Secured Communication of Speech Signal Using the Discrete Cosine

Transform Based on Hyperchaos-System

Noureddine Aissaoui

1 a

, Fethi Demim

2 b

, Abdenebi Rouigueb

3 c

, Hadjira Belaidi

6 d

Ali Zakaria Messaoui

4 e

, Kahina Louadj

5 f

, Abdelkrim Nemra

2 g

, Ahmed Allam

, Yasmine Saidi

Said Sadoudi

and Mohamed Salah Azzaz

Laboratoire Syst

emes Electronique et Num

eriques, Ecole Militaire Polytechnique, Bordj El Bahri, Algiers, Algeria

Laboratory of Guidance and Navigation, Ecole Militaire Polytechnique, Bordj El Bahri, Algiers, Algeria

Laboratory of Artiﬁcial Intelligence and Virtual Reality, Ecole Militaire Polytechnique, Bordj El Bahri, Algiers, Algeria

Complex Systems Control and Simulators Laboratory, Ecole Militaire Polytechnique, Bordj El Bahri, Algiers, Algeria

Laboratoire d’Informatique, Mathmatiques, et Physique pour l’Agriculture et les For

ets, Universit

e de Bouira, Algreia

Signals and Systems Laboratory, Institute of Electrical and Electronic Engineering,

University M’Hamed Bougara of Boumerdes, Algeria

Ecole Nationale Polytechnique, Algiers, Algeria

Keywords:

Chaos-Based Cryptography, Secure Chaotic Communication, Speech Signal, Discrete Cosine Transform,

Encryption-Based Diffusion.

Abstract:

This paper proposes a novel approach that combines chaos-based encryption and Discrete Cosine Transform

(DCT) to ensure high-level speech security and robustness against attacks. In this approach, the encryption

process is based on Lorenz’s hyperchaotic system, which utilizes the One Time Pad approach to encrypt the

speech DCT coefﬁcients. The effectiveness of this approach has been validated through experiments on two

PCs interconnected via real-time serial communication links (USB-RS232), which showed that the original

speech is effectively hidden, and the proposed solution is highly resistant to possible attacks. Moreover, the

proposed solution can be implemented in real-time applications using technologies such as FPGA.

1 INTRODUCTION

Speech-based communication has taken an important

place in many ﬁelds, such as civil applications like

trade, military defence, telephone banking and tele-

conferences, etc. This type of communication is in

continuous evolution with the continuous investiga-

tion for better speeds, improved mobility and espe-

cially high conﬁdentiality. The exchange of impor-

tant conﬁdential information via unsecured commu-

nication provides easy access to secret information

https://orcid.org/0000-0002-5328-2414

https://orcid.org/0000-0003-0687-0800

https://orcid.org/0000-0001-5699-2721

https://orcid.org/0000-0003-2424-626X

https://orcid.org/0000-0001-5753-5776

https://orcid.org/0000-0002-4203-6357

https://orcid.org/0000-0001-9237-9449

and facilitates the hacking process. Today’s secu-

rity devices are mainly based on encrypted systems

that guarantee speciﬁc requirements. On one hand,

the current encryption methods such as Rivest Shamir

Adleman (RSA), Data Encryption Standard (DES)

systems, and Rivest Cipher 4 (RC4) have already been

defeated and are therefore no more secured. In this

context, it is necessary to provide using other meth-

ods that are secure and have not yet been broken with

the generation of microprocessors, such as Cipher 5

or the Advanced Encryption Standard (AES). Thus,

chaotic cryptography is one of the alternatives created

in the last ten years, particularly quantum cryptogra-

phy, even though it cannot be opened due to its char-

acteristics and obscures the data from its original form

into difﬁcult-to-understand data.

In 1990, L.M. Pecora et al. established the practi-

cal feasibility of synchronization between two identi-

Aissaoui, N., Demim, F., Rouigueb, A., Belaidi, H., Messaoui, A., Louadj, K., Nemra, A., Allam, A., Saidi, Y., Sadoudi, S. and Azzaz, M.

