1: %\documentclass[showpacs,preprintnumbers,amssymb]{revtex4}
2: \documentclass[twocolumn,showpacs,preprintnumbers,amssymb]{revtex4}
3: \usepackage{graphicx}
4: \usepackage{dcolumn}
5: \usepackage{bm}
6:
7: \begin{document}
8: \preprint{CCOC-02-01}
9: \title{Periodic Phase Synchronization in coupled chaotic oscillators}
10: \author{Won-Ho Kye$^1$}
11: \email{whkyes@empal.com}
12: \author{Dae-Sic Lee$^{1,2}$}
13: \author{Sunghwan Rim$^1$}
14: \author{Chil-Min Kim$^1$}
15: \email{chmkim@mail.paichai.ac.kr}
16: \author{Young-Jai Park$^2$}
17: \affiliation{$^1$National Creative Research Initiative Center for Controlling Optical Chaos,\\
18: Pai-Chai University, Daejeon 302-735, Korea}
19: \affiliation{$^2$Department of Physics, Sogang University, Seoul 121-742, Korea}
20: \begin{abstract}
21: We investigate the characteristics of temporal phase locking states
22: observed in the route to phase synchronization.
23: It is found that before phase synchronization there is a periodic phase synchronization
24: state characterized by periodic appearance of temporal phase locking state and that
25: the state leads to local negativeness in one of the vanishing Lyapunov exponents.
26: By taking a statistical measure, we present the evidences of the phenomenon
27: in unidirectionally and mutually coupled chaotic oscillators, respectively.
28: And it is qualitatively discussed that the phenomenon is described
29: by a nonuniform oscillator model in the presence of noise.
30: \end{abstract}
31: %\narrowtext
32:
33: \pacs{05.45.Xt, 05.45.Pq}
34: \maketitle
35: %{\Large \bf Introduction}
36: Over the past decade, synchronization in chaotic oscillators \cite{SyncOrg0,SyncOrg1}
37: has received much attention because of its fundamental importance
38: in nonlinear dynamics and potential
39: applications to laser dynamics \cite{PSExp}, electronic circuits \cite{Circuit},
40: chemical and biological systems \cite{NeuronSync},
41: and secure communications \cite{Secure}.
42: Synchronization in chaotic oscillators is characterized by the loss of exponential
43: instability in the transverse direction
44: through interaction.
45: In coupled chaotic oscillators, it is known, various types of synchronization
46: are possible to observe, among which are complete synchronization (CS) \cite{SyncOrg0,SyncOrg1},
47: phase synchronization (PS) \cite{PhaseSync,Lee},
48: lag synchronization (LS) \cite{Lag}
49: and generalized synchronization (GS) \cite{GenSync}.
50:
51: One of the noteworthy synchronization phenomena in this regard is PS
52: which is defined by the phase locking between nonidentical chaotic oscillators
53: whose amplitudes remain chaotic and uncorrelated with each other:
54: %\begin{equation}
55: $|\theta_1-\theta_2| \leq \mbox{const}.$
56: %\end{equation}
57: Since the first observation of PS in mutually coupled chaotic oscillators \cite{PhaseSync},
58: there have been extensive studies in theory \cite{Lee}
59: and experiments \cite{PSExp}.
60: The most interesting recent development in this regard is the report that
61: the interdependence between physiological systems
62: is represented by PS and temporary phase-locking (TPL) states, e.g.,
63: (a) human heart beat and respiration \cite{HeartBeat},
64: (b) a certain brain area and the tremor activity \cite{BrainMuscle},
65: etc \cite{Visual, Wavelet}.
66: Application of the concept of PS in these areas sheds light
67: on the analysis of nonstationary bivariate data coming
68: from biological systems which was thought to be impossible
69: in the conventional statistical approach.
70: And this calls new attention to the PS phenomenon.
71:
72: Accordingly, it is quite important to elucidate a detailed
73: transition route to PS
74: in consideration of the recent observation of a TPL state in
75: biological systems.
76: What is known at present is that TPL\cite{Lee} transits to PS
77: and then transits to LS as the coupling strength increases.
78: On the other hand, it is noticeable that the phenomenon from
79: nonsynchronization to PS have hardly been studied,
80: in contrast to the wide observations of the TPL states in
81: the biological systems.
82:
83:
84: The chief goal of this Letter is to study the characteristics of
85: TPL states observed in the regime
86: from nonsynchronization to PS in coupled chaotic oscillators.
87: We report that there exists a special locking regime in which
88: a TPL state shows maximal periodicity,
89: which phenomenon we would call {\it periodic phase synchronization} (PPS).
90: We show this PPS state leads to local negativeness
91: in one of the vanishing Lyapunov exponents, taking the measure by which
92: we can identify
93: the maximal periodicity in a TPL state.
94: We present a qualitative explanation of the phenomenon
95: with a nonuniform oscillator model in the presence of noise.
96:
97: We consider here the unidirectionally coupled non-identical Ro\"ssler oscillators
98: for first example:
99: \begin{eqnarray}
100: \dot{x}_{1}&=&-\omega_{1} y_{1} -z_{1},\nonumber\\
101: \dot{y}_{1}&=& \omega_{1} x_{1} +0.15 y_{1},\nonumber\\
102: \dot{z}_{1}&=& 0.2 +z_{1}(x_{1} -10.0), \nonumber\\
103: \dot{x}_{2}&=&-\omega_{2} y_{2} -z_{2},\nonumber\\
104: \dot{y}_{2}&=& \omega_{2} x_{2} +0.165 y_{2}+\epsilon(y_{1}-y_{2}),\nonumber\\
105: \dot{z}_{2}&=& 0.2 +z_{2}(x_{2} -10.0),
106: \end{eqnarray}
107: where the subscripts imply the oscillators 1 and 2, respectively,
108: $\omega_{1,2}$ $(=1.0 \pm 0.015)$ is the overall
109: frequency of each oscