1 Electronically tunable Chaotic Oscillators using OTA-C Derived

2 from Jerk Model

3 Pipat Prommee, Khunanon Karawanich and Montree Kumngern
4 *Department of Telecommunications Engineering,
5 Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520,
6 Thailand
7 Corresponding Email: [email protected] §

8Abstract

9This article presents the new chaotic oscillator circuit using jerk model structure in

10combination with the nonlinear function. The chaotic signal generator is achieved by

11adjusting only one coefficient. The proposed circuit works in current-mode with

12electronically tunability by using Operational Transconductance Amplifiers (OTAs) and

13grounded capacitors without resistors. There are three main components: lossless integrator,

14second-order low-pass filter circuit, and non-linear function. The proposed circuit can adjust

15the signal origin condition through the quality factor of the second order low-pass filter. The

16simulation results are performed by with PSpice and MATLAB programs. The experimental

17results are also included by using the commercially available OTA (LT1228) for confirming

18the simulation results and further integrable. The simulation results of bifurcation graph

19found that the chaos signal behavior had a certainly period doubling totally consistent with

20the experimental results.

21Keywords: Chaotic signal, Jerk model, Bifurcation, Non-linear function, OTA,

22 Electronically tunable

231. Introduction

24 The chaotic system is an interesting dynamic system and useful in interdisciplinary such

25as telecommunications, control system, power system, IT network, transportation, biological

26and not even the financial system. It can be applied as data transmission security purposes.

27The chaotic system contains three key characteristics: non-linear system, dynamic system, and

1
28sensitive to the initial state. There are two types of dynamic systems: linear dynamic system

29and nonlinear dynamic system. For chaotic systems, system changes are in the form of non-

30linear dynamics. According the system changes is depended on the current state or the state-

31variables by exponential and non-constant characteristics. The pattern of changing cannot be

32clearly predicted and also sensitive to various initial conditions. In chaotic system that changes

33in dynamics not linearly, there are three variables for controlling the system by state-variable,

34constant parameter and auxiliary function or nonlinear function for obtaining the chaos

35behavior patterns.

36 In the study of the state-variable structure used to realize chaos system by Edward

37Norton Lorenz called Lorenz's model [1]. Three dimensions (3D) state model as shown in

38equation (1).

x   ( y  x)
y  x  xz  y
39 z  xy   z (1)

40Where  ,  ,   0 are state-variable control parameters. The parameters xz and xy are

41nonlinear terms with state-variable parameters in Lorenz's model causing a non-periodic

42oscillation system. This oscillation values are operated in the certain boundary called attractor

43and values that are displayed in phase space trajectory as shown in Fig1.

44
45 Fig.1 The Lorenz's model chaotic attractor

2
46

47From a Lorenz model that uses nonlinear terms with the multiplication of two state-variables:

48xz and xy which are nonlinear quadratic functions. In 1976, Otto Eberhard Rössler developed

49a simplified state model by reducing the number of multiplications of state-variables [2].

50Only xz was required and control parameters (a, c) are used to control the system. The three

51dimension of state model can be expressed as

x   y  z
y  x  ay
52 z  a  z ( x  c) (2)

53where the control parameters, a=0.2 and c=5.7 are assigned. The variable values from

54Rossler's model are plotted in a 3D axis. The trajectory line on the phase space is shown as

55attractor in Fig. 2. Based on more simplification of Rossler's model by reducing controlled

56parameters and reducing the multiplication of state variables, makes Rössler's model to be

57popular for developing and furthering chaos applications.

58

59 Fig.2 The Rossler's model chaotic attractor

60Later in 1983, the first chaos circuit has coincidentally been discovered in the research room

61by Leon Ong Chua [3-4], as shown in Figure 2.5.

3
 R

Chua¶s
L C2 C1 Diode

62
63 Fig.3 The first chaos circuit invented by Chua
64Figure 3 is a prototype circuit of Leon Ong Chua that can be applied to create a chaos

65oscillator based on Chua’s diode modifications to achieve different nonlinear function

66characteristics. From Chua's circuit, a state variable can be written as a set of 3D state model

67used to create a chaos oscillator as shown in Eq. (3).

x   ( x  y  f ( x))
y  x  y  z
68 z    y (3)

1
f ( x )  Gb  x  (Ga  Gb)  x  E  x  E 
69Where 2 , α and  are controlled parameters.