Secured Communication of Speech Signal Using the Discrete Cosine Transform Based on Hyperchaos-System.

DOI: 10.5220/0012091600003543

In Proceedings of the 20th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2023) - Volume 2, pages 69-76

ISBN: 978-989-758-670-5; ISSN: 2184-2809

cal chaotic systems, thus proving the exploitation and

efﬁcient application of chaotic systems in communi-

cations (Pecora and Carroll, 1990). In 1993, K.M.

Cuomo et al. proved that for some chaotic systems,

Lorenz for example, the synchronization property, ac-

cording to the technique of L.M. Pecora et al. is ro-

bust despite small perturbations in the coupling sig-

nal. This property allowed these researchers to pro-

pose the ﬁrst scheme of a chaotic communication sys-

tem based on the principle of Chaotic Addition Mask-

ing (Cuomo and Oppenheim, 1993).

In 2003, A.S. Dmitriev developed a new ap-

plication base-chaos in communications called Di-

rect Chaotic Communication (DCC) (Dmitriev et al.,

2003). This new chaotic-based technology has

opened a new ﬁeld of application of chaos, namely

ultra-wideband communications. In 1997, T. Yang et

al. published the ﬁrst work presenting the concept of

chaotic encryption systems (Yang and Chua, 1997).

In this context, they proposed to integrate cryptogra-

phy into chaotic communication systems to improve

the security degree of the latter. Chaotic pseudo-

random sequences have desirable cryptographic prop-

erties, such as good randomness, deterministic dy-

namics, structure complexity, and sensitivity to ini-

tial conditions. Therefore, cryptanalysts have adopted

several cryptographic standards based on chaos the-

ory.

Subsequently, P.G. Vaidya et al. proposed chaotic

cryptography with chaotic timing in 1998 (Vaidya and

Ronge, 1998). The chaos-based digital cryptography

system, based on pseudo-random number generators,

does not depend on the chaos timing; it uses the initial

conditions and control parameters as the secret key

(Li et al., 2001). The one-dimensional logistic map

is often used as a Pseudo-Random Number Generator

(PRNG) in encryption, for example M.H.A. Samah

et al. introduced a method where the logistic map is

used to scatter the samples and the one-dimensional

circular map is used to confuse the samples (Samah

and Eihab, 2013). E. Mosa et al. (Mosa et al., 2009)

presented cryptography in the transformation domain,

based on the two-dimensional Baker map. The three-

dimensional Lorentz map as a pseudo-random num-

ber generator is discussed by B.S. Sattar (Sadoudi and

Azzaz, 2009).

In 2009, S. Sadoudi and M.S. Azzaz, developed

a hardware implementation of the Rossler chaotic

system to secure communication (Sadoudi and Az-

zaz, 2009) (Sattar and Rana, 2015), followed by a

new auto-switched chaotic system and its Field Pro-

grammable Gate Array (FPGA) implementation is

presented in (Azzaz et al., 2013b), as well as some

work based on synchronized hybrid chaotic genera-

tors is proposed in (Azzaz et al., 2013a). In the same

year, S. Sadoudi et al. developed a wireless hyper-

chaotic communication system for secure real-time

image transmission (Sadoudi et al., 2013). Subse-

quently, in the same year, they proposed an FPGA

real-time implementation of Chen’s chaotic system

to secure chaotic communications (Sadoudi et al.,

2009). Several works are proposed founded on real-

time FPGA implementation of Lorenz’s chaotic gen-

erator, Dufﬁng’s chaotic attractor and experimental

synchronization technique for chaotic communica-

tions (Azzaz et al., 2009), (Sadoudi et al., 2014) and

(Sadoudi et al., 2015).

Currently, from 2020 to 2022, several researchers

have been working on implementing FPGA-based

real-time chaos secure encryption systems, as well

as chaos-based video encryption algorithms (Azzaz

et al., 2019)-(Hadjadj et al., 2022). Secure commu-

nication is crucial in the digital age, with data pri-

vacy and conﬁdentiality being paramount concerns.

One promising method for encrypting speech sig-

nals is the use of the Discrete Cosine Transform

(DCT) based on hyperchaos-system. Hyperchaos-

system employs hyper-chaotic systems, which are

even more unpredictable than traditional chaotic sys-

tems, adding an extra layer of security to the encryp-

tion process (Zghair et al., 2021). Combining the

DCT with hyperchaos-system allows efﬁcient encryp-

tion of speech signals while ensuring accurate recon-

struction at the receiver’s end. This novel approach

presents an effective solution for secure speech sig-

nal communication (Sathiyamurthi and Ramakrish-

nan, 2022).

In this paper (Abdallah and Meshoul, 2022), au-

thors propose multilayer cryptosystems for audio

communication encryption by continuously fusing

audio signals with speech signals without silent pe-

riods. Three levels of encryption (fusion, substitu-

tion, and permutation) are considered, and the pro-

posed approach shows increased security compared to

one-dimensional logistic map-based encryption tech-

niques.

Cryptography is a mathematical discipline allow-

ing one to perform operations on an intelligible text

to ensure some security properties of the informa-

tion. This work aims to secure speech communication

via an encryption system based on the chaotic PRNG

technique using a transmission channel-based USB-

RS232. Hybrid chaotic PRNG-based cryptography

is secure and has powerful confusion and diffusion

properties necessary for strong encryption. Each unit

in the confusion and scattering area wants to block

the derivation of the secret write key or the possible

prevention of the original message. While scattering

ICINCO 2023 - 20th International Conference on Informatics in Control, Automation and Robotics

increases the repetition of the clear text over the en-

crypted text key to make it obscure, confusion is used

to make the invalid encrypted text. Only confusion

can make stream encryption work; otherwise, stream

encryption and block encryption require to scatter.

This paper is organized as follows: Section II de-

scribes hyper-chaotic systems and presents a hyper-

chaotic system based-speech encryption using a Dis-

crete Cosine Transformation. While, in Section III,

the speech encryption process is proposed. Then, in

Section IV, the simulations and experiment results are

presented. At last, Section V concludes the paper and

gives some future works.

2 HYPERCHAOTIC SYSTEM

BASED-SPEECH ENCRYPTION

A dynamical system can generate hyperchaos if it has

at least four state variables. It is modeled by at least

four nonlinear differential equations. The mathemati-

cal model of a nonlinear dynamical system represents

a four dimensional hyper chaotic system as follows:











∂x

∂t

= hy − ax + yz

∂y

∂t

= hx − by − xz

∂z

∂t

= −dz + xy + w

∂w

∂t

= xy + cw

(1)

where (x, y, z, w) are state variables and (a, b, c, d, h)

represent system parameters. The initial states and the

eventual control parameters show that this system is

chaotic as: x(0) = 0.8, y(0) = 0.3, z(0) = 10.1, w(0) =

4.5, a = 5, b = −0.5, c = 2.2, d = 1, h ≥ 4.85. The ﬁ-

nal state of a chaotic system is extremely sensitive to

small changes in the initial state. With chaotic attrac-

tor, the phase diagram is the set of trajectories that are

the solutions of the hyper-chaotic system as shown

in Eq. (1), represented in the phase plane initiated

from different initial conditions. A chaotic switching-

rule allows the speciﬁed chaotic system to change its

behaviour, while still producing complicated chaotic

attractors. The properties of the chaotic system are

so interesting for data encryption. Thus, we have

developed an encryption system based on a hyper-

chaotic function. The hyperchaotic system is used as

a PRNG. Scattering of speech samples in this domain

is performed only by Discrete Cosine Transformation

(DCT) using PRNG, as shown in Figure 1. Before

the transmission of the encrypted speech signal using

DCT, it is converted into the time domain by the In-

verse Discrete Cosine Transformation (IDCT). In the

reception phase,, we will apply the inverse operation

of encryption to recover the original speech signal. It

is useful to note that the generation of the key stream

encryption is done by the same initial values, as well

as the same control parameters.

Figure 1: Encryption process ﬂowchart.

Nowadays, telecommunication technologies are

involved everywhere. The large-scale use of mod-

ern speech communication technologies via the In-

ternet or mobile telephony requires securing informa-

tion exchanges. The concept of security involves per-

forming transformations on the original speech signal

to mask it, making it unintelligible to unauthorized

users, while ensuring the possibility of reconstituting

it for authorized users. To meet these requirements,

several methods of processing the speech signal have

been proposed in the literature.

In the case of key management for streaming

encryption, it has been generated using a four-

dimensional hyper-chaotic system as seen in Figure

2 which represents the simulation results. This ﬁgure

shows the change of variables over time, all variables

of the states whose starting point is the point men-

tioned in the initial conditions. These initial condi-

tions and parameters of the system are the key func-

tions, which are given as input to the hyper-chaotic

system. At the output of the PRNG, the state vari-

ables are given random values, and the normalization

operation of each key is presented as follows:

¯x =

x − x

min

max

− x

min

(2)

where x

min

is the minimum value of the generated key

and x

max

is the maximum value of the generated key

stream encryption. The normalized key is converted

to a 16-bit integer as follows:

˜x

= (round( ¯x

× 10

))mod 2

(3)

where i = {1, 2, 3, . .. , N

} and N

is the number of

samples of the speech signal.

Secured Communication of Speech Signal Using the Discrete Cosine Transform Based on Hyperchaos-System

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-10

-5

Normalized vectors

Normalized X-vector

Normalized Y-vector

Normalized Z-vector

Normalized W-vector

Figure 2: Normalized key to different variables x, y, z and w.

3 SPEECH ENCRYPTION

PROCESS

To perform this process, we have chosen an audio

recording of 4.5 seconds with a sampling rate of

8000 HZ. In the example illustrated in Figure 3 (a),

the original speech signal is recorded and then con-

verted to the Wav format using a time domain rep-

resentation. By deﬁnition, the audiogram graphically

represents the evolution of the capacity or intensity

of a speech to see its temporal envelope. A Dis-

crete Cosine Transform (DCT) describes a ﬁnite se-

quence of data points in terms of the sum of cosine

functions varying in frequency. It is used in signal

and image processing and especially in data compres-

sion. Indeed, it has an exceptional energy aggregation

property, and the low-frequency coefﬁcients carry the

information. In this case, Figure 3 (b) shows the

original speech signal in transformation mode (Azzaz

et al., 2019). The chaotic signals in the process of

masking the speech signal are based on the streaming

process. At the output of the PRNG, the keys stream

encryption has given random values, followed by a

normalization operation for each key; after this op-

eration, each normalized key will be converted into

a 16-bit coded integer and the encryption operation

is performed by the XOR operator between the key

stream encryption and the transformed speech scram-

bles to completely encrypt the speech signal. How-

ever, before this operation, the transformed speech

scrambles are converted into equivalent decimal val-

ues using 16-bit quantization. The samples are pro-

cessed in a ﬁxed-size block, where the block size de-

pends on the random key size (Sadoudi et al., 2015)

(Kaibou et al., 2021). The following process presents

the diffusion using the XOR function between each

number of samples of the intercepted speech signal

and the key converted to a 16-bit integer, as follows:

˜e

= ˜m

+ ˜x

(4)

where ˜x

presents the key converted to a 16-bit inte-

ger, ˜m

deﬁnes the samples of the original speech, and

˜e

presents the encrypted samples. If the correlation

Table 1: Correlation coefﬁcient and SNR results.

Key Correlation

m,e

SNR (dB)

x(t) 0.000003713 -263.9514

y(t) 0.0003711 -263.9649

z(t) -0.002972 -263.9557

w(t) 0.003713 -263.9682

coefﬁcient is zero, the original and encrypted signals

are different. The smaller the values of the correla-

tion coefﬁcient, the better the cryptographic process.

It can be calculated as follows:

m,e

cov(m, e)

(5)

where cov(m, e) is the covariance between the original

signal m and the encrypted signal and σ

, σ

are the

standard deviations of the signal m and signal e.

The determination of the key is based on the best

performance in terms of the correlation coefﬁcient.

The analysis of the correlation coefﬁcient measures

the quality of the encryption system. By analysis, if

the correlation coefﬁcient is equal to zero, the origi-

nal signal and the encrypted signal are considered to

be completely different. When the Signal-to-Noise

Ratio (SNR) is a negative value in dB, it means that

the signal intensity is lower than the noise intensity,

making the encrypted data not detectable, and the key

represents a good correlation (Hadjadj et al., 2022).

In this case, we compute the noise between the orig-

inal speech signal and the encrypted speech signal as

follows:

SNR = 10 × log10

∑

i=1

∑

i=1

− e

)

(6)

4 SIMULATION RESULTS AND

DISCUSSION

Several experimental analyses are performed to test

the effectiveness of the speech-based cryptography

technique. In this analysis, we will measure the qual-

ity of our cryptographic system through the analysis

of the coefﬁcient. The latter is given by the correlation

between the similar segments in the original speech

signal and the encrypted one. Table 1 shows different

correlation coefﬁcients for the different keys, and we

notice that the key x(t) converges to zero, which indi-

cates a good correlation. These results are close to

each other, but we choose the key having the small-

est negative value, which is x(t). Self-correlation is

a mathematical method to analyze a periodic signal

using the cross-correlation of a signal with itself (see

ICINCO 2023 - 20th International Conference on Informatics in Control, Automation and Robotics

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-1

-0.5

0.5

Magnitude (v)

Input speech

(a) Audiogram of the original speech.

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-15

-10

-5

Magnitude (v)

Transformed speech

(b) Original speech using a discrete cosine transformation.

Figure 3: Original speech and its DCT over time.

Figure 4: Self-correlation analysis of the key x(t).

Figure 4). The proposed key ﬂow gave better corre-

lation values. After selecting the encryption key, the

next step is to use this key in masking the speech sig-

nal using the diffusion process. An example of a tem-

poral representation of an encrypted signal is shown

in Figure 5, which is almost impossible to identify the

transmitted speech signal. The obtained results from

the simulation demonstrate the disappearance of the

speech signal completely, which proves the success

of the function of the encryption process. The advan-

tages of chaotic masking based on diffusion are the

simplicity of its realization and the possibility of us-

ing it to mask analogue or digital signals with high

SNR channels. As far as the security analysis of the

speech signal is considered, we will test the efﬁciency

and robustness of our cryptographic system for the

encryption of a speech signal by analyzing the two

signals presented in Figure 6. It is a measure of the

similarity of the signal to itself over time. Typically,

the correlation is maximal for a zero-time delay. The

principle of correlation is to extract future predictive

information from the predicted signal values. In this

part, we tested the efﬁciency and robustness of our

adopted encryption system. From Figure 7, we can

notice that the autocorrelation of the encrypted signal

is like white noise converging to zero. Therefore, pre-

dicting the signal from its predicted state is not possi-

ble. Figure 8 illustrates the diffusion of information in

the crypto-system. The histogram is a statistical study

tool that is used to illustrate the efﬁciency of the dif-

fusion of information in the encrypted data brought

by the used cryptosystem. As for the original signal

history, it is presented in the form of a stripe, as well

as with a uniformly presented encrypted signal. In

the case of the spectrogram analysis signal, the anal-

ysis of the time-frequency distribution of the original

speech signals is shown in Figure 9, along with an en-

Table 2: Perceptual evaluation of speech quality based tech-

nique.

Original

speech

Encrypted

speech

Decrypted

speech

4.176 1.260 3.952

crypted signal that has a uniform frequency behavior

over time. In the encryption process, the encrypted

signal in the transformation domain is converted into

the time domain by the IDCT and in the decryption

process step, the speech signal is recovered at the re-

ceiver. The latter must generate the same key stream

encryption with the same initial values and control pa-

rameters to restore the original speech signal. The

received signal is transformed into a frequency rep-

resentation using DCT. As for the diffusion opera-

tion, it is performed between the samples of the en-

crypted signal and its counterpart of the encryption

key. These encrypted samples are converted to equiv-

alent 16-bit decimal values. Note that the key is a 16-

bit encoded integer, and the XOR operation between

the encrypted samples and the encryption key yields

the original signal in its frequency representation, as

shown in Figure 10 (a). To recover the original sig-

nal, a transformation from the frequency to the time

domains is necessary using the inverse DCT, as shown

in Figure 10 (b).

In the decryption process step, the signal recov-

ery at the receiving end is still with a scattering op-

eration, but this time it is performed between the en-

crypted signal samples and its counterpart of the en-

cryption key. These encrypted samples are converted

into equivalent 16-bit decimal values.

Figure 11 is the recovered speech signal in the

time domain. According to this example, the recon-

structed speech signal corresponds well to the original

transmitted original signal. We can see that the orig-

inal speech signal has been recovered successfully in

terms of audio quality with its appropriate character-

istics, as seen in Figure 3. Table 2 presents the Percep-

tual Evaluation of Speech Quality (PESQ) based eval-

uation technique. The results for the PESQ are almost

similar to the original speech, which shows the perfor-

Secured Communication of Speech Signal Using the Discrete Cosine Transform Based on Hyperchaos-System

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

Magnitude (v)

Encrypted speech

(a) Encrypted speech using DCT.

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-1

Magnitude (v)

Encrypted speech transformation

(b) Encrypted speech using IDCT.

Figure 5: Encrypted speech presentation using transformation process.

0 1 2 3 4

Algorithm iterations

-0.6

-0.4

-0.2

0.2

0.4

0.6

Normalized magnitude

Original speech signal

0 1 2 3 4

Algorithm iterations

-1

-0.8

-0.6

-0.4

-0.2

0.2

0.4

0.6

0.8

Normalized magnitude

Encrypted speech signal

Figure 6: Speech encryption normalization.

-5 0 5

Delay time (s)

-0.4

-0.2

0.2

0.4

0.6

0.8

Autocorrelation

Speech autocorrelation

-5 0 5

Delay time (s)

-0.2

0.2

0.4

0.6

0.8

Autocorrelation

Speech autocorrelation

Figure 7: Presentation of speech signal autocorrelation.

Original speech

-1 -0.5 0 0.5 1

Magnitude

500

1000

1500

2000

2500

3000

3500

4000

Number of samples

Speech distribution probability

Normal distribution

Encrypted speech

-1 -0.5 0 0.5 1

Magnitude

100

200

300

400

500

600

700

800

900

Number of samples

Speech distribution probability

Normal distribution

Figure 8: Histogram analysis.

Original speech spectrogram

2 4 6 8

Time (s)

Frequency (Hz)

-140

-120

-100

-80

-60

-40

Magnitude (dB)

Encrypted speech spectrogram

2 4 6 8

Time (s)

Frequency (Hz)

-140

-120

-100

-80

-60

-40

Magnitude (dB)

Figure 9: Spectrogram analysis presentation.

mance of our proposition. We used the PESQ method

for evaluation in our simulation. Based on PESQ

analysis, the results shown in Table 2 depict the rate

of loss of the frames. The PESQ scale (PESQ > 3),

which offers good speech quality and high accuracy,

preserves the intelligibility of the signal. The PESQ

score obtained between the original and the encrypted

speech is lower than the value of 2, which is consid-

ered satisfactory.

