ISSN (Print) : 0974-6846

Indian Journal of Science and Technology, Vol 11(21), DOI: 10.17485/ijst/2018/v11i21/123003, June 2018 ISSN (Online) : 0974-5645

A Note on the Exponents of Lyapunov for a System

with Unknown Dynamic
Pedro Pablo Cárdenas Alzate1*, José Gerardo Cardona Toro2 and Luz María Rojas Duque3
Departamento de Matemáticas and Research Group GEDNOL, Universidad Tecnológica de Pereira,
1

Pereira, Colombia; ppablo@utp.edu.co

2
Departamento de Matemáticas, Universidad Tecnológica de Pereira, Pereira, Colombia; gerardo7@utp.edu.co
3
Fundación Universitaria del Área Andina, Pereira, Colombia; lmrojas@areandina.edu.co

Abstract
Objective: To determine the exponents of Lyapunov for a system with unknown dynamics. Methods: The determination
of chaos presence in any system have a great importance to right analysis, thus there are indicators that permit quantifying
this phenomenon, one of them and perhaps more relevant are the Lyapunov exponents. That is why this paper presents
a methodology to calculate these quantifiers from the system dynamically. (differentials equations). Findings: In many
occasions, we know the system dynamics, but we do not have necessary instruments to measure all the present states,
thus we need to reconstructthe phase space from the times series scalar that we have. If one wants to analyze the original
dynamics of the dynamic system from the said series, the attractor in the multivariable space is necessary, which is known
as reconstruction of the phase space, originally proposed by Ruelle. This technique converts a scalar series x (t) in a vector
v of dimension m (Embedding dimension) using a delay time. Application: As mentioned, in various physical situations
there is no measurement of all the variables of the system’s state, therefore, it is necessary to reconstruct the dynamics
of the system from the available measurements. Assuming that only the voltage measurement of capacitor one (Vc1) is
available from the Chua circuit.

Keywords: Chua Circuit, Embedment Dimension, Exponents of Lyapunov, Multivariable Dynamic System,
Phase Space

1. Introduction The chaos has been widely studied by mathemati-

cians and engineers since its discovery by Edward Lorenz
Dynamic systems can be described by a set of ordinary in 1970s. The main characteristic of these systems is the
differential equations, which when going into detail extreme sensitivity to the initial conditions, that is, for
in terms of their modeling increases their complexity, nearby trajectories initially their distance in the state
becoming non-linear. The analysis of linear dynamic sys- space diagram will increase exponentially, being quanti-
tems in phase space has reduced behaviors (fixed points, fied by the exponents of Lyapunov1.
periodic or divergent solutions); in terms of non-linear If you want to determine the exponents of Lyapunov
systems they contain a variety of scenarios, stationary or for a system with unknown dynamics, you must recon-
convergent solutions (fixed points), periodic solutions struct the phase space that captures the original attractor
(limit cycles), quasi-periodic (attractor bull) or chaotic of the system. The most familiar method for the recon-
solutions (strange attractors). struction of phase space from a scalar time series is based

*Author for correspondence

 A Note on the Exponents of Lyapunov for a System with Unknown Dynamic

on the determination of a delay time and an embedding

Initially defined by the identity matrix In The variational
dimension2.
expressions give rise to the tangent space, defined as:

2. Estimation of the Lyapunov φ(x,t)=J(x,t)φ (x,t),(5)

Exponents
Where J(x,t) represents the Jacobian of F(x,t). The system
2.1 Phase Space of equation of state together with the variational expres-
When observing a real multivariable dynamic system, it is sions are integrated from x 0 with φ (x 0 )=In for a time T
normal that all it’s variables cannot be measured simulta- and thus obtain the divergent vectors transformed by the
neously, in the worst case there is only one measurement application of the tangent plane, that is, for the first expo-
(time series). nent