70Chua's state model can freely create nonlinear functions by defining the characteristics of the

71f(x). The elements of the f(x) can be exhibited in Fig.4 and configured the control parameters

72and nonlinear functions for creating Chua’s chaos signal. The state model can be plotted in

733D phase space as the attractor in Fig.5.

f ( x)

Gb

(Gb  Ga) E

E
Ga x
0
E

(Ga  Gb ) E
Gb

74
75 Fig.4 Characteristic of nonlinear function Chua’s diode
76

4
 y

77
78 Fig.5 Chua’s model chaotic attractor

79Considering between Lorenz's model in 1963 and Chua's model in 1983, chua's model has

80evolved into an easier-to-design state model. Because only two control parameters α and 

81are adjusted, non-linear function elements are freely defined. Lorenz's state model requires

82three control parameters α, λ and  >0 to be adjusted and use nonlinearity from the

83multiplication of state variables (x, y, z). Non-linear function elements cannot be modified

84that makes it limited and difficult to design. Rössler's model could not define the nonlinear

85functions. Only quadratic functions can be defined by the multiplication of state variables.

86Chua's model is therefore popular and is a model of the Chaotic oscillator.

87 In 2000, chaos signal generator based on non-linear function has been described by

88Julien Clinton Sportt's by using state-variable model [5]. The variables were formulated by

89three of first-order equations. Nonlinear system definition in mathematic is a system which

90output changes are not proportionate to input changes. The non-linear systems in biology,

91astronomy, physics, engineering are sophisticated systems, since non-linear systems are

92difficult to solve and cannot be analyzed as accurately as linear systems. The mathematical

93equations described in the nonlinear system are therefore the higher derivative equations or the

94combinations of derivative equations with a multiplication of constant coefficients for

95obtaining the chaotic signal. Nonlinearity is important for periodic or irregular trajectory. This

5
 96causes a unique swing, but the pattern of swing can be determined by nonlinear function

97elements.

98 Many chaos circuits have been invented based on several techniques. The chaotic

99Colpitts oscillator circuit with non-linear resistor was presented by Kennedy [6] who

100discovered that sinusoidal oscillator can generate chaos signal. The chaotic oscillator can be

101realized by using a modify the sinusoidal oscillator with non-linear resistor. The modified

102OPAMP-based Twin-T and Wein-bridge oscillators are the main circuit and replace a linear

103resistor by JFET as a nonlinear resistor and incorporated by an energy storage device (inductor

104and capacitor) [7-8]. The CFOA-based (Current Feedback OPAMP) [9] chaotic oscillator in

105third-order form is also introduced with the same non-linear resistor (JFET) technique. The

106modified Lorenz's model using CFOA was introduced which enables to change the attractor

107by MOS switches [10-11]. Chua’s network incorporated with non-linear resistor based on

108CFOA was introduced to generate chaos signal generator [12]. The Lorenz-like attractor based

109on modified shimizu-morioka model by using many OPAMP and two multipliers was

110implemented [13]. The complex structure was suffered depend on the complex attractor. The

111chaotic oscillator based on jerk model with exponential function by using OPAMPs and diode

112was presented [14]. The jerk model was implemented by single OPAMP and RLC passive

113elements while the nonlinear function was produced by JFET [15]. The chaotic oscillator

114based on simplified Chua’s diode incorporated with Wein-bridge and BPF by using OPAMP

115and was presented [16-17]. The CCCII-based and multi-scrolls attractor had been presented.

116The four non-linear functions were used for achieving multi-scrolls with the high complicate

117structure [18]. It can be seen that a many chaos circuits were invented in variety methods.

118There are non-linear devices that are independently separated which possibly to realize the

119circuits. However, the Lorenz's model, Rossler's model and Chua’s model are relatively high

120complicate structure and excessively parameters to adjust for achieving the chaos signal.

6
121Unfortunately, such circuits have not been able to be electronically adjusted and not suitable

122for integration.

123 Due to not complicate model of jerk function [5], it therefore preferred to create the

124chaos signal generator. The system used only one third derivative equation in the jerk format

125with a coefficient (k), which equation can be described as follows:

d 3x d 2 x dx
k 2   G ( x)
126 dt 3 dt dt (4)

127 Jerk's equation is derived from the displacement mechanisms of x that contain the

 dx   d 2x   d 3x 
   2   3
dt  dt 
128derivatives of x relative to the time as acceleration  dt  , speed  and jerking  .

129Various nonlinear function G(x) can be used in the Jerk format that can be possible, for

130example absolute function x, hyperbolic tangent function (tanh(x)) or signum function

131sgn(x)) [19]. This article presents a chaotic oscillator using the OTA and grounded capacitor

132based on jerk model structure. The OTA-based signum function is preferred to use as non-

133linear function cooperated with the third order differential equation. The circuit can

134electronically adjust the conditions of the G(x) nonlinear function and chaos attractor without

135the affecting to each other. The simulation results of both MATLAB and PSpice are displayed

136comparing with the experimental results.

137

1382. Theory and Principles

139 This section consists of three parts: principle of jerk model, nonlinear function, and

140Bifurcation diagram.

141 2.1. Jerk Model

7
142 Sportt offers a 3D dynamic system based on jerk equation that can show chaotic

143signal. The Jerk equation incorporates with a non-linear function that can independently

144adjust a constant coefficient. The condition of chaotic oscillator can be adjusted from a single

145coefficient (k). It simplifies the application to electronic circuits [14].

146 This article contains the design of a chaotic signal generator with a prototype from the

147Jerk model as described in the equation (4). Only one coefficient is used to control the

148chaotic oscillation condition. The non-linear function G(x), which when modifying this

149nonlinear function, can easily modify the chaotic oscillator form. Jerk's three-dimensional

150nonlinear system can be described with the following ordinary differential equations:

dw 
 x, 
dt

dx 
 y, 
dt 
dy 
 kz  y  Aw  BG ( w), 
151 dt  (5)

152 Where k is a coefficient to adjust the condition of oscillation, A and B are nonlinear

153function constants. In this case, non-linear function G(w) is a signum function [sgn(w)], A=1,

154B=0.3 and the initial conditions

 w0 , x0 , y0    0.6, 0.3, 0  are configured, respectively. From

155(4), jerk diagram can be represented by a typical state variable, as shown in Fig.6. The three

156dimensions state variable can be composed by MATLAB program for jerk equations, written

157as

wn 1  xn , 

xn 1  yn , 
yn 1   kzn  yn  wn  0.3sgn wn , 
158 (6)

8
 z 1 y 1 x 1 w
as bs cs
Integrator1 Integrator2 Integrator3

Gain k
nonlinear
function
 
 v G  w
159

160 Fig.6 Jerk diagram with coefficient k

161 Considering the jerk model in Fig. 6 only x-z states, it can be rewritten as Fig.7. It was

162found that to behave as a biquadratic function in the form of a second order low-pass filter

163(LPF) in equation (3).

x 1 ab
 2
164
w s  sk b  1 ab (7)
165

v  z 1 y 1 x

  as bs
Integrator1 Integrator2
Gain

166
167 Fig.7 Fragmented block diagram of Jerk model
168

169 2.2. Nonlinear Function

170 There are several non-linear functions that can be applied to the Jerk model, as

171discussed in section 1. The nonlinear function used in this article use the Signum function,

9
172that incorporated of symbol and linear functions. In Fig. 8, it is a nonlinear function that

173freely adjusts the condition with variables A and B suitable for chaos, jerk structure.

G  x   Ax  B sgn  x 
A 1 G(x)
B  0.3 0.3

-0.3 0.3
x

-0.3

174
175 Fig. 8 The signum nonlinear function used in jerk model
1
Period doubling : 4-cycle

Period doubling : 2-cycle

3  x 1
xn 0.5

Chaos

x 1
0
0 1 2 3 3.4 3.8 4

176 k
177 Fig. 9 The bifurcation diagram of logistic map equation

178 2.3. Bifurcation diagram

179 Regarding the chaotic oscillator in dynamic system, there are many theoretical effects

180describing chaos such as Poincare map, LE (Lyapunov Exponent), Bifurcation Diagram, etc.

181In this article, the results of the period doubling result in a unique spreading pattern

182(Bifurcation). To prove the chaos behavior, the theory of the logistic map equation in Eq.(8)

183can be plots the bifurcation diagram as Fig.9. The period doubling behaviors are apparent at

184k=3.0, 3.4 and it is evident that chaos behavior is occurred when k>3.8.

10
 xn 1  kxn  1  xn 
185 (6)

186

1873. Realization of Chaotic Oscillator based on OTA

188 In the realization of a chaotic oscillator using the principle in Figure 1, the main

189circuit is a lossless integrator that can be constructed by OTA as shown in Figure 4 and its

190transfer function can be expressed in (6).

Iin
IO1 IO1
Iin gm gm
IO2 C
sC IO2

191

192 Fig.10 Lossless integrator based on OTA

I o1 g m

193
I in sC (7)

Io2 g
 m
194
I in sC (8)

195 From the separate part of block diagram of the Jerk model in Figure 2, it is obviously

196seen that second-order low-pass filter that can independently adjust the quality factor with the

197frequency response. It can be designed using OTA and grounded capacitors, as shown in

198Figure 5, with the current transfer function as equation (8).

11
 IB3

gm3
Iin IB1

C1 gm1 IB2

C2 gm2
Iout
199
200 Fig.11 Current-mode second order low-pass filter based on OTA-C

g m1 g m 2

I out C1C2

I in g  g g
s 2   m 3  s  m1 m 2
201  C1  C1C2
(9)

202 Considering an equation (8) and (3) and if gm1=gm2=gm, C1=C2=C, the relevant

a  b  C g m , and k  g m3 g m
203variables can be summarized to be

IOA
Iin IB

gm5 B·IB
A Iout
gm6
IOB

204
205 Fig.12 Current-mode signum function based on OTA

206 From the diagram block of The Jerk Model Figure 1, a nonlinear function generator

207G(x) is required, which can be realized by 2 OTAs as shown in Figure 6. The OTA 5 acts as a

208current follower, while OTA6 connected to the output of OTA 5, provides the current

209comparator. From Fig.3, the nonlinear function

G ( w)  wn  0.3sgn wn is required. This can

210be done by adjusting the OTA5 and OTA6 bias currents to IB and 0.3IB, respectively. The

211current equation of the Figure 6 circuit can be summarized as

12
212
I OA  I in (10)

213
VA  Vsat sgn( I in ) (11)

214
I OB  0.3I B sgn( I in ) (12)

215
I out  I in  0.3I B sgn( I in ) (13)

216
IB3 IB
gm3 B·IB gm5
IB1 gm6
IB2
C1 gm1
IB4
C2 gm2
IO1 C3 gm4
IO2
IO3
217
218 Fig.13 Current-mode chaotic oscillator based on OTA

219 According to the diagram block of the Jerk model in Fig.1, the sub-circuit in Figs. 4-6

220are incorporated. It is a chaotic oscillator circuit that can adjust the signal conditions and non-

221linear coefficients electronically and independently of each other. Three dimensions chaos

222output signals can be obtained from the current outputs of OTA1, OTA2, and OTA4.

2234. Simulation and Experimental Results

224 To confirm the operation of the proposed current-mode chaotic oscillator, it is divided

225into three parts: the MATLAB bifurcation and chaotic signal simulation, the PSpice

226simulation result of circuit realization and experimental results based on commercially

227available OTA (LT1228) compared with the simulation results.

228 4.1. The bifurcation diagram

229In the results of the experiment to prove chaos behavior, use the equation (5) as an equation.

230The state in 3D (Dimension) is graphed with a twee value. Doubling period in bifurcation

13
231diagram with MATLAB based on the The default values initial condition of the three state

232variables (w0, x0, y0)=(0.6, 0.3, 0), respectively. From Figure 8, it plots the w, x, and y

233positions against the k parameter axis with A=1 and B=0.3. The results of bifurcation are

234demonstrated that the range 0.75-0.8 was the beginning of k when the output would be double

235that called period doubling behavior. The unique value occurs in one interval, from the range

236k is between 0.7 and 0.3. During this period, chaos is clearly displayed.

1.2 1.2
1 1
0.8 0.8
0.6 0.6

w 0.4 x 0.4
0.2 0.2
0 0
-0.2 -0.2
-0.4 -0.4
0.3 0.4 0.5 0.6 0.7 0.8 0.3 0.4 0.5 0.6 0.7 0.8
237 k k
238 (a) (b)
1.2
1
0.8
0.6
0.4
y
0.2
0
-0.2
-0.4
0.3 0.4 0.5 0.6 0.7 0.8
239 k
240 (c)
241 Fig.14 Period doubling behavior of jerk model verified by bifurcation diagram
242

243 4.2. PSpice simulation Results

244In the simulation of this research with the PSpice program, there are three results, second

245order LPF, nonlinear circuit and chaotic oscillator. The simple multi-output CMOS-OTA is

14
246used in a circuit [10] with TSMC 0.25m model and 1V power supplies. The bias currents

247(IB) of all OTAs are given 100A equally except OTA3 and OTA6 and all capacitors are 50nF.

248Figure 15 shows the magnitude response of the second-order low-pass filter circuit. The

249frequency response is obtained around 1kHz when all bias currents are equally assigned. If

250the IB3=kIB is decreasingly adjusted, the quality factor (Q) of the low-frequency filter circuit is

251higher but does not affect the frequency response agreed with the theories. Figure 16 shows

252the simulation result of a signum function circuit in Fig. 12 by assigning the bias current

253IB=100uA and BIB=30A for OTA5 and OTA6, respectively. The output of signum circuit can

254be obtained according to the theory.

20

IB1=IB2=100mA

0
Gain (dB)

-20
IB3=30mA
IB3=50mA
IB3=100mA

-40
10Hz 100Hz 1.0KHz 10KHz
255 Frequency

256 Fig. 15 The frequency response of the second order low-pass filter
40uA

IOA
20uA

Iout
0A

-20uA
IOB
-40uA
-40uA -20uA 0A 20uA 40uA
257 Iin

15
258 Fig. 16 The output current of signum circuit (a) BIB=30A

259 According the bifurcation diagram in Fig.8, the chaos behavior would be apparent in

260the range of k from 0.3 to 0.7. In simulation basis, the bias current of all OTAs are IB=100A

261except OTA3 and OTA6 are manipulated at IB3=50A (k=0.5) and BIB=30A, respectively.

262The outputs IO1, IO2 and IO3 of the circuit in Fig.7 are exhibited in Fig.11. The 3D trajectory

263can be plots as the three attractors in Fig.12 (a) IO2 against IO3, (b) IO1 against IO2 and (a) IO1

264against IO3.

50uA
IO1
0A

-50uA
50uA
IO2
0A

-50uA
100uA
IO3

0A

-100uA
0s 5ms 10ms 15ms 20ms 25ms 30ms
265 time

266 Fig. 17 Three current outputs of proposed chaotic oscillator

70uA 70uA

35uA 35uA

IO3 0A IO1 0A

-35uA -35uA

-70uA -70uA
-70uA -35uA 0A 35uA 70uA -70uA -35uA 0A 35uA 70uA
267 IO2 IO2

268 (a) (b)

16
 70uA

35uA

IO1 0A

-35uA

-70uA
-70uA -35uA 0A 35uA 70uA
269 IO3
270 (c)
271 Fig. 18 Three attractor of proposed chaotic oscillator (a) IO2:IO3 (b) IO2:IO1 (c) IO3:IO1

272To confirm the operation of the chaotic oscillator will conduct experiments by constructing

273off-the-shelf passive and active components in PCB as shown in Fig.13 (a) and the

274measurement of output signals are shown in Figure 13(b). The circuit used in the experiment

275used commercially available OTA (LT1228) and all capacitors are 50nF.

276
277 (a) (b)
278 Fig.19 The chaos circuit implementation based on LT1228 (a) Circuit (b) Experimental Setup

279 The properties of the LPF in Fig.5 can be measured by using Bode 100 Vector

280Network Analyzer (VNA) as shown in Figure 14. The pole-Q of the LPF is increased when

281decreasing bias current of OTA3 (IB3), in agreement with the result of the simulation of Fig.9.

17
 10

Trace 1: Gain Magnitude (dB)

0

kIB=30uA
-10
kIB=50uA
kIB=80uA
-20

-30

-40
10 100 1k 10k
282 Frequency (Hz)

283 Fig.20 The LPF based on LT1228 with manipulating IB3

284

285In signum circuit in Fig. 12, the bias currents IB=100A and BIB=30A are assigned for OTA5

286and OTA6, respectively. Figure 21 shows the experimental result of a signum function circuit.

287The output of signum circuit can be obtained according to the theory and simulation result.

288
289

290

18
291
292

293
294

295
296

19
297

298
299
300
301

302
303 K=0.9

20
304
305 K=0.8

306
307 K=0.7

308
309 K=0.5
310Acknowledgments

311This paper is supported by KMITL fund no. 2563-02-01-024.

21
312References

313[1] E. N. Lorenz, “Deterministic Nonperiodic Flow,” Journal of the Atmospheric

314 Sciences, vol.20, no.2, pp.130-141, 1963.

315[2] O. E. Rössler, “An Equation for Continuous Chaos,” Physics Letters, vol.57A, no. 5,

316 pp. 397-398, 1976.

317[3] L. O. Chua, “Dynamic nonlinear networks: state of the art,” IEEE Trans. Circuits

318 Syst., vol. CAS-27, pp. 1059-1087, 1980.

319[4] T. Matsumoto, “A Chaotic Attractor from Chua’s Circuit,” IEEE Trans. Circuits

320 Syst., Vol. CAS-31, no.12, 1984.

321[5] J. C. Sprott, “A new class of chaotic circuit,” Physics Letters, vol.226A, no. 1, pp. 19-

322 23, 2000.

323[6] M. P. Kennedy, “Chaos in the Colpitts oscillator”, IEEE Trans. Circuit Syst.-I,

324 Vol.41, pp.771–774, 1994.

325[7] A. S. Elwakil, and A. M. Soliman, “Two twin-T based op amp oscillators modified

326 for Chaos”, J. Franklin Inst, Vol.335B, pp.771-787, 1998.

327[8] A. S. Elwakil, and M. P. Kennedy, “A Semi-Systematic Procedure for Producing

328 Chaos from Sinusoidal Oscillators Using Diode-Inductor and FET-Capacitor

329 Composites,” IEEE Trans. Circuit Syst.-I, Vol.47, pp.582-590, 2000.

330[9] A. S. Elwakil, and M. P. Kennedy, “Chaotic oscillators derived from sinusoidal

331 oscillators based on the current feedback opamp,” Analog Integr Circuits Signal

332 Process., Vol.24, pp.239-251, 2000.

333[10] A.S. Elwakil, S. Ozoguz, and M. P. Kennedy, “Creation of a complex butterfly

334 attractor using a novel lorenz-type system,” IEEE Transactions on Circuits and

335 Systems—I: Fundamental Theory and Applications, vol. 49, no. 4, pp. 527–530, Apr.

336 2002.

22
337[11] A. S. Elwakil, S. Ozoguz, and M. P. Kennedy, “A four-wing butterfly attractor from a

338 fully autonomous system,” Int. J. Bifurcation Chaos, vol. 13, no. 10, pp. 3093–3098,

339 2003.

340[12] Tlelo-Cuautle, E., & Duarte-Villaseñor, M. A., “Designing Chua’s circuit from the

341 behavioral to the transistor level of abstraction,” Applied Mathematics and

342 Computation,” vol.184, no.2, pp.715–720, 2007.

343[13] S. Yu, W.K.S. Tang, J. Lu, and G. Chen, “Generation of n x m-wing Lorenz-like

344 attractors from a modified shimizu-morioka model,” IEEE Trans. Circuit Syst.—Part

345 II: Express Briefs, vol. 55, no. 11, pp. 1168–1172, 2008.

346[14] J. C. Sprott, “A New Chaotic Jerk Circuit,” IEEE Trans. Circuit Syst.—Part II:

347 Express Briefs, vol.58, no4, pp.240–243, 2011.

348[15] R. Tchitnga, T. Nguazon, P.H. Louodop Fotso, and J.A.C Gallas, “Chaos in a Single

349 Op-Amp–Based Jerk Circuit: Experiments and Simulations,” IEEE Trans. Circuit

350 Syst.—Part II: Express Briefs, vol.63, no.3, pp.239–243, 2016.

351[16] B. Bao, N. Wang, M. Chen, Q. Xu, and J. Wang, “Inductor-free simplified Chua’s

352 circuit only using two-op-amp-based realization,” Nonlinear Dynamics, vol. 84, no. 2,

353 pp. 511–525, Apr. 2016.

354[17] J. Jin and L. V. Zhao, “Low voltage low power fully integrated chaos generator,” J.

355 Circuits Syst. Comput., vol. 27, no. 10, p. 1850155, Sep. 2018.

356[18] X. Zhang and C. Wang, “A novel multi-attractor period multi-scroll chaotic integrated

357 circuit based on CMOS wide adjustable CCCII,” IEEE Access, vol. 7, pp. 16336–

358 16350, Feb. 2019.

359[19] J. C. Sprott, “Simple chaotic systems and circuits,” American. J. Phys. Vol.68, No.8,

360 pp.758-763, 2000.

361

23