As for the realization of secure speech com-

munication, speech encryption is done directly in

real-time using a USB-RS232 serial communication

converter. Figure 12 illustrates this implementa-

tion via this transmission channel. In this part,

an encryption device is used to transmit speech via

two PCs via a graphical interface which is con-

sidered as a transmission device for recording and

processing the speech signal which will be inter-

cepted by a decryption device with a reverse oper-

ation, as shown in Figure 13. In the emission pro-

cess, we used an HP-PC with Pentium (R) Dual-core

CPU T 4500 @ 2.30 GHz, 2.00 GB RAM and 64-bit

operating system with windows10, while in the recep-

tion process, we used a DELL-PC with Pentium (R)

Dual-core CPU T 4500 @ 2.30 GHz, 2.00 GB RAM

and 32-bit operating system with windows7.

5 CONCLUSIONS

This study advances secure communication through

hyper-chaotic systems, leveraging chaotic behavior

for effective speech signal encryption using logistic

and sinusoidal maps. Validation via stringent trans-

mission tests on encrypted speech using RS232 pro-

tocol between two PCs demonstrates strong encryp-

tion efﬁciency, signal-to-noise ratio, and bit error rate

performance, conﬁrming the method’s robustness in

securing speech communication.

Our work’s utilization of hyper-chaotic systems

and its successful implementation of real-time en-

cryption signify a compelling step towards enhanc-

ing speech communication security. The integration

ICINCO 2023 - 20th International Conference on Informatics in Control, Automation and Robotics

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-1

Magnitude (v)

Encrypted speech transformation using DCT

(a) Encrypted speech using DCT.

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-15

-10

-5

Magnitude (v)

Decrypted speech transformation using IDCT

(b) Decrypted speech using IDCT.

Figure 10: Encrypted and Decrypted speech presentation using transformation process.

Figure 13: Graphical interface for transmission and reception of encrypted speech.

0 0.5 1 1.5 2 2.5 3 3.5 4

Algorithm iterations

-1

-0.5

0.5

Magnitude (v)

Final step with decrypted speech

Figure 11: Final step with the decrypted speech signal.

Figure 12: Experimental validation of the proposed ap-

proach of secured speech.

of chaotic systems in this context paves the way for

further advancements and applications in secure com-

munication methods. This research opens up exciting

possibilities for developing novel and practical solu-

tions to safeguard communication in various domains.

In addition, future work could extend the proposed

approach beyond speech signals to other types of sig-

nals, such as images and videos. Adapting the algo-

rithm for different signal types would allow for ad-

dressing security concerns in various domains and ap-

plications.

Moreover, implementing the algorithm on differ-

ent hardware platforms, such as FPGAs and micro-

controllers, will be a key focus to achieve real-time

performance and ensure portability for practical de-

ployments. This will allow the method to be inte-

grated into various systems and devices, enhancing

its usability in real-world scenarios.

Additionally, exploring the design space of hyper-

chaotic systems is planned to further optimize the pro-

posed algorithm. Evaluating its performance in terms

of security and computational complexity under vari-

ous conditions will help in understanding its strengths

and limitations. This evaluation will be crucial in as-

sessing the potential advantages of the method com-

pared to existing solutions.

To facilitate a comprehensive evaluation, per-

formance metrics and execution times will be pro-

vided. These measurements will enable researchers

and practitioners to assess the efﬁciency and effec-

tiveness of the proposed method for various applica-

tions and scenarios. Making informed decisions about

the best implementation approach will be essential for

successful deployment. Investigating error correction

codes and compression techniques will enhance the

system’s robustness and signal integrity during trans-

mission and storage.

REFERENCES

Abdallah, H. and Meshoul, S. (2022). A multilayered audio

signal encryption approach for secure voice commu-

nication. pages 1–18.

Secured Communication of Speech Signal Using the Discrete Cosine Transform Based on Hyperchaos-System

Azzaz, M., Tanougast, C., and Benssalah, A. (2019). An ef-

ﬁcient and lightweight multi-scroll chaos-based hard-

ware solution for protecting ﬁngerprint biometric tem-

plates. pages 1–13.

Azzaz, M., Tanougast, C., Sadoudi, S., and Bouridane,

A. (2013a). Synchronized hybrid chaotic generators:

Application to real-time wireless speech encryption.

pages 2035–2047.

Azzaz, M., Tanougast, C., Sadoudi, S., and Dandache, A.