 δ1x(1) =φ (x,t)u1(0)
v(t)=[v 1 (t)v 2 (t)...v n (t)]T


v(t)=[x(t)x(t+τ )...x(t+(m-1)τ )]T ∈ R m (1) δ1x(1)
Where u1(0) = and the superscript represents the
δ1x(1)
where T is the transpose. current iteration of the calculation. By repeating this pro-
The proper choice of delay time and embedment cedure of integration and normalization K times the i-th
dimension is very important for the correct reconstruc- exponent of Lyapunov can be written in the following
tion of the attractor. There are methods to determine way:
the delay time, such as the false neighboring neighbors.
On the other hand, the dimension of embedding of a 1 K
n-dimensional system is obtained according to the for-
λi = lim
K →∞ KT
∑
k =1
ln δ ix(k ) (6)

mulation of Takens3.
In each iteration made the direction of the vectors
m2 n+1 (2) {δ ix (t), δ 2 x (t),...,δ nx (t)} it must be aligned with the expan-
sion direction of the system dynamics, for this reason the
2.2 Lyapunov Exponents Gram-Schmidt method must be used, as shown below
For an n-dimensional system described by
n =δ nx -〈δ nx ,u1 〉 u1 -...〈δ nx ,u n-1 〉 u n-1
v (k) (k) (k) (k) (k) (k) (k) (k)

x(t) = F(x,t)
v (k)
n (7)
u(k)
n =
the i-th exponent of Lyapunov is defined λi depending v (k)
n

on the variations of the trajectories from a starting point

x 0 in t 0 , these variations can be seen as the degeneration Once the exponents of Lyapunov have been calculated,
of the radius of a hyper sphere δ 0 (t) in a hyper-ellipsoid the maximum number of them must be identified, since it
of radio δ i (t) is the one in charge of determining the presence of chaos
or not in said dynamic system. For this, it is clear that
δ i (t)=δ 0 eλ (t-t ) (3)
i 0
if you have λmax > 0 the initially close trajectories in the
phase space will tend to separate, that is, the system pres-
The classic procedure for the estimation of the exponents ents a high sensitivity to the initial conditions.
of Lyapunov was proposed by Wolf in 19854, which con-
sists of evaluating the local divergences from the applica-
tion of the tangent plane associated with the system of
3. Application to Chua’s Circuit
state equations on n orthogonal vectors Since its discovery in 1983, the Chua circuit has served as
the basis for the study of the chaos in electronic systems5.
{δ ix (t), δ 2 x (t),...,δ nx (t)} (4) This circuit is composed of two capacitors (C1 and C2), an

2 Vol 11 (21) | June 2018 | www.indjst.org Indian Journal of Science and Technology
 Pedro Pablo Cárdenas Alzate, José Gerardo Cardona Toro and Luz María Rojas Duque

inductance (L), a resistance (R) and a non-linear resistance 1

(Rn), in Figure 1 the canonical circuit can be observed. f (x1 )= m1 x1 + (m0 − m1 )(| x1 + E | − | x1 − E |) (9)
2
Here, E is a parameter dependent on the polarization volt-
age of the circuit, when implemented with OPAMs. The
mathematical model of the Chua circuit can be normal-
ized according to the proposal by Matsumoto5, for greater
ease in the simulation

x α ( y − h(x ))
=
y = x − y + z (10)
z = −β y

Figure 1. Chua’s circuit. α, β being system parameters with typical values are α = 9,
β = 14.286. Here, h(x) it is represented in the form:
Solving the circuit by state variables, the system is
described by differential equations: 1
h(x )= m1 x + (m0 − m1 )(| x + 1 | − | x − 1 |) (11)
1 2
x1 = (x 2 − x1 ) − f(x1 )
C1 R
The values of the constants m0 and m1 are assumed
for this analysis with -1/7 and 2/7 respectively.
1
x2 = (x1 − x 2 ) + x 3 By integrating the system (9) using the Runge-Kutta45
C2R algorithm, the time series is obtained (Figure 2).
Graphing the phase diagram between Vc1 and Vc2
-1 shows that the resulting attractor is a limit cycle.
x2 = x 2 (8)
L By varying the control parameter α=9 (Figure 3), a
Where x1 =V= V= iL= large change in the dynamics of the system is observed,
C1 , x 2 C 2 , x3 f (xi ) f (VC1 ) is the func-
tion that characterizes non-linear resistance, and is defined becoming a chaotic system (strange attractor) as seen in
as follows? the phase diagram of Figure 4.

Figure 2. Time series Vc1, Vc2 and iL with α=7.5 .

Vol 11 (21) | June 2018 | www.indjst.org Indian Journal of Science and Technology 3
 A Note on the Exponents of Lyapunov for a System with Unknown Dynamic

Figure 3. Time series with α=9.

Figure 4. Strange attractor for α=9.

3.1. Reconstruction of the Phase Space of Taking as dimension of embedment m=3 and a delay time of
Chua’s Circuit τ =5s , the attractor of the system seen in Figure 6 is obtained.
As mentioned, in various physical situations there is no
3.2. Determination of Lyapunov Exponents
measurement of all the variables of the system’s state,
therefore, it is necessary to reconstruct the dynamics of for the Chua Circuit
the system from the available measurements. Assuming For the calculation of the Lyapunov exponents6 of the
that only the voltage measurement of capacitor one (Vc1) Chua circuit, we start from the system (9), from which
is available from the Chua circuit, is (Figure 5): the Jacobian matrix is ​​identified:

4 Vol 11 (21) | June 2018 | www.indjst.org Indian Journal of Science and Technology
 Pedro Pablo Cárdenas Alzate, José Gerardo Cardona Toro and Luz María Rojas Duque

Figure 5. Measurement of Vc1.

Figure 6. Comparison reconstructed vs. Original.

 dh  Which is the basis to start with the algorithm proposed

 α dx α 0 by Wolf4, for the determination of the exponents of
  Lyapunov. The simulation (Figure 7) was performed with
J(x,t) =  1 -1 1 
 0 T=200, α =9 and initial conditions [0.15264 - 0.02281
-β 0 
  0.38127].
 

Vol 11 (21) | June 2018 | www.indjst.org Indian Journal of Science and Technology 5
 A Note on the Exponents of Lyapunov for a System with Unknown Dynamic

Figure 7. Spectrum of Lyapunov exponents for α=9.

Figure 8. Limit cycle attractor α=7.5 .

Figure 9. Divergence for α=10.

6 Vol 11 (21) | June 2018 | www.indjst.org Indian Journal of Science and Technology
 Pedro Pablo Cárdenas Alzate, José Gerardo Cardona Toro and Luz María Rojas Duque

The values ​​obtained are shown in Table 1. It is observed that that is why the phase space must be reconstructed
the maximum exponent of Lyapunov = λmax 1.920399 > 0 , when this dynamic is not known.
therefore the system will behave chaotic and its future
states cannot be determined from the initial states. From
the results of Figures 8-9, 4 a high dependence of the
5. Acknowledgements
dynamics is observed with respect to the parameter α. We would like to thank the referee for his valuable sug-
gestions that improved the presentation of this paper and
Table 1. Exponents of Lyapunov for α=9 which shows our gratitude to the Department of Mathematics of the
that the phase diagram is a strange attractor Universidad Tecnológica de Pereira (Colombia) and the
α λ1 λ2 λ3 group GEDNOL.
9 1.920399 -0.686016 -0.909888
6. References
4. Conclusions 1. Strogatz SH, Fox R. Nonlinear dynamics and chaos: With
applications to physics, biology, chemistry and engineering.
The characterization of the dynamics of any system is Physics Today. 1995; 48(3):48–93. Crossref
determined by exponents of Lyapunov, which, depend- 2. Eckmann JP, Ruelle D. Ergodic theory of chaos and strange
ing on their value, allow a qualitative analysis, estab- attractors. Reviews of Modern Physics. 1985; 57:1–617.
lishing that dynamic is converging to a fixed point of Crossref Crossref
the state space, or if it is periodic (limit cycle and bull) 3. Takens F. Detecting Strange Attractors in Turbulence.
or simply if the future states cannot be determined Dynamical Systems of Turbulence. 1981; p.366–81. Crossref
from the initial states (strange chaotic-attractor sys- 4. Wolf A, Swift J, Swinney H. Determining Lyapunov expo-
tem). Therefore, the constructive form of the circuit, nents from a time series. Physica D: Nonlinear Phenomena.
for values between 7 and 11 there is a highly variable 1985; 16 (3):285–317. Crossref
5. Elwakil A, Kennedy M. Generic realizations of Chua’s cir-
spectrum of the exponents of Lyapunov as it passes
cuit. International Journal of Bifurcation and Chaos. 2000;
from one moment to operate stably under a limit cycle
10(8):1981–5. Crossref Crossref
to be divergent. Finally, based on the knowledge of the 6. Matsumoto T, Chua L, Komuro M. The double scroll.
dynamics of a system, a control law can be proposed to IEEE Transactions on Circuits and Systems CAS-32. 1985;
stabilize the system around the desired operating point (8):798–818.

Vol 11 (21) | June 2018 | www.indjst.org Indian Journal of Science and Technology 7