Statistic and Dynamic Analysis of Hysteretic System

Carmen Grigoraş1,2, Victor Grigoraş3

1
‘Gr.T. Popa‘ University of Medicine and Pharmacy of Iaşi, 700115, Romania
2
Institute of Computer Science, Romanian Academy, Iaşi, 700481, Romania
3
‘Gheorghe Asachi’ Technical University of Iaşi, 700506, Romania
[email protected]

Abstract — One of the attractive applications of In the referenced paper [1], the a, b, c, and d coefficients are
nonlinear systems is noise generation, based on the chaotic deduced from circuit component values, but in the following
behavior. The present contribution studies an analog analysis they are treated as generic values.
hysteretic system exhibiting chaos and aiming at using the The algebraic nonlinear function sat(.) is given by the input-
unpredictable generated signals for stochastic applications. output characteristic of the nonlinear function:
Using different parameters for the state equations of the
system reduces the possibility of identifying its behaviors. The
statistical analysis presented complements the dynamic results 1 if v  1

better estimating the noise generation using hysteresis. sat (v )  v if v  1 ,
I. INTRODUCTION 1 if v  1
 (2)
The chaotic dynamical behavior of analog nonlinear systems
is studied due to applications in noise generation [1], random The analog nonlinear system described by equations (1) is
number generation [2], secure communication [3], data dissipative, for usual coefficients values, similar to those used in
conversion circuits analysis [4] and data encryption [5]. the referenced paper:
The present paper presents the design of a four-order 4
f n (x)
nonlinear system with hysteresis behavior, based on a f ( x )     a  0  c  d  (3)
previously reported complex dynamical circuit [6]. The n 1 xn
nonlinear dynamic analysis presented and the statistical  0.31  2.23  3.51  0.11  5.94  0
properties of the resulting circuit show high performance in
noise generation. The sensitivity of the analyzed system presented in the
The following section presents the properties of the hysteretic examples in Fig. 1 for initial conditions and Fig. 2 for parameter
nonlinear system based on dynamic analysis. The statistical values. The topological transitivity of the nonlinear system is
properties presented in the third section confirm the noise suggested by the example 3D projections of the strange attractor
generation possible application. The final section draws in Fig. 3.
resulting conclusions and highlights necessary future research
for a more detailed design.
II. DYNAMICS OF THE NONLINEAR SYSTEM
The analyzed analogue nonlinear system, first introduced in
[1], is developed around the model of the Schmitt trigger
hysteretic nonlinearity that leads to complex dynamic behavior.
The system is characterized by analog order four state equations
(1):

dx   a  x  0  y  z Figure 1. State variables errors for 10-4 initial condition diference
 dt
dy
 dt  0  x  z

dz   x  y  0  z  sat (v) (1)
 dt
dv  b  x  c  v  d  sat (v)
 dt
Out (t )  sat (v)

Figure 2. State variables errors for 10-3 parameter a diference

 These dynamic results highlight the chaotic behavior of the The analysis was performed on the state variables depicted in
system, with noise generation possible application. Fig. 4, in comparison with the output signal. The suggestion
offered by the graphical representations reflects the expected
uncorrelated state variables and dependence of the output with
the instantaneous signals value referenced to the statistical
averages.
The first order statistic is reflected by the box plots for the
state variables in Fig. 5 and the histograms shown in figures 6
and 7.

Figure 5. State variables box plot

Figure 3. 3-D projections of the strange attractor

III. STATISTICAL ANALYSIS

In order to obtain the desired noise generator using the
chaotic nonlinear system analyzed, statistical estimations were
performed.

Figure 6. State variable histograms for low 0 = 1.1

Especially histogram suggest the importance of system

parameters related to the probability distribution. Comparing the
histograms for the system with the free oscillating frequency w0
of a lower value, 1.1 in this example, with the case of a higher
value, as 3.1 we may notice reducing the center decreased
values, still keeping large differences to the gaussian
distribution, represented in thicker graph.

Figure 4. State variables and output signal

 Window Length
Figure 7. State variable histograms for higher 0 = 3.1

The second order statistic can be estimated by power spectral

density, that is related to the autocorrelation function, as
presented in the examples in Fig. 8. The obtained results are
showing a relatively flat power distribution in the low frequency
range, below 1 Hz. This confirms the possibility of applying the
analyzed system in generating unpredictable noise signals.

Figure 8. Power spectra for the state variables of the nonlinear system Figure 9. Frequency Kurtosis for the x,y and z state variables

More detailed results can be obtained using the Kurtosis

statistic methods as the order three frequency Kurtosis sown in IV. CONCLUSIONS
Fig. 9 and the fourth order spectral Kurtosis in Fig 10. The detailed analysis performed on the fourth order analogue
Similar suggestions are offered by the multicomponent nonlinear system show, both by dynamic and by statistical
analysis with example results depicted in Fig. 10 means that it is designed with good properties to be applied as
noise generator. The system analysis highlights chaotic
dynamics. The dissipative character of the system was
mathematically demonstrated. Sensitivity to the initial
conditions and parameters as well as topologic transitivity of the
designed systems were also checked by simulation.
The statistical results dependence on the exact system
coefficients shows the necessity of deeper parametric study.
 International Symposium on Circuits and Systems, ISCAS 2007, pp. 213
– 216, 2007.
Spectral Kurtosis x(t) [5] V. Grigoras, C. Grigoras: ‘Digital Data Encryption Based on Discrete

Spectral Kurtosis y(t)

Additive Systems’, in Topics on Chaotic Systems, Selected Papers from
CHAOS 2008 International Conference, World Scientific Publishing Co.,
pp.149-157, 2009.
[6] C. Grigoras, V. Grigoras, “Complex dynamics in hysteretic nonlinear
oscillator circuit”, Proceedings of the Romanian Academy – Series A
Mathematics, Physics, Technical Sciences, Information Science, Vol. 18,
No. 4, pp. 370-377, 2017.

Spectral Kurtosis v(t)

Spectral Kurtosis z(t)

Figure 10. Spectral Kurtosis for the x,y,z and v state variables
Values

Figure 11. Multicomponent analysis for the x,y,z and v state variables

The statistical results confirm the proposed application for

noise and spread spectrum generator, with possible use in
biomedical signal transmission and encryption, analog signal
processing and secure clock architectures.

REFERENCES
[1] T. Endo, J. Yokota, “Generation of White Noise by Using Chaos in
Practical Phase Locked Loop Integrated Circuit Module”, Proceedings of
the International Symposium on Circuits and Systems, ISCAS 2007, pp.
201 – 204, 2007.
[2] H. Nejati, A. Beirami, W. H. Ali, “Discrete-time chaotic-map truly
random number generators: design, implementation, and variability
analysis of the zigzag map”, Analog Integr. Circ. Sig. Process, vol. 73,
no. 1, pp. 363-374, 2012
[3] Cuomo, K.M., Oppenheim, A.V., Strogatz, S. H. “Synchronization of
Lorenz-based chaotic circuits with applications to communications”,
IEEE Transactions on Circuits and Systems II: Analog and Digital Signal
Processing, Volume 40, Issue 10, 1993, pp. 626 – 633
[4] S. Callegari, G. Setti, “ADCs, Chaos and TRNGs: a Generalized View
Exploiting Markov Chain Lumpability Properties”, Proceedings of the