(2009). Real-time fpga implementation of lorenz’s

chaotic generator for ciphering telecommunications.

In Proceedings of Joint IEEE North-East Workshop

on Circuits and Systems and TAISA Conference, pages

1–4.

Azzaz, M., Tanougast, C., Sadoudi, S., Fellah, R., and Dan-

dache, A. (2013b). A new auto-switched chaotic sys-

tem and its fpga implementation. pages 1792–1804.

Cuomo, K. and Oppenheim, A. (1993). Circuit implementa-

tion of synchronized chaos with applications to com-

munications. In Physical Review Letters.

Dmitriev, A., Kyarginsky, B., Panas, A., and Starkov, S.

(2003). Experiments on direct chaotic communica-

tions in microwave band. pages 1495–1507.

Hadjadj, M., Sadoudi, S., Azzaz, M., Bendecheche, H., and

Kaibou, R. (2022). A new hardware architecture of

lightweight and efﬁcient real-time video chaos-based

encryption algorithm. pages 1–14.

Kaibou, R., Azzaz, M., Benssalah, M., Teguig, D., Hamil,

H., Merah, A., and Akrour, M. (2021). Real-time

fpga implementation of a secure chaos-based digital

crypto-watermarking system in the dwt domain using

co-design approach.

Li, S., Mou, X., and Cai, Y. (2001). Pseudo-random Bit

Generator Based on Couple Chaotic Systems and Its

Applications in Stream-Cipher Cryptography. Lecture

Notes in Computer Science book series, London.

Mosa, E., Messiha, N., and Zahran, O. (2009). Chaotic

encryption of speech signals in transform domains.

In Proceedings of IEEE International Conference on

Computer Engineering and Systems, Cairo, Egypt.

Pecora, L. and Carroll, T. (1990). Synchronization in

chaotic systems. In Physical Review Letters.

Sadoudi, S. and Azzaz, M. (2009). Hardware implementa-

tion of the rossler chaotic system for securing chaotic

communication. In Proceedings of 5th International

Conference of Sciences of Electronic, Technologies of

Information and Telecommunications.

Sadoudi, S., Azzaz, M., Djeddou, M., and Benssalah, M.

(2009). An fpga real-time implementation of the

chen’s chaotic system for securing chaotic communi-

cations. pages 467–474.

Sadoudi, S., Azzaz, M., and Tanougast, C. (2014). Novel

experimental synchronization technique for embed-

ded chaotic communications. In Proceedings of IEEE

International Conference on Control, Decision and

Information Technologies, pages 669–672.

Sadoudi, S., Tanougast, C., and Azzaz, M. (2015). Journal

of engineering science and technology review.

Sadoudi, S., Tanougast, C., Azzaz, M., and Dandache, A.

(2013). Design and fpga implementation of a wireless

hyperchaotic communication system for secure real-

time image transmission. pages 1–18.

Samah, M. and Eihab, B. (2013). Speech scrambling based

on chaotic maps and one time pad. In Proceedings of

IEEE International Conference on Computing Electri-

cal and Electronics Engineering, Khartoum, Sudan.

Sathiyamurthi, P. and Ramakrishnan, S. (2022). Speech

encryption using hybrid-hyper chaotic system and bi-

nary masking technique. pages 6331–6349.

Sattar, B. and Rana, S. (2015). A proposed voice encryp-

tion based on random lorenz map with dct permuta-

tion. page 90.

Vaidya, P. and Ronge, H. (1998). Implementation of chaotic

cryptography with chaotic synchronization. In Physi-

cal Review Letters.

Yang, T. and Chua, L. (1997). Impulsive stabilization for

control and synchronization of chaotic systems: The-

ory and application to secure communication.

Zghair, H. K., Mehdi, S., and Sadkhan, S. (2021). Speech

scrambler based on discrete cosine transform and

novel seven-dimension hyper chaotic system. pages

1–14.

ICINCO 2023 - 20th International Conference on Informatics in Control, Automation and Robotics