Phys. Scr. 99 (2024) 075238 https://doi.org/10.

1088/1402-4896/ad5476

PAPER

Cap like trajectories in 5D chaotic tangent hyperbolic memristive

RECEIVED
7 April 2024
model: fractional calculus and encryption
REVISED
20 May 2024
ACCEPTED FOR PUBLICATION
Muhammad Ali Qureshi1 , Najeeb Alam Khan2 , Salman Raza1 and
5 June 2024 Sheikh Muhammad Zeeshan Iqbal1
1
PUBLISHED Department of Physics, University of Karachi, Karachi 75270, Pakistan
14 June 2024 2
Department of Mathematics, University of Karachi, Karachi 75270, Pakistan
E-mail: [email protected](M A Qureshi)

Keywords: chaos, Lyapunov exponents, bifurcation, Atangana-Baleanu Caputo derivative, memristor, encryption

Abstract
This research aims to investigate the influence of model parameters and fractional order on a novel
mathematical model with tangent hyperbolic memristor. This investigation conducted by applying
Lyapunov exponents and bifurcation analysis. We utilize the Lyapunov exponent theory to
understand and characterize these chaotic behaviors under fractional indices. The Lyapunov
exponent, bifurcation, and phase diagrams have been depicted to explore the intricate dynamics of the
chaotic system governed by the chaotic equation. A novel approach termed Atangana-Baleanu-
Caputo (ABC) fractional derivative (FD) to generate phase portraits and gain insights into the system’s
behavior. The random numbers generated by the chaotic system are employed to distort the image
through an amalgamated image encryption (AIE) algorithm. Subsequently, the integrity of the
scrambled image has been assessed using various image security evaluation methods to reinforce the
notion that combining the chaotic system and image can constitute a valuable encryption key. Finally,
the chaotic model circuit realization uses active and passive components, and the outcomes are
compared with the numerical simulations.

1. Introduction

Several papers have been published on modeling chaotic and hyperchaotic systems, with widespread application
across diverse fields, including electrical circuits, mathematics, and physics [1–6]. It has emerged as a prominent
trend in recent years. Many of these publications have concentrated on incorporating chaotic systems into
electrical circuit modeling, a prevalent challenge in chaotic applications. The intricate nature of real-world
phenomena and their integration into chaotic modeling makes precise predictions formidable. However,
developing novel algorithms and computational techniques has significantly improved the accuracy of
predictions in chaotic systems. These advancements have led to significant breakthroughs in the understanding
and harnessing of chaos for practical applications. Consequently, chaotic systems have become valuable tools for
studying complex dynamics and nonlinear behavior in various disciplines. With further advancements in
technology and computational power, the potential for exploring the intricacies of chaotic systems and their
applications has become vast and promising.
Various methodologies exist for examining chaotic systems, including techniques such as phase portraits
employed to analyze the influence of model parameters on system behavior and assess Lyapunov exponents and
both chaotic and hyperchaotic tendencies. Additionally, an algorithm has been designed to scrutinize the impact
of the initial conditions by utilizing specialized tools. These tools can identify critical and bifurcation points in a
system, helping researchers understand the stability and predictability of chaotic behavior. Furthermore,
analyzing attractors, such as strange attractors and fractal dimensions, provides insights into chaotic systems’
long-term behavior and complexity. By combining these methodologies, researchers can comprehensively
understand chaotic systems and predict their future behavior.

© 2024 IOP Publishing Ltd

Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Classical models are extensions of the fractional models. As we know, classical calculus provides powerful
modeling tools for many phenomena inspected in the applied sciences, it does not offer a complete description
of some aberrant dynamics inside complex systems observed in nature. Experimental results showcase that
different phenomena associated with these complex systems exhibit non-local dynamic behavior, including
long-term effects, and fractional derivatives have been shown to share some of these traits with fractional
processes [7–13]. In recent years, significant developments have been made in the application of fractional
calculus, giving rise to a novel field of study. Various fractional operators have emerged in this domain. Notably,
Caputo derivatives have garnered widespread usage owing to their effectiveness in solving real-world problems
[14, 15], with the generalized mathematical tools such as Mittag-Leffler kernels and exponential kernels [16–18].
The utilization of the Atangana-Baleanu derivative was explored and elaborated, offering advantages in
simulating real-world scenarios [19]. As evidence, there has been growing interest in exploring the impact of
memory modeling on chaotic and hyperchaotic systems, which has revealed a new avenue of research [20]. It is
important to note that even slight adjustments to the initial conditions of chaotic systems can potentially
mitigate chaotic or hyperchaotic behavior.
The authors in this investigation delineated a fractional-order chaotic system and its attenuation of certain
intricacies inherent in fractional-order derivatives, in contrast to the alternative formulations [21–24]. The
primary emphasis revolves around Atangana-Baleanu derivative, bifurcations, and Lyapunov analyses,
electronic circuit development and image encryption. Moreover, the tangent hyperbolic memristor nonlinearity
is fused in the 5D chaotic model to enhance the dynamical system complexity [25–27]. To begin with, this study
includes the well-known Adams-Bashforth approach, which is modified for fractional differential equations and
was first used to create phase pictures. The study’s most significant addition is its use of a Numerical Scheme for
the ABC Fractional Derivative (ABC-FD) approach. The novel five-dimensional nonlinear dynamical system
nature is investigated with Eigenvalues and Lyapunov exponents and it is established that the system is chaotic in
nature. Moreover, the systems parameters were investigated with bifurcation analysis for maximum and
minimum in x2. The system’s parameter-based sensitivity is demonstrated in geometrical analysis where the
parameter a is changed and is plotted to illustrate the variation in trajectories, which are essential in the
generation of random numbers and is beneficial for the application in cryptography. In applications, the image
encryption and electronics circuits are presented for our new engendered model. The image encryption is
demonstrated with proper key utilization along with statistical image tests and electronic circuits is developed
with equivalent tangent hyperbolic memristor.

