenberg, M. .1. (1989).
Identification of MIMO nonlinear systems using a Casdagli, M., Eubank, S., Farmer,.1. D., and Gibson,.1.
forward-regression orthogonal estimator. In1. J. Con- (1991). State space reconstruction in the presence of
tra!,49(6):2157-2189. noise. Phy,sica. D, 51:52-98.

62 5BA Controle & Automação jVol.7 no. 1jJan., Fev., Mar. e Abril 1996
Casdagli, M., Jardins, D. D., Eubank, S., Farmer, J. D., Dent.on, T. A. anel Diamond, G. A. (1991). Can the ana-
Gibson, .1., Theiler, .1., and Hunt.er, N. (1992). Non- lytic t.echniques of nonlinear elynamics elistinguish peri-
linear modeling of chaot.ic time series: theory and ap- oelic, ranelom anel chaotic signals? Compt. ElOl. M ed.,
plications. In Kim, J. H. anel Stringer, .1., editors, Ap- 21( 4):243-264.
plied Chaos, chapter 15, pages 335-380. John "'hley &
Ditto, \tV. L. anel Pecora, L. M. (1993). Mastering chaos.
Sons., New York.
8cientific A merican, pages 62-68.
Cenys, _A. and Pyragas, K. (1988). Estimation of t.he
Ditto, W. L., Rauseo, S. N., anel Spano, M. L. (1990).
number of c1egrees of freedom from chaotic time series.
Experiment.al control of chaos. Ph:ljs. Re·v. Leit.,
Phys. Leit., 129A(4):227-230. 65(26 ):3211-:3214.
Chang, H. C. (1992). Chaotic mixing for heat transfer Dorning, J . .1., Decker . .1., anel P., H. J. (1992). Con-
enhancement. In lúm, J. H. and St.ringer, .1., editors, trolling the elynamics of chaotic convective ftows. In
Applied Chaos, chapter 6, pages 175-188. John Wiley Kim, J. H. anel St.ringer, .1., eelitors. Appl-ied Chaos,
& Sons, Inc. chapter 7, pages 189-206..John Wiley & Sons, Inc.
Chen, G. and Dong, X. (1993a). From chaos t.o order - Eckmann, .1. P. anel Ruelle, D. (1985). Ergoelic theory
Perspectives anel methoelologies in cont.rolling chaotic of chaos anel st.range attractors. Rev. M od. Phys.,
nonlinear elynamical systems. Int. J. Bifurcation and 57( 3):617-656.
Chaos,3(6):1363-1409.
Elsner, J. B. (1992). Preelicting time series using a neu-
Chen, G. anel Dong, X. (1993b). On feeelback cont.rol of ral net.work as a met.hoel of elistinguishing chaos from
chaotic continuous-time systems. IEEE Trans. Cir- noise. J. Phys. A: Math. Gen., 25:843-850.
ntits 8yst., 40(9):591-601.
Farmer, .1. D. anel Sielorowich, J. J. (1987). Preelicting
Chen, S., Billings, S. A., Cowan, C. F. N., anel Grant, chaot.ic time series. Ph:ljs. Rev. Leit., 59(8):845-848.
P. M. (1990). Practical ielentificat.ion of NARMAX Farmer, J. D. anel Sielorowich, J. J. (1 988a). Exploiting
moelels using radial basis functions. Int. J. Control, chaos to preelict the future anel reeluce noise. In Lee,
52(6):1327-1350. Y., eelitor, Evolution, Learning and Cognition. W'orlel
Chen, S., Billings, S. A., anel Luo, W. (1989). Orthogonal Scientific, Singapore.
least squares methoels anel their application to nonlin- Farmer, J. D. anel Sidorowich, J ..1. (1988b). Preelieting
ear system ielentification. Int. J. Control, 50(5): 1873- chaotic elynamics. In Kelso, J. A. S., Manelell, A. .1.,
1896. and Shlesinger, M. F., eelitors, Dynamic paiterns in
complex systems, pages 265-292. Worlel Scientific, Sin-
Cheng, B. anel Tong, H. (1992). On consistent nonpara-
gapore.
metric oreler eletermination anel chaos. J. R. 8tat. 80c.
B, 54(2):427-449. Farmer, J. D. anel Sielorowich,.1. J. (1991). Optimal shael-
owing anel noise reeluction. Physica D, 47 :373-392.
Chennaoui, A., Pawelzik, K., Liebert, ·VV., Schuster, G. H.,
anel Pfister, G. (1990). Attractor reconstruction from Fraser, A. M. (1989). Information anel entropy in
filtereel chaotic time series. Phys. Rev. A,41(8):4151- strange att.ractors. IEEE Trans. Informatian Theary,
4159. 3.5(2):245-262.

CIma, L. O., Itoh, M., Kocarev, L., anel Eckert, K. (1993). Fronzoni, L., Gioconelo, M., anel Pettini, M. (1991). Ex-
Chaos synchronization in Chua's circuit. J. Circuits perimental evielence of suppression of chaos by res-
8yst. Comput., :3(1):93-108. onant parametric perturbations. Phys. Rev. A,
43( 12) :648:3-6487.
Cremers, .1. anel Hübler, A. (1987). Construction of elif-
ferential equations from experimental elata. Z. N atur- Garfinkel, A., Spano, rvI. L., Ditto, W. L., anel Weiss, J. N.
forsch, 42a:797-802.
(1992). Controlling cardiac chaos. 8cience, 257:1230-
1235.
Crutchfielel, J. P. anel McNamara, B. S. (1987). Equations
Genesio, R. anel Tesi, A. (1991). Chaos preeliction innon-
of motion from a elata series. Complex 8ystems, 1:417-
linear feeelback systems. IEE Prac. Pt. D,138(4):313-
452.
320.
Cuomo, K. M., Oppenheim, A. V., anel Strogatz, S. H. Genesio, R. anel Tesi, A. (1992). A harmonic balance
(1993). Robustness anel signal recovery in a synchro- methoels for the analysis of chaotic elynamics in non-
nizeel chaotic system. Int. J. Bifurcation and Chaos, linear syst.ems. Automatica,28(3):531-548.
3(6):1629-1638.
Genesio, R. anel Tesi, A. (1993). Distortion cont.rol of
Davies, M. E. (1992). An iterateel function approximation chaotic systems: the Chua's circuito J. Circuits 8yst.
in shaelowing time series. Phys. Leit., 169 A:251-258. Cornput.,3(1):151-171.
Deelieu, H. anel Ogorzalek, M. (1994). Controlling chaos in Gibson, J. F., Farmer, J. D., Caselagli, M., anel Eubank, S.
chua's circuito via sampleel inputs. Int. J. Bifurcatioll (1992). An aelnalytic approach t.o praetical st.ate space
and Chaos, 4(2):447-455. reconstruction. Physica D, 57:1.

SBA Controle & Automação jVol.7 no. 1jJan., Fev., Mar. e Abril 1996 63
 Giona, M., Lentini, F., anel Cimagalli, V. (1991). Func- Hunter, N. F. (1992). Application ofnonlinear time-series
tional reconstruction anel local preeliction of chaotic moelels to elriven systems. In C aselagli, M. anel Eu-
time series. Phys. Rev. A,44(6):3496-3502. bank, S., eelitors, Nonlinear Moeleling anel Forecasting,
pages 467-491. Aelelison Wesley, New York.
Glass, L., Golelberger, A., Courtemanche, M., anel Shrier,
A. (1987). Nonlillear elynamics, cimos anel complex Jackson, E. A. (1990). The entrainment anel migration
careliac arrythmias. Pr'OC. R.. Soe. Lonel. A, 41:3:9-26. controls of multiple-attractor systems. Phys. Léu.,
151A(9):478-484.
Glass, L. anel Mackey, M. (1988). F'rom Clocb to Chaos
- The Rhythms of Life. Princeton University Press, Jackson. E. A. anel Grosu, I. (1994). An op~li-plus­
Prillceton. closeel-Ioop (OPCL) control of complex elynamic sys-
tem. Technical Report CCSR-94-7, University of Illi-
Gleick, .1. (1987). Chaos; making a ne'W science. Abacus,
nois, Center for Complex Systems Research.
Lonelon.

Golelberger, A., Ringney, D., anel West, B. (19~)0). Chaos Johansen, T. A. anel Foss, B. A. (1993). Constructing
anel fractaIs in human physiology. Scientific "4merican, NARMAX moelels using ARMAX moelels. Int. J.
pages 34-41. Control,58(5):1125-115:3.

Golden, M. P. anel Yelstie, B. E. (1992). Smal! amplituele Kaeltke, .l. B., Brush, .1., anel Holzfuss, .l. (1993). Global
chaos anel ergoelicity in aelaptive contro!. A utoTl7atica, elynamical equations anel Lyapunov exponents fro111
28(1):11-25. noisy chaotic time series. Int. J. Bifurcation anel
Chaos, 3(3):607-616.
Gooijer, J. G., Abraham, B., Goulel, A., anel Robinson,
L. (1985). Methoels for eletermining the oreler of an Kapitaniak, T. (1991). The loss of chaos in a
autoregressive-moving average process: a survey. Int. quasiperioelically-forceel nonlinear oscil!ator. 1nt. .J.
Statistical Re'view, 53( 3 ):301-329. Bif1lrcation anel Chaos, 1(2):357-362.

Gottwalel, .1. A., Virgin, L. N., anel Dowel!, E. H. (1992). Kapitaniak, T., Kocarev, L..1., anel CIma, L. O. (1993).
Experimentalmimicry of Duffing's equation . .J. Sounel Control!ing chaos without feeelback anel control signals.
Vibr., 158(3):447-467. Int . .J. Bifurcation anel Chaos, 3(2):459-468.

Gouesbet, G. anel Maquet, .1. (1992). Construction ofphe- Kashyap, R. L. (1977). A bayesian comparison of elifferent
nomenological moelels from numerical scalar time se- classes of elynamica moelels using empirical data. IEEE
ries. Physica D, 58:202-215. Trans. Automat. Contr., 22(5):715-727.
Grassberger, P., Hegger, R., Kantz, H., anel Schaffrath, C. Kember, G. anel Fowler, A. C. (1993). A correlation func-
(1993). On noise reeluction methoels for chaotic elata. tion for choosing time elelays in phase portrait recon-
Chaos, 3(2):127-141. structions. Phys. LeU., 179A(2):72-80.
Grassberger, P., Schreiber, .1., anel Schaffrath, C. (1991). Kennel, M. B., Brown, R., anel Abarbanel, H. D. I. (1992).
Nonlinear time sequence analysis. Int. .J. Bifurcation Determining embeelding elimensions for phase-space re-
anel Chaos, 1(3) :521 ~547. construction using geometrical construction. Phys.
Guckenheimer, J. (1982). Noise in chaotic systems. N a- Rev. A, 45(6):3403-3411.
ture, 298:358-361.
Kennel, M. B. anel Isabelle, S. (1992). Methoel to elis-
Haber, R. anel Unbehauen, H. (1990). Structure ielentifi- tinguish possible chaos from coloureel noise anel to
cation of nonlinear elynamic systems - A survey on determine embeeleling parameters. Phys. Rev. A,
inputjoutput approaches. A utomatica, 26(4) :651-677. 46( 6):3111-3118.

Hammel, S. H. (1990). A noise reeluction methoel for King, G. P. anel Stewart, I. (1992). Phase space recon-
chaotic systems. Phys. LeU., 148 A(8,9):421-428. struction for symmetric elynamical systems. Physica
D, 58:216-228.
Hartley, T. T. anel Mossayebi, F. (1993). ContraI of chua's
circuito J. Circuits Syst. Comput., ;3(1):173-194. Kocarev, L., Shang, A., anel CIma, L. O. (1993). Tran-
sitions in elynamical regimes by elriving: A unifieel
Hénon, M. (1976). A two-elimensionalmap with a strange
methoel of control anel synchronization of cimos. In1.
attractor. Commun. M ath. Phys., 50:69-77.
.J. Bifurca.tion anel Chaos, 3(2):479-483.
Holzfuss, J. anel Kaeltke, J. (1993). Global nonlinear noise
Korenberg, M. .l., Billings, S. A., Liu, Y. P., anel McIl-
reeluction using raelial basis functions. Int. J. Bifurca-
roy, P . .l. (1988). Orthogonal parameter estimation
tion anel Chaos, 3(3):589-596.
algorithm for nonlinear stochastic systems. Int. J.
Hunt, E. R. (1991). Stabilizing high-performance orbits Control, 48(1):193-210.
in a chaotic system: the elioele resonator. Phys. Rev.
LeU., 67(15):1953-1955. Korenberg, M..1. anel Paarmann, L. D. (1991). Orthogo-
nal approaches to time-series analysis anel system ielen-
Hunt, E. R. anel Johnson, G. (1993). Keeping chaos at tification. IEEE S'ignal Processing Magazine, 8(3) :29-
bay. IEEE Spectr1l.1n, pages 31-36. 43.

64 SBA Controle & Automação jVol.7 no. 1jJan., Fev., Mar. e Abril 1996
 Kostelich, E ..1. and Yorke, .I. A. (1988). Noise reduction Nicolis,.l. S. (1984). The role of chaos in reliable informa-
in dynamicalsystems. Phys. Rev. A, 38(3):1649-1652. tion processing. J. Franklin Institue, 317(5):289-307.

Kostelich, E ..1. and Yorke, .I. A. (1990). Noise reduction: Nitsche, G. anel Dressler, U. (H)92). Controlling chaotic
Finding the simplest elynamical system consistent with elynamical systems using time elelay coorelinates. Ph:lJs-
the data. Physica D, 41:183-196. iea D, 58:153-164.
Leontaritis, I..l. anel Billings, S. A. (1985a). Input-output Ogorzalek, M . .I. (1993). Taming chaos: Part I - synchro-
parametric models for nonlinear systems part I: cleter- nization. IEEE Trans. CirC'lIits Syst. I, 40( 10):693-
,'I !~ninistic nonlinear systems. Int. J. Contrai, 41(2) :303- 699.
328.
ou, E., Grebogi, C., anel )'orke, .I. A. (1990). Controlling
Leontaritis, I. .I. anel Billings, S. A. (1985b). Input- chaos. Phys. Re11. Leit.,64(11):l1DG-1199.
output parametric moelels for nonlinear systems part
11: stochastic nonlinear systems. Int. J. Contrai, OUino, .J. M. (1992). New applications of chaos in chem-
41(2):329-344. ical engineering: intuition versus preeliction. In Kim,
.I. H. anel Stringer, .I., editors, Applied Chaos, chap-
Liebert, W. anel Schuster, H. G. (1989). Proper choice of ter 5, pages 14:3-174. .lohn Wiley & Sons, Inc.
the time delay for the analysis of chaotic series. Phys.
Leit., 142A(2,3):107-111. Ozaki, T. (1993). A locallinearization approach to non-
linear filtering. Int. J. Control,57(1):75-96.
Lima, R. anel Pettini, M. (1990). Suppression of chaos
by resonant parametric perturbatios. Phys. Rev. A, Packard, N. H., Crutchfielel, .I. P., Farmer, .I. D., and
41(2):726-33. Shaw, R. S. (1980). Geometry from a time series.
Phys. Rev. Leit.,45(9):712-716.
Linsay, P. S. (1991). An efficient methoel of forecasting
chaotic time series using linear regression. Phys. Leit., Parlitz, U., CIma., L. O., Kocarev, L., Halle, K. S., and
153A(6,7):35:3-356. Shang, A. (1992). Transmission of eligital signals by
chaotic synchronization. Int. J. Bif. Choos, 2(4):973-
Liu, Y., Barbosa, L. C., anel Rios-Leite, .I. R. (1994). Sup-
977.
pression of chaos in a CO 2 laser by feedback. Phys.
Leit., 19:3A:259-262. Pecora, L. M. Carroll, T. L. (1990). Synchronization in
chaotic systems. Phys. Rev. Leit., 64(8):821-824.
Lorenz, E. (1963). Deterministic nonperioelic flow. J. A t-
mos. Sei., 20:282-293. Pool, R. (1989). Is it healthy to be chaotic? Seienee,
243:604-607.
Mackey, M. C. anel Glass, L. (1977). Oscillation anel chaos
in physiological control systems. Seienee, 197:287-289. Principe, .I. C., Rathie, A., and Kuo, .I. M. (1992). Pre-
eliction of chaotic time series with neuralnetworks anel
Mahfouz, I. A. and Baelrakhan, F. (1990a). Chaotic be-
haviour of some piecewise-linear systems part I: Sys- the issue of elynamic moeleling. Int. J. Bif'll.reation and
Chaos, 2(4):989-996.
tems with set up spring 01' with unsymetric elasticity.
J. SO'llnd Vibr., 143(2):2.5.5-288. Pyragas, K. (1992). Continuous control of chaos by self-
Mahfouz, I. A. and Baelrakhan, F. (1990b). Chaotic be- controlling feeelback. Phys. Leit., 170A:421-428.
haviour of some piecewise-linear systems part 11: Sys- Romeiras, F . .I., Grebogi, C., Ott, E., anel Dayawansa,
tems with clearance. J. Sound Vibr., 143(2):289-328. W. P. (1992). Controlling chaotic elynamical systems.
Mareels, I. M. Y. and Bitmeael, R. R. (1986). Non-linear Physiea D, 58:165-192.
elynamics in adaptive control: cha.otic and periodic sta-
Rosenstein, M. T., Collins, .I . .I., anel De Luca, C . .I.
bilization. A'I1tomotiea, 22(6):641-655.
(1994). Reconstruction expansion as a geometry-baseel
Mareels, I. M. Y. and Bitmeacl, R. R. (1988). Non-linear framework for choosing proper delay times. Physica D,
dynamics in aelaptive control: perioelic anel chaotic sta- 73:82-98.
bilization - 11. Analysis. Automotiea, 24(4) :485-497.
Roy, R., Murphy, .Jr., T. W., Maier, T. D., and Gills,
Marteau, P. F. and Abarbanel, H. D. I. (1991). Noise Z. (1992). Dynamical control of a chaotic laser: ex-
reduction in chaotic times series using scaleel proba- perimentaI stabilization of a globally coupled system.
bilistic methoels. J. Nonlinear Sei., 1:313-343. Phys. Re·v. Leit.,68(9):1259-1262.

Martinerie, .I. M., Albano, A. M., Mees, A. 1., anel Rapp, Sauer, T., Yorke, .I. A., and Caselagli, M. (1991). Em-
P. E. (1992). Mutual information, strange attractors, bedology. J. of Stat-istieal Physies, 65(3/4):579-616.
anel the optimal estimation of dimensiono Phys. Rev.
Schreiber, T. and Grassberger, P. (1991). A simpIe noise-
A, 45(10):7058-7064.
reduction methoel for real data. Phys. Leit., 160
Mees, A. I. (1993). Parsimonious dynamical reconstruc- A(5):411-418.
tion. Int. J. Bifureation and Chaos, 3(3):669-675.
Schreier, R. (1994). On the use of chaos to reduce ielle-
Mitschke, F. (1990). Acausal filters for chaotic signals. channel tones in eleIta-sigma modulators. IEEE Trans.
Phys. Rev. A, 41(2):1169-1171. CirC'l/its Syst. 1,41(8):539-547.

SBA Controle & Automação IVo!.7 no. 1/Jan., Fev., Mar. e Abril 1996 65
 Shinbrot, T., Ott, E., Grebogi, C., anel Yorke,.l. A. (1990).
Using chaos to elirect trajectories to targets. PII.:/j.s.
Rev. Leit., 65(26) :3215-3218.

Sinha, S., Ramaswamy, R., anel Rao, .l. S. (1990). Aelap-

tive control in nonlinear elynamics. Phy.siea D, 43:118-
128.

Soelerstrom, T. anel Stoica, P. (1989). Sy.stem Identifiea-

t-ion. Prentice Hall, Lonelon.

Spano, M. L., Ditto, W. L., anel Rauseo, S. N. (1991).

Exploitation of chaos for active control: an experimento
J. of Intel!. 111ater. Slj.st. and Struet., 2:482-49:3.

Stokbro, L. anel Umberger, D. K. (1992). Forecasting with

weighteel maps. In Caselagli, M. anel Eubank, S., eeli-
tors, NonlinearModeling and Foreea.st-lng, pages 73-93.
Aelelison Wesley, New York.

Takens, F. (1980). Detecting strange attractors in tur-

bulence. In Ranel, D. A. anel Young, L. S., eelitors,
Dynamieal .s:lj.stems and turbulenef, Leeture N ote.s in
Mathemat-ie.s, 1'01. 898, pages 366-381. Springer Ver-
lag, Berlin.

Ushio, T. anel Hsu, C. S. (198T). Chaotic rouneling error

in digital control systems. IEEE Tran.s. Cireuit.s Sy.st.,
:34(2): 13:3-139.
Vassiliadis, D. (1993). Parametric aelaptive control of an
NMR laser moelel. In i. 1. Bifureation and Chao.s,
3(3):T93-T96.

'Wayland, R., Pickett, D., Bromley, D., anel Passamante,

A. (1993). Measuring the preelictability of noisy re-
current time series. Int. J. Bifureation and Chao.s,
3(3):797-802.

Whaba, G. (1992). Multivariabe function anel operator es-

timation, baseel on smoothing splines anel reproelucing
kernels. In Casdagli, M. anel Eubank, S., eelitors, Non-
linear M odeling and Foreea.sting, pages 95-112. Aeleli-
son Wesley, New York:

Wu, C. W. anel CIma, L. O. (1993). A simple way to

synchronize chaotic systems with applications to se-
cure communication systems. Int. J. Bifureation and
Chao.s, 3(6):1619-1627.

Yau, C. H., Bajaj, A. K., anel Nwokah, O. D. I. (1993).

Robust control of a chaotic vibratory system. Int. 1.
Bifurcation and Chao.s, 3(4): 1075-1081.

66 SBA Controle & Automação jVol.7 no. 1jJan., Fev., Mar. e Abril 1996