Secure Digital Signal Transmission by Multistep Parameter

Modulation and Alternative Driving of Transmitter Variables

P. Palaniyandi and M. Lakshmanan

Centre for Nonlinear Dynamics, Department of Physics

arXiv:nlin/0011005v1 [nlin.CD] 3 Nov 2000

Bharathidasan University, Tiruchirapalli - 620 024, India.

(November 1, 2018)

Abstract

The idea of secure communication of digital signals via chaos synchro-

nization has been plagued by the possibility of attractor reconstruction by

eavesdroppers as pointed out by Pérez and Cerdeira. In this Letter, we wish

to present a very simple mechanism by which this problem can be overcome,

wherein the signal is transmitted via a multistep parameter modulation com-

bined with alternative driving of different transmitter variables, which makes

the attractor reconstruction impossible. The method is illustrated by means

of the Lorenz system and Chua’s circuit as examples.

Typeset using REVTEX

1
The work of Pecora and Carroll [1990,1991] on synchronization of chaotic signals has
suggested the possibility of secure communication using chaos synchronization. They have
shown that two chaotic systems can be synchronized when they are linked by a common

signal (driving signal) provided the Lyapunov exponents of the subsystem are all negative.
Following this suggestion a number of methods have been proposed for secure communication
[Lakshmanan & Murali, 1996; Hayes et al., 1993; Cuomo & Oppenheim, 1993; Murali &
Lakshmanan, 1993a,1993b; Kocarev & Parlitz, 1995; Zhang et al., 1998] with the help of

chaotic signals. Cuomo and Oppenheim [1993] have shown that a small difference between
the corresponding parameters in the drive and response systems will cause synchronization
frustration between the transmitter and receiver variables. Using this property, they have
suggested a very simple method for digital signal transmission and illustrated it with the

Lorenz system. However, Pérez and Cerdeira [1995] have pointed out that it is possible to
reconstruct the message by an eavesdropper from a simple return map formed by the extrema
of the modulated driving signal (Here, the driving signal is used as a carrier signal in the
digital signal transmission. In our method the digital message is imposed on this carrier

signal through parameter variation. So, we call the driving signal bearing the digital message
as modulated driving signal and the modulation as parameter modulation.), even without
synchronization with the receiver system. To overcome the above problem several methods
have been suggested by many authors recently [Murali & Lakshmanan, 1998; Mensour &
Longtin, 1998; Minai & Pandian, 1998]: Private communication using compound chaotic

signal technique, using delay-differential equations technique, communication through noise

and so on. However, most of the above methods are very complicated and are often difficult
to implement. In this Letter, we describe a simple method by which the digital signal can be
highly masked with the chaotic signal by using a multistep parameter modulation and show

how it can be further complicated by introducing two different drive signals alternatively
instead of a single drive signal. As a result, reconstruction of the message becomes virtually
impossible and this provides further development in the technique of secure communication
using chaotic signals.

2
In order to present our method, we make use of the dynamical system used by Cuomo
and Oppenheim [1993] namely the Lorenz system. The associated evolution equations at
the transmitter end are
dxs
= σ(ys − xs ),
dt
dys
= rxs − ys − xs zs , (1)
dt
dzs
= xs ys − bzs ,
dt
while at the receiver end, corresponding to an x-drive, they read as
dxr
= σ(yr − xr ),
dt
dyr
= rxs − yr − xs zr , (2)
dt
dzr
= xs yr − bzr ,
dt
where σ = 16.0, r = 45.6 and the parameter b chosen for the purpose of modulation can have
either the value 4.0 or 4.4. The numerical value of the parameter b at the receiver end is kept
at a constant value 4.0, while in the transmitter circuit it is changed between two values,

namely 4.0 and 4.4, when the digital signal is zero and one, respectively. The transmitted
digital message is reconstructed by using the synchronization error power (xr − xs )2 . It is
negligible when the transmitter and the receiver systems are synchronized, while it has some
finite value when they are not synchronized. Thus the receiver will synchronize with the
transmitter when the parameter b in the latter takes a value 4.0, while asynchronization

takes place when b = 4.4 at the transmitter. It may be noted that instead of the x-drive at
the receiver end one can as well use the y-drive in which case the eq.(2) will be modified as
dxr
= σ(ys − xr ),
dt
dyr
= rxr − yr − xr zr , (3)
dt
dzr
= xr ys − bzr .
dt
In the work of Pérez and Cerdeira [1995], it has been shown that even without any
receiver circuit, one can unmask the message from the modulated drive signal. In this