2. Preliminary concepts

The fundamental terminology and background information required for this study is discussed. The most
typical definitions of derivatives and fractional integrals are provided, and it is shown that only a limited number
of characteristics of classical derivatives can be extrapolated to the fractional situation.
Definition 1 [28]: The ABC fractional derivative operator of function y (t ) in the Caputo equation is given by
ABC (a) d t a
ò0
ABC a
0 Dt y (t ) : = y ¢ ( t ) Ea ⎛ (t - t )a ⎞ dt , 0 < a < 1, (1)
1 - a dt ⎝1 - a ⎠
for, y  H1 (0, t ) , t > 0, a  (0, 1].In this expression ABC (a) satisfies the condition
ABC (0) = ABC (1) = 1. H1 (0, t ), t > 0, is a space of square-integrable functions and is defined as: H1 (0, t )
a -a a +a
= {y (s ) Î L2 (0, t )∣y¢ (s ) Î L2 (0, t )} and ABC (a) = .
a
Definition 2: The Mittag-Leffler function (MLF) is as follows:
k=0
tk
Ea (t ) = å G (ak + 1)
, (2)
¥

where G (.) represents the gamma function and we set the order as 0 < a < 1.
Definition 3: Feedback flux controlled memristor is mathematically written as: The memristor emulator is
applied to our novel system which has a current, voltage relation as,
i = W (v0 ) v, . (3)
W (v0 ) v = [a - b tanh (v0 )] v , (4)
t ( v 0) = f ( v 0, v ) = v 0 - v . (5)
here, i and v, represents the current and the voltage for the input and output of the memristor. a , b are the
positive constants and v0 is the inner state variable. t is the integral time constant and W (v0)is the memductance
which is a nonlinear function defined as,

2
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

W (v0 ) v = [a - b tanh (v0)]. (6)

The above expression, implies that the memristor is a passive voltage controlled device [29].

3. The 5D chaotic model

The 5D fractional-order memristor based chaotic system is introduced. In this section, the memristive based
chaotic model is presented with ABC operator of fractional order. That are formulated as a set of five coupled
differential equations as given below:
ABC  a
0 Dt x1 (t ) = a (x2 - x1) + 0.25x3 + g (d + g tan h (x 5 )) x3
ABC  a
0 Dt x 2 (t ) = bx1 - x1 x3 + cx2 - x 4
ABC  a
0 Dt x 3 (t ) = dx1 x2 - ex3
ABC  a
0 Dt x4 (t ) = x1

ABC  a
0 Dt x5 (t ) = x3 (7)

The novel model is five-dimensional coupled differential equation infused with tangent hyperbolic memristor
with a , b, c , dand e are equation parameters and g = 1, d = 1 and g = 0.5 are memristor parameters with
state variables (x i (i = 1 - 5)) and with the initial conditions x1 (0) = 0, x2 (t ) = 0, x3 (0) = 8, x4 (0) = 6
and x5 (0) = 0.

4. Numerical scheme for the Atangana-Baleanu Caputo—fractional derivative (ABC-FD)

The purpose of section is to investigate a novel scheme for the (ABC-FD), which is introduced in [30]:
ABC a
0 D0 x1 (t ) = E (t , x k (t ))
ABC a
0 D0 x2 (t ) = H (t , x k (t ))
ABC a
0 D0 x3 (t ) = P (t , x k (t ))
ABC a
0 D0 x 4 (t ) = Q(t , x k (t ))
ABC a
0 D0 x 5 (t ) = S (t , x k (t )) (8)

Utilizing the theorems of fractional calculus, we have:

( 1 - a) a t
x1 (t ) - x1 (0) =
ABC (a)
E (t , x k (t )) +
G (a) ´ ABC (a) ò0 E (t , x k (t ))(t - t )a- 1 dt

( 1 - a) a t
x2 (t ) - x2 (0) =
ABC (a)
H (t , x k (t )) +
G (a) ´ ABC (a) ò0 H (t , x k (t ))(t - t )a- 1 dt

( 1 - a) a t
x3 (t ) - x3 (0) =
ABC (a)
P (t , x k (t )) +
G (a) ´ ABC (a) ò0 P (t , x k (t ))(t - t )a- 1 dt

( 1 - a) a t
x 4 (t ) - x 4 (0) =
ABC (a)
Q(t , x k (t )) +
G (a) ´ ABC (a) ò0 Q(t , x k (t ))(t - t )a- 1 dt

( 1 - a) a t
x 5 (t ) - x 5 (0) =
ABC (a)
S (t , x k (t )) +
G (a) ´ ABC (a) ò0 S (t , x k (t ))(t - t )a- 1 dt (9)

Where t n + 1, n = 0, 1, 2 ¼ , and reformulated as

( 1 - a)
x1 (tn + 1) - x1 (0) = E (tn, x k (tn))
ABC (a)
a tn + 1
+
ABC (a) ´ G (a) ò0 E (t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
= E (tn, x k (tn))
ABC (a)
a n tk + 1
+
ABC (a) ´ G (a) k = 0
å òt k
E (t , x k (t ))(tn + 1 - t )a - 1 dt

3
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

( 1 - a)
x2 (tn + 1) - x2 (0) = H (tn, x k (tn))
ABC (a)
a tn + 1
+
ABC (a) ´ G (a) ò0 H (t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
= H (tn, x k (tn))
ABC (a)
a n tk + 1
+
ABC (a) ´ G (a) k = 0
å òt k
H (t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
x3 (tn + 1) - x3 (0) = P (tn, x k (tn))
ABC (a)
a tn + 1
+
ABC (a) ´ G (a) ò0 P (t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
= P (tn, x k (tn))
ABC (a)
a n tk + 1
+
ABC (a) ´ G (a) k = 0
å òt k
P (t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
x 4 (tn + 1) - x 4 (0) = Q(tn, x k (tn))
ABC (a)
a tn + 1
+
ABC (a) ´ G (a) ò0 Q(t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
= Q(tn, x k (tn))
ABC (a)
a n tk + 1
+ å
ABC (a) ´ G (a) k = 0 òt k
Q(t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
x 5 (tn + 1) - x 5 (0) = S (tn, x k (tn))
ABC (a)
a tn + 1
+
ABC (a) ´ G (a) ò0 S (t , x k (t ))(tn + 1 - t )a - 1 dt

( 1 - a)
= S (tn, x k (tn))
ABC (a)
a n tk + 1
+ å
ABC (a) ´ G (a) k = 0 òt k
S (t , x k (t ))(tn + 1 - t )a - 1 dt (10)

with [tk, tk + 1], the functions E (tn, xk (tn)), H (tn, xk (tn)) , P (tn, xk (tn)), Q (tn, xk (tn)), and S (tn, xk (tn)) the
approximation can be done using two-step Lagrange polynomial interpolation:
E ( tk , x k ) E (tk - 1, x k - 1)
P1k (t )  (t - tk - 1) - ( t - tk )
h h
H ( tk , x k ) H (tk - 1, x k - 1)
P 2k ( t )  (t - tk - 1) - ( t - tk )
h h
P ( tk , x k ) P (tk - 1, x k - 1)
P 3k ( t )  (t - tk - 1) - ( t - tk )
h h
Q ( tk , x k ) Q(tk - 1, x k - 1)
P4k (t )  (t - tk - 1) - ( t - tk )
h h
S ( tk , x k ) S (tk - 1, x k - 1)
P 5k ( t )  (t - tk - 1) - ( t - tk ) (11)
h h
where, xk (k = 1 - 5). The above approximation can therefore be included and produce.
(1 - a )
x1n + 1 = x10 + ABC (a )
E (t n , x k (tn))
E(t , x ) tk + 1
òtk (t - tk - 1)(tn+ 1 - t )a - 1 dt ⎞
k k
n ⎛
a h
+ ABC (a ) ´ G (a ) å ⎜ E(tk-1, xk-1) tk+1 ⎟
a - 1 dt ⎟
k = 0 ⎜- ò ( t - t k )( t n + 1 - t )
⎝ h tk ⎠

4
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 1. The 2D trajectories of system (14) for a = 1.00; (a) (x1, x2), (b) (x1, x3), (c) (-x2 , -x3): Cap shape and (d) (x1, x2, x3).

5
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 2. The 2D trajectories of system (14) for a = 0.99; (a) (x1, x2), (b) (x1, x3), (c) (-x2 , -x3): Cap shape and (d) (x1, x2, x3).

6
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

(1 - a )
x 2n + 1 = x 2 0 + ABC (a )
H (t n , x k (tn))
H(t , x ) tk + 1
òtk (t - tk - 1)(tn+ 1 - t )a - 1 dt ⎞
k k
⎛ n
a h
+ ABC (a) ´ G(a) å ⎜ ⎟
H(tk - 1, xk - 1) tk + 1 a - 1 dt ⎟
k = 0 ⎜- ò ( t - t k )( t n + 1 - t )
⎝ h tk ⎠
(1 - a )
x 3n + 1 = x 3 0 + ABC (a )
P (t n , x k (tn))
P (t , x ) tk + 1
òtk (t - tk - 1)(tn+ 1 - t )a - 1 dt ⎞
k k
n ⎛
a h
+ ABC (a ) ´ G (a ) å ⎜ P(tk-1, xk-1) tk+1 ⎟
k = 0 ⎜-
h ò (t - tk)(tn + 1 - t )a - 1 dt ⎟
⎝ tk ⎠
(1 - a )
x 4n + 1 = x 4 0 + ABC (a )
Q (t n , x k (tn))
Q (t , x ) tk + 1
òtk (t - tk - 1)(tn+ 1 - t )a - 1 dt ⎞
k k
⎛ n
a h
+ ABC (a) ´ G(a) å ⎜ ⎟
Q (tk - 1, xk - 1) tk + 1 a - 1 dt ⎟
k = 0 ⎜- ò ( t - t k )( t n + 1 - t )
⎝ h tk ⎠
(1 - a )
x 5n + 1 = x 5 0 + ABC (a )
S (t n , x k (tn))
S (t , x ) tk + 1
òtk (t - tk - 1)(tn+ 1 - t )a - 1 dt ⎞
k k
n ⎛
a h
+ ABC (a ) ´ G (a ) å ⎜ S(tk-1, xk-1) tk+1 ⎟ (12)
k = 0 ⎜-
h ò (t - tk)(tn + 1 - t )a - 1 dt ⎟
⎝ tk ⎠
When you solve the integrals on the right, you get the following numerical scheme:

( 1 - a)
x n+ 1 = x 0 + E (tn, x k (tn))
ABC (a)
a n a
+ å ⎛ h E (tk , x k ) ((n + 1 - k)a (n - k + 2 + a)
⎜

ABC (a) k = 0 ⎝ G (a + 2)
- (n - k)a (n - k + 2 + 2a))
haE (tk - 1, x k - 1)
- ((n + 1 - k)a + 1 - (n - k)a (n - k + 1 + a)) ⎞ ⎟ (13)
G (a + 2) ⎠

5. Simulation results

To inspects the effect of the Atangana-Baleanu-Caputo derivative on the way our fractional system behaves
dynamically on equation (7). By examining several 2D phase portraits an illustration of this technique is
presented. The study investigates the fractional derivative with values for different fractional values. The
development of chaos may be observed with these different orders at t = 200 with the interval of simulation is
0.001. The equation (7) with ABC-FD is presented as
ABC a
0 Dt x1 (t ) = a (x2 - x1) + 0.25x3 + g (d + g tan h (x 5 )) x3
ABC a
0 Dt x2 (t ) = bx1 - x1 x3 + cx2 - x 4
ABC a
0 Dt x3 (t ) = dx1 x2 - ex3
ABC a
0 Dt x 4 (t ) = x1
ABC a
0 Dt x 5 (t ) = x3 (14)
When a = 78, b = -16, c = 58, d = 35, e = 15, g = 1, d = 1 and g = 0.5 with initial conditions
[0,0,8,6,0] and a = 1 the 2D trajectories are presented in figures 1(a)–(d) with phase portraits (x1, x2), (x1, x3),
(-x2, -x3) and (x1, x2, x3) respectively. Further, for fractional values with ABC-FD the a = 0.99 is simulated
and its results are presented in figures 2(a)–(d) with same state variables as mentioned above.

6. Analysis and simulation

Here, the focus is on the equilibrium points, Jacobian matrix corresponds to Eigenvalues for stability analysis
and the nature of chaos with Lyapunov exponents for the novel system (14). Further, the parametric analysis is

7
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 3. The chaotic system’s lyapunov exponents with (a) a = 1.00, and (b) a = 0.99.

Table 1. LE and DKY values, according to fractional order a .

a LE1 LE 2 LE3 LE 4 LE5 DKY

1 0.99150 −0.01697 −0.13940 −3.08469 −30.75108 4.7348

0.99 0.62018 −0.05681 −0.15619 −3.97487 −29.81673 4.2738

considered into the account via the bifurcation diagrams for variations in different parameters in a state variable
along with the dissipativity calculation to ensure the chaos and calculated Lyapunov exponents values.

6.1. Equilibria and eigenvalues

The equilibrium points of the system (14) must be identified. For the fractional differential equations in the
model, which is calculated as follows.

0 = a (x2 - x1) + 0.25x3 + g ( + tanh (x 5 )) x3

0 = bx1 - x1 x3 + cx2 - x 4

0 = dx1 x2 - ex3

0 = x1

0 = x3 (15)

With equation (15) the equilibrium points are evaluated for novel chaotic model (14) with parameters a = 78,
b = -16, c = 58, d = 35, e = 15, g = 1, d = 1 and g = 0.5, which are H0 (-0.2992, -0.3012, 0.2094,
-4.76x1017 , -12.6224) . The Jacobian matrix is evaluated for the model which is written as:

8
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 4. Bifurcation diagram maximum and minimum in x2 of system (14); (a) variation with a, (b) variation with parameter b, (c)
variation with parameter c, (d) variation with parameter d, and (e) variation with parameter e .

⎛ -a a 0.5 0 0⎞
⎜ b - 0.103 c 0.2109 - 1 0⎟
J0 = 0.2115 d 0.2019 d - e 0 0⎟ (16)
⎜
⎜ 1 0 0 0 0⎟
⎝ 1 0 1 0 0 ⎠ H0

The Jacobian matrix J0 has Eigenvalues:

(l1, l2, l3, l 4, l5) = ( - 66.08, 48.09, - 15.03, 0.0237, 0) (17)

The Eigenvalues calculated has two negative, one positive and one zero outcome signifying that system is
unstable and has potential to produce random numbers.

6.2. Lyapunov exponents and Kaplan Yorke dimensions

The characterization of chaos takes place whether the system is chaotic in nature or hyperchaotic for non-integer
orders of derivative. Five Lyapunov exponents (LE) were present: LE1,LE 2, LE3, LE4 and LE5. When at least one is
positive, the dynamic behaviors are chaotic, and when there are two, they are described as hyperchaotic. The
outcomes suggest the systems nature is chaotic as the first LE is positive and second LE is small, approximately
approaches to zero.

9
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 4. (Continued.)

10
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 5. (a) Geometrical control with parameter a variation for the generation of different Cap like structures (-x2, -x3 ) and (b)
the inversion symmetry in system (14).

The dimension of the attractor is calculated using Kaplan Yorke dimension (DKY ) [31] and is presented as
LE1 + LE2 + LE3 + LE4
Dky = 4 + (18)
LE5

Table 1 summarizes the LE’s and DKY , for different fractional values. Phase pictures shows the chaotic
behaviors are conspicuous when the order converges to a = 1 compared to the case where a = 0.99. This is
because the positive Lyapunov exponent is greater at a = 1. Figures 3(a)–(b) illustrates the time-based
spectrum of LE’s for a = 1.00 and 0.99.

6.3. Bifurcation analysis and dissipativity

The theory of Bifurcation refers to the analysis of qualitative changes to the state of a system as a parameter is
varied. It can be practically used to dynamical systems, or to steady state system and as a mathematical model it
can be understood at its level best typically applied to continuous coupled models. System (14) is analyzed by
varying parameters a , b, c , d and e one by one keeping one parameter variable and other constant. Firstly, the
parameter a = 78 is varied from values (70 to 140 with interval of 0.1) keeping other parameters b, c , d , e and k
constant over the time interval of (1 to 10 with the interval of 0.001) results in figure 4(a). The figure 4(a) analysis
in state x2 suggests the system shows chaos for the all the values of a. The bifurcation analysis outcome also
suggests with the variation in value a the system is exhibiting coexisting attractors starting from a » 75 to
a = 140 for x2 max. Secondly, the figure 4(b) is for variational analysis of parameter b which shows that for

11
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 6. Amalgamated image encryption’s algorithm.

b [-100, -61) the system is generating periodic attractor and after b [-62, 0] the chaotic model is generating
chaos with coexisting attractors. In figure 4(c) the bifurcation analysis is conducted for the system’s parameter c
which outcomes shows that for c [0,39) the system is periodic in nature, and after that c [40, 60] system is
generating chaos with coexisting attractors. Moreover, the system is analyzed again with respect to parameter d
keeping other parameters constant suggests that the system is all the way chaotic for the value d [10,120] as
illustrated in figure 4(d). Lastly, the state variable x2 is checked for the parameter e with plot presented in
figure 4(e). The variation in e shows that the system is chaotic when e is varied from the values 10 to 28 and after
the e = 29 to e = 70 the system is exhibiting periodic nature.
The dissipativity of the 4D chaotic system is computed with the expression.
¶E ¶H ¶P ¶Q ¶S
divF = + + + + (19)
¶x1 ¶x2 ¶x3 ¶x 4 ¶x 5
here, ψj (xi(t)) j=1–5 and for each j, i = 1–5. The outcome of the equation (19) for model (14) is less than zero which is
−33 evident the system is dissipative and is unstable.

6.4. Geometric control and inversion symmetry

The model (14) is probed with bifurcation analysis and the results shows that the system is rich with parameters
variations, slight change in the parameters changes the geometry of the phase portraits and data generation. The
geometric control is illustrated in figure 5(a) where the parameter a(72, 74, 76, 78, 80) is varied in the intervals of
2, along with the graphical modifications to avoid the trajectories intersection, showing diverse Cap like
structures in the phase diagrams of (-x2, -x3). The dynamical richness of the system (14) is explored. By just
changing the mathematical sign of g from +g to -g the inversion symmetry is generated and is illustrated in
figure 5(b). The richness in the chaotic pseudorandom number is greatly affected with change in parameters and
initial conditions which can be utilized in various applications especially encryption.

7. Chaotic data generated based encryption

The chaotic system that can generate random numbers utilizing encryption are deeply depends upon chaotic
model, its parameters, and initial conditions. It becomes a core part in the field of cryptography, and
cryptanalysis [32–35]. In the following portion the utilization of chaotic data in cryptography is presented and is
divided into two parts.

7.1. Amalgamated image encryption

In this work we utilized the newly coined algorithm called amalgamated image encryption. The central pillar of
this algorithm is the enhancement of the image security by using two keys: (i) starting with the 3D flow chaotic
data and (ii) the plain image itself. The above two keys combine to decrypt the image successfully; if don’t, the

12
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 7. Plain images utilized in encryption.

Figure 8. Baboon I (Encryption of Plain Image I with Plain Image II and decrypted through Plain Image II with chaotic data).

Figure 9. Histogram of plain image I original, encryption and decryption for baboon I.

Table 2. PSNR, NPCR, UAIC and pixel correlation analysis results.

Pixel correlation
Figure Dimensions MSE PSNR NPCR % UAIC %
Original figure Decrypted figure Encrypted figure

Baboon I 256 × 256 0.00001 99.6261 29.889 0.87207 0.87207 0.16880

Baboon II 256 × 256 303319723 27.5657 99.5168 29.889 0.87207 0.63896 0.16880

13
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 10. Pixel correlation of Image I original, encryption and decryption for baboon I.

Figure 11. Encryption of plain image I with plain image II and decrypted through plain image III in baboon II with chaotic data.

Figure 12. Histogram of plain image I original, encryption and decryption for baboon II.

results are unrecognizable and highly distorted image. The amalgamated image encryption algorithm of the
encryption process is illustrated in figure 6 and the Plain Images are shown in figures 7(a)–(c). The effectiveness
of above algorithm is shown by demonstrating both successful and unsuccessful decryption outcomes utilizing
the CV2 module represented as Baboon I and Baboon II respectively. In figure 8, Baboon I presents the
encryption of Plain Image I, that encrypt/decrypted by utilizing the chaotic data and Plain Image II along with
histogram and correlation analysis illustrated in figures 9 and 10 respectively. The decrypted image and the
original image are in good agreement. In tables 2 and 3 the security tests are implemented to further analyzed the
health of encryption and key utilization. However, as depicted in figure 11, when Plain Image I is encrypted with
Plain Image II but decrypted with Plain Image III using chaotic data, unrecognizable and heavily distorted image

14
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 13. Pixel correlation of image I original, encryption and decryption for baboon II.

Figure 14. Equivalent circuit of tangent hyperbolic memristor.

Table 3. Entropy results.

Original in Encrypted in Decrypted in

Figure percent percent percent

Baboon I R 7.69160 7.76458 7.544348

G 7.40504 7.78647 7.649765
B 7.72243 7.75463 7.730009
Baboon II R 7.69160 7.76458 7.88373
G 7.40504 7.78647 7.87892
B 7.72243 7.75463 7.84847

is formed. The histogram and correlation analysis for Baboon II varies a lot as compared to the Baboon I as
shown in figures 12 and 13. It is worth noting that figure 11(c) demonstrates the distortion that occurs when an
incorrect image key is used for decryption.

7.2. Security analysis of image encryption

Further in this work, utilizing multiple metrics such as Shannon’s entropy, histogram, unified average changing
intensity (UACI), number of pixels change rate (NPCR), correlation between pixels and peak signal to noise ratio
(PSNR), conducting tests on methods decryption and encryption of image. The indication of the successful
encryption is the uniformly distributed data in the encrypted images, that can be verified by the histograms. The

15
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 15. Chaotic circuit with tangent hyperbolic memristor.

histograms of figures 9 and 12 are analyzed to evaluate the algorithm. In Baboon I, the decrypted and original
images histograms are agreed well, indicating the successful decryption. However, in Baboon II, the decrypted
and original images histograms are not in good agreement, indicating the unsuccessful decryption due to
incorrect utilization of keys. Furthermore, the mean square error (MSE) is computed between decrypted and
original images for Baboon I and II. The results for Baboon II presented a significant deviation in the decryption
process, suggesting that when the key II is incorrect the decryption fails. Also, computed the PSNR between the
encrypted and original images, and the values are same, indicating no deviation between the decrypted data and
encrypted data. The UACI was computed between the encrypted and original images, and again, no deviation
was observed between the encrypted data and original data. NPCR is utilized to measure the dissimilarity
between decrypted images and encrypted images, and since the decrypted images for Baboons I and II are not in
good agreement, the values are deviated. Shannon’s information entropy is computed for the original,
decrypted, and encrypted images in the red, green, and blue color channels, as shown in table 3.

8. Chaotic system circuit implementation

In this section, we simulate the results of the memristor-based chaotic system (14) via active and passive
electronic components. Equation (14) is considered and set the circuit frequency to 1000 Hz with capacitors
C1 = C2 = C3 = C4 = C5 = 1nF , then we applied Kirchhoff’s law and the connection between the current and

16
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

Figure 16. Numerical simulated and circuital generated results for system (14) for fractional index 1.00; (a)–(b) for (x1, x2 ), (c)–(d) for
(x1, x3 ), and (e)–(f) for (-x2, -x3 ).

voltage of memristor described by equation (6). With given system parameters a = 78, b = -16,
c = 58, d = 35, e = 15, g = 1, d = 1andg = 0.5, the equation (14) in the form of circuital equation can be
written as

17
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

 1 (t) = 78000(x2 - x1) + 250x3 + 1000(1000 + 500 tanh (x 5 )) x3

Dx
 2 (t) = - 16000x1 - 100x1 x3 + 58000x2 - 1000x 4
Dx
 3 (t) = 35000x1 x2 - 15000x3
Dx
 4 (t) = 1000x1
Dx
 5 (t) = 1000x3
Dx (20)
In terms of resistor and capacitor the above equation is written as

 1 (t) = 1 1 1 1 ⎛ R 45
Dx x2 - x1 + x3 + ⎜ + tanh (x 5 ) ⎞ x3
⎟

R1 C1 R2 C1 R3 C1 Rmem C1 ⎝ R32 ⎠
 2 (t) = - 1 x1 - 1 x1 x3 + 1 x2 - 1 x 4
Dx
R4 C2 R5 C 2 R6 C 2 R7 C 2
 3 (t) = 1 1
Dx x1 x2 - x3
R8 C3 R 9 C3
 4 (t) = 1 x1
Dx
R10 C 4
 5 (t) = 1 x3
Dx (21)
R32 C5
With comparison of equations (20) and (21), the values of resistors and capacitors are, R1 = R2 = 12820.512W,
R3 = 799999.999, R 4 = 62499.999, R5 = 99999.999, R 6 = 17241.379, R7 = R10 = 999999.999,
R8 = 2857.142, R8 = 66666.666 with all capacitors are valued Ci = 1nF . The values plugged in designed circuit
illustrated in figures 14 and 15 in the software MULTISIM and simulated. The outcomes of the circuital
generated 2D trajectories are compared with the equation (14) fractional index 1.00 generated 2D phase portraits
in figure 16 and are compared with each other’s, shows an excellent agreement between the generated data from
two different schemes showing chaotic model (14) application in electronics.

9. Conclusions

The main aim of this work is to investigate the effects of chaotic model parameters and fractional order on a
fusion of recently developed chaotic model and tangent hyperbolic memristor. The investigation relied on
Lyapunov spectra, bifurcation diagrams, geometrical control, and phase diagrams to unravel the complex
dynamics. Throughout the study, it became evident that both the system parameters and the order of the
fractional derivative had a profound influence, leading to a rich tapestry of complex behaviors. To gain a deeper
understanding of the system’s nature, an innovative approach called ABC-FD is implemented. This method
allowed us to create phase portraits and gain a clearer understanding of the system’s behavior, thereby enhancing
our comprehension of its inner workings. As the system is sensitive to initial condition, model parameters and
fractional values the random numbers generated by the chaotic system is utilized to manipulate images using an
Amalgamated Image Encryption algorithm. The scrambled images subjected to rigorous evaluation using various
image security assessment methods. This step reaffirmed our belief that the combination of the chaotic system
and the image itself could serve as a potent and reliable encryption key, thereby enhancing the security of
sensitive visual data. The codes and data for re generating and deep understanding of this work are available for
the scientific community on GitHub and Zenodo [36]. Application of novel chaotic model in electronics is
presented. The circuit design and simulation are carried out using resistor, capacitors and operational
amplifiers. The numerical simulated chaotic model results circuital simulated design results are compared
showing an excellent agreement with each other.

Data availability statement

No new data were created or analysed in this study.

ORCID iDs

Muhammad Ali Qureshi https://orcid.org/0009-0001-1672-3073

Najeeb Alam Khan https://orcid.org/0000-0001-6242-0658
Salman Raza https://orcid.org/0000-0002-1880-6796
Sheikh Muhammad Zeeshan Iqbal https://orcid.org/0000-0001-9571-9728

18
Phys. Scr. 99 (2024) 075238 M A Qureshi et al

References
[1] Ahmed E and Elgazzar A 2007 On fractional order differential equations model for nonlocal epidemics Physica A 379 607–14
[2] Dudkowski D et al 2016 Hidden attractors in dynamical systems Phys. Rep. 637 1–50
[3] Lei T et al 2021 Dynamics analysis and synchronous control of fractional-order entanglement symmetrical chaotic systems Symmetry
13 1996
[4] Matouk A et al 2023 Existence of self-excited and hidden attractors in the modified autonomous van der pol-duffing systems
Mathematics 11 591
[5] Rahman Z-A S et al 2021 High-security image encryption based on a novel simple fractional-order memristive chaotic system with a
single unstable equilibrium point Electronics 10 3130
[6] Xu G et al 2018 A new chaotic system with a self-excited attractor: entropy measurement, signal encryption, and parameter estimation
Entropy 20 86
[7] Monje C A et al 2010 Fractional-Order Systems and Controls: Fundamentals and Applications (Springer Science & Business Media)
[8] Caponetto R 2010 Fractional Order Systems: Modeling And Control Applications 72 (World Scientific)
[9] Alqahtani Z, El-Shahed M and Mottram N 2019 Derivative-order-dependent stability and transient behaviour in a predator–prey
system of fractional differential equations Letters in Biomathematics 6 32–49
[10] Ahmed I et al 2020 Analysis of caputo fractional-order model for COVID-19 with lockdown Advances in Difference Equations 2020 394
[11] Tuan N H, Mohammadi H and Rezapour S 2020 A mathematical model for COVID-19 transmission by using the Caputo fractional
derivative Chaos, Solitons Fractals 140 110107
[12] Diethelm K 2013 A fractional calculus based model for the simulation of an outbreak of dengue fever Nonlinear Dyn. 71 613–9
[13] Khan N et al 2023 Probing 3D chaotic thomas’ cyclically attractor with multimedia encryption and electronic circuitry Arch. Control
Sci. 33 239
[14] Podlubny I 1999 Fractional Differential Equations, Mathematics In Science And Engineering (Academic press New York)
[15] Kilbas A A, Srivastava H M and Trujillo J J 2006 Theory and Applications of Fractional Differential Equations 204 (elsevier)
[16] Gómez-Aguilar J F et al 2017 Chaos in a cancer model via fractional derivatives with exponential decay and mittag-leffler law Entropy
19 681
[17] Atangana A and Baleanu D 2016 New fractional derivatives with nonlocal and non-singular kernel: theory and application to heat
transfer model Thermal Science 20 763–9
[18] Caputo M and Fabrizio M 2015 A new definition of fractional derivative without singular kernel Progress in Fractional Differentiation &
Applications 1 73–85
[19] Uçar S et al 2019 Mathematical analysis and numerical simulation for a smoking model with Atangana–Baleanu derivative. Chaos,
Solitons Fractals 118 300–6
[20] Atangana A and Gómez-Aguilar J 2017 Hyperchaotic behaviour obtained via a nonlocal operator with exponential decay and mittag-
leffler laws Chaos, Solitons Fractals 102 285–94
[21] Rajagopal K, Karthikeyan A and Duraisamy P 2017 Hyperchaotic chameleon: fractional order FPGA implementation Complexity 2017
8979408
[22] Diouf M and Sene N 2020 Analysis of the financial chaotic model with the fractional derivative operator Complexity 2020 1–14
[23] Chen W-C 2008 Nonlinear dynamics and chaos in a fractional-order financial system Chaos, Solitons Fractals 36 1305–14
[24] Sene N 2021 Analysis of a fractional-order chaotic system in the context of the Caputo fractional derivative via bifurcation and
lyapunov exponents Journal of King Saud University-Science 33 101275
[25] Khan N A et al 2022 From chaos to encryption using fractional order lorenz-stenflo model with flux-controlled feedback memristor
Phys. Scr. 98 014002
[26] Khan N A, Qureshi M A and Khan N A 2023 Evolving tangent hyperbolic memristor based 6D chaotic model with fractional order
derivative: analysis and applications Partial Differential Equations in Applied Mathematics 7 100505
[27] Qureshi M A 2023 Bird-like trajectories in 6D chaotic system integrated with fractional order Derivative, memristor, and encryption
Int. Conf. on Nonlinear Dynamics and Applications (Springer)
[28] Atangana A and Koca I 2016 Chaos in a simple nonlinear system with atangana–baleanu derivatives with fractional order Chaos,
Solitons Fractals 89 447–54
[29] Bao B et al 2017 Coexisting behaviors of asymmetric attractors in hyperbolic-type memristor based Hopfield neural network Front.
Comput. Neurosci. 11 81
[30] Toufik M and Atangana A 2017 New numerical approximation of fractional derivative with non-local and non-singular kernel:
application to chaotic models The European Physical Journal Plus 132 1–16
[31] Khan N A et al 2020 Emulate the chaotic flows of fractional jerk system to scramble the sound and image memo with circuit execution
Phys. Scr. 95 065217
[32] Khan N A et al 2020 Behavioral effects of a four-wing attractor with circuit realization: a cryptographic perspective on immersion
Commun. Theor. Phys. 72 125004
[33] Qureshi M 2021 Encryption-Python-Codes-0: Release of Video Encryption in Python (Zenodo)
[34] Qureshi M 2020 Encryption-Python-codes: Release of Voice and Image Encryption in Python (Zenodo)
[35] Khan N A et al 2021 Stumped nature hyperjerk system with fractional order and exponential nonlinearity: Analog simulation,
bifurcation analysis and cryptographic applications Integration 79 73–93
[36] Qureshi M A 2022 Amalgamated Image Encryption, Circuit Realization for Lorenz-Stenflo Chaotic System doi.org/10.5281/
zenodo.7046557 (Zenodo)

19