3
reference, the maxima Xm and minima Ym are collected from the modulated driving signal
and the new varibles Am = (Xm + Ym )/2 and Bm = Xm − Ym are defined. The plot of
the return map for Am Vs Bm is shown in fig. 1. Note that there are 3 segments in the

attractor each one of which is further split into two strips (one corresponding to the high
state and the other corresponding to low state of the digital message). It is obvious that
the split in the attractor is due to the change of the parameter b at the sender end between
the values 4.0 and 4.4. From the return map one can easily unmask the message by finding

the strip on which the point(Am , Bm ) falls in each segment of the attractor. We can call
the above procedure as a single-step parameter modulation, since b can have only one value
for one state of the message (that is, b can take only a single value 4.4 for any high state
in the digital message and only 4.0 for the low state digital signal at the transmitter). The

crux of the problem is then that from the nature of the chaotic attractor and the return
map of the Lorenz system, one can have only eight possibilities of message reconstruction
from which one can easily discover the actual message. Then, one possible way to overcome
such a reconstruction is to increase the number of possible ways by which message can be

reconstructed to an unmanageable level and thereby eliminating the possibility of identifying

the correct message.
In our work, we have effected two changes in the mode of transmission of the digital
message by the parameter modulation in order to complicate the attractor in the return
map. Firstly, in the sender part, instead of using a single-step parameter modulation, we

use a multi-step parameter modulation, where for one state of the digital message (either 1 or
0), the parameter b in the sender equation is assigned to have any one of the ’n’ preassigned
values (n > 1 and sufficiently large), while at the receiver part we use n receiver subsystems
with different parameters. We call the number n as the step of the modulation. Secondly,

we will use both x and y signals to drive the receiver subsystems alternatively during the
transmission as indicated in eqs.(2) and (3).
The block diagram of the transmitter and the receiver with modulation step n is shown
in fig. 2, where R1 ,R2 ,...Rn are the subsystems of the receiver. Each one of the Ri′ s is

4
assigned a specific value among the n chosen parameter values which are used to impose the
high state of the digital message on the driving signal. The synchronization error power at
the output of each subsystem is fed into low pass filters(LPFs) separately. Then the filtered

signal is converted into a digital signal at threshold detectors(TD)[Cuomo & Oppenheim,

1993], which is then passed through NOT gates in order to invert and to confirm with the
original assignment of high and low states of the digital signal. An OR gate at the receiver
combines all the outputs from the NOT gates. Thus, when the digital message imposed on

the driving signal is ’1’, any one of the n subsystems at the receiver will synchronize and the
output at the OR gate will be ’1’. None of the subsystems will synchronize, if the modulated
digital message is zero, and hence ’0’ will be the output of the OR gate.
We illustrate our method in the case of the Lorenz system with the same parameters

used by Pérez and Cerdeira [1995] and with the modulation step as small as n = 5. The
Lorenz system (1) is known to exibit chaotic behavior when the parameter b takes any value
between 1.5 and 6.8. One can assign a specific value to b within this range for the purpose
of modulation. We select the region b = {3.0, 4.0} for the illustration of our method. As an

example, the parameter b in the transmitting system is allowed to take any one of the five
values 3.1, 3.3, 3.5, 3.7, 3.9 for high state of the digital signal, while it can have any one the
values 3.2, 3.4, 3.6, 3.8, 4.0 for the low state. In the receiver part, we use 5 subsystems with
the modulation parameter b fixed at 3.1, 3.3, 3.5, 3.7, 3.9 respectively.
The method works as follows. Suppose we have a high state in the digital message, to

start with. Then, the receiver is driven by the modulated driving signal with modulation
parameter b = 3.1. For the next high state in the message the modulation is done with
b = 3.3 and this process continues upto the b value 3.9. Then the value of b is reset
to 3.1. This procedure is also applicable to the low state of the message but with the

modulation parameter(b) taking the values 3.2, 3.4, 3.6, 3.8, 4.0 in that order. In the receiver
part, all the subsystems are driven by the modulated driving signal. However, the subsystems
R1 , R2 , R3 , R4 and R5 will synchronize only for the values of modulation parameter at the
transmitter, 3.1, 3.3, 3.5, 3.7, 3.9 respectively. When the synchronization is achieved in a

5
subsystem, the error power will be below the threshold level and so the message obtained
from TD will be at the low state. It is inverted to the high state using the NOT gate. But
a low state is attained by the digital message at the NOT gate, if the subsystem is not

synchronized. Thus a high state message is obtained as an output at the OR gate if any
one of the subsystems is synchronized and a low state digital message is recovered when
asynchronization takes place at all the subsystems. A test digital message transmitted and
recovered by this scheme by numerical simulation is shown in fig. 3.

Actually the above possibility is only a test case and has been discussed for illustrative
purpose. One can modulate b in a random way also (that is, at the transmitter circuit,
b can be assigned any one of the n chosen values in a random manner) rather than in a
predetermined way. The method presented here is applicable equally well to such possibilities

also.
The return map constructed using the modulated driving signal is shown in fig. 4. Here
we have 10 strips in each segment. This corresponds to 2 × n strips, where n is the step of
the parameter modulation. Because of the complicated nature of the return map it is very

difficult to find the strip on which the point(Am , Bm ) falls in each segment of the attractor
and so it is difficult to unmask the message from the return map. Even if one can identify
the location of the points in the strip, still one has a serious difficulty in assigning ’0’ and
’1’ to the strips since there are (22n − 2)3 − 1 chances to make a mistake, for a successful
reconstruction of the message. For n = 1, there are 7 chances [Pére & Cerdeira, 1995] of

making a mistake during the unmasking of the message from the modulated driving signal
and for n = 5, this chances increases to a high value of 1,067,462,647 (of the order of 109 ).
So it is almost impossible to unmask the message masked by our scheme, even at as low
value as n = 5.

We can complicate the attractor in the return map further by the second method. In the
Lorenz system, synchronization [Pecora & Carroll, 1991] is possible by both x driving and
y driving, since the conditional Lyapunov exponents are all negative in both the cases. We
use this property in the second method. Here the principle of transmitting and receiving

6
the message is the same as described above, but instead of using a single driving signal
x(t), we now use both x(t) and y(t) as the driving signals alternatively. For the odd digital
bits, we drive the receiver by modulated x(t) whereas the even digital bits are transmitted

using modulated y(t) as the driving signal. The block diagram of the transmitting and
receiving systems with modulation step n is shown in fig. 5. Here we use 2n subsystems
in the receiver, 2 subsystems for each value of modulation parameter (which are used for
the transmission of high state of the digital message), one with modulated x driving and

the other with modulated y driving. The recovered message is similar to the one obtained
in our previous method. The return map constructed from the modulated drive signals
(modulated x(t) and modulated y(t)) for the modulation step n=5 is shown in fig. 6. From
the return map, it is impossible to unmask the message since the attractor of the y drive

signal is merged with the attractor of the x driving signal.

We have also applied our scheme to the simplest autonomous Chua circuit generator of
the chaotic signal [Madan, 1993; Chua et al., 1992]. We have used the unnormalized circuit
equations
dv1
C1 = (1/R)(v2 − v1 ) − f (v1 ),
dt
dv2
C2 = (1/R)(v1 − v2 ) + iL , (4)
dt
diL
L = −v2 ,
dt
where f (v1 ) = Gb v1 + 0.5(Ga − Gb )[|v1 + Bp | − |v1 − Bp |] during our verification with R as the

modulation parameter. The other parameters value used are C1 = 10.0 nF , C2 = 100.0 nF
and L = 18.0 mH. The Chua circuit can also be synchronized either with the driving signal
v1 or with v2 and so we can again use them as alternative drive variables. One can assign
any value to the parameter R between 1555 Ω to 1960 Ω, except in some small regions where

periodic windows occur. Here the largest Lyapunov exponent calculated has positive values
and hence the circuit exibits chaotic behaviour. We find from our numerical studies that
it is even more difficult to extract message in this case with multiparameter modulation
compared to the Lorenz system, due to the nature of the chaotic attractor.

7
Thus, in this Letter, we have described two simple and straightforward methods to com-
plicate the attractor in the return map of the modulated driving signal: One with multi-step
parameter modulation by using a number of values to the modulation parameter b and the

other by using alternative x and y driving combined with multistep parameter modula-
tion. We have illustrated them with the Lorenz system and Chua’s circuit as examples. We
conclude that it is almost impossible to reconstruct the digital message transmitted in our
scheme by the use of the return map formed by the modulated driving signal.

ACKNOWLEDGMENTS

This work has been supported by the National Board of Higher Mathematics, Department
of Atomic Energy, Government of India and the Department of Science and Technology,
Government of India through the research projects. The authors thank Dr. K. Murali for

many valuable suggestions.

REFERENCES

Chua, L.O., Kocarev, Lj., Eckert, K. & Itoh, M. [1992] “ Experimental chaos synchroniza-
tion in Chua’s circuit” Int. J. Bifurcation and Chaos 2 705-708.

Cuomo, K.M. & Oppenheim, A. V. [1993] “Circuit implementation of synchronized chaos

with application to communication” Phys. Rev. Lett. 71 65-68.

Hayes, S., Grebogi, C. & Ott, E. [1993] “Communicating with chaos” Phys. Rev. Lett.
70, 3031-3035.

Kocarev, L. & Parlitz, U. [1995] “General approach for chaotic synchronization with ap-
plications to communication” Phys. Rev. Lett. 74 5028-5031.

Lakshmanan, M. & Murali, K. [1996] Chaos in Nonlinear Oscillators: Controlling and

Synchronization (World Scientific, Singapore).

8
Madan, R. N., [1993] Chua’s circuit: A Paradigm for Chaos (World Scientific, Singapore).

Mensour, B. & Longtin, A. [1998] “Synchronization of delay-differential equations with

application to private communication” Phys. Lett. A244, 59-70.

Minai, A. A. & Pandian, T. D. [1998] “Communicating with noise: How chaos and noise
combine to generate secure encryption keys” Chaos 8 621-628.

Murali, K. & Lakshmanan, M. [1993] “Transmission of signals by synchronization in a

chaotic van der pol-duffing oscillator” Phys. Rev. E48, R1624-1626

Murali, K. & Lakshmanan, M. [1993] “Synchroning chaos in driven chua’s circuit” Int. J.
Bifurcations Chaos 3, 1057-1066.

Murali, K. & Lakshmanan, M. [1998] “Secure communication using a compound signal

from generalized synchronizable chaotic systems” Phys. Lett. A241, 303-310.

Pecora, L. M. & Carroll, T. L. [1990] “Synchronization in chaotic systems” Phys. Rev.

Lett. 64, 821-823 (1990);

Pecora, L. M. & Carroll, T. L. [1991] “Driving systems with chaotic signals” Phys. Rev.
A44, 2374-2383.

Pérez, G. & Cerdeira, H.A. [1995] “Extracting messages masked by chaos” Phys. Rev.
Lett. 74, 1970-1973.

Zhang, Y., Dai, M., Hua, Y., Ni, W, & Du, G. [1998] “Digital communication by active-
passive-decomposition synchronization in hyperchaotic systems” Phys. Rev. E58

3022-3027.

9
FIGURES

60

50

40
Bm

30

20

10

0
-15 -10 -5 0 5 10 15 20

Am
FIG. 1. Return map between Am and Bm in single-step parameter modulation for the Lorenz

system.

10
Transmitter
x(t)

R1 R2 − − − Rn

LPF LPF − − − LPF

TD TD − − − TD

NOT NOT − − − NOT

Receiver

OR

FIG. 2. The block diagram of transmitter and receiver in multi-step parameter modulation

with x driving (LPF-Low pass filter, TD-Threshold detector).

11
Message Transmitted 2

1.5

0.5

−0.5

−1
0 200 400 600 800 1000 1200

2
Recovered Message

1.5

0.5

−0.5

−1
0 200 400 600 800 1000 1200
Time (prenormalized)

FIG. 3. Transmitted and recovered message in multi-step parameter modulation(n = 5) with

x driving.

12
60

50

40
Bm

30

20

10

0
-15 -10 -5 0 5 10 15 20

Am
FIG. 4. Return map between Am and Bm in multi-step parameter modulation(n = 5) with x

driving for the Lorenz system.

13
Transmitter
x(t) or y(t)

RB1 RB2

Receiver
OR

FIG. 5. The block diagram of transmitter and receiver in multi-step parameter modulation

with x and y driving (RB1 -receiver block in fig.2 with x driving subsystems, RB2 -receiver block

in fig.2 with y driving subsystems).

14
70

60

50

40
Bm

30

20

10

0
-15 -10 -5 0 5 10 15 20

Am
FIG. 6. Return map between Am and Bm in multi-step parameter modulation(n = 5) with

alternate x and y driving for the Lorenz system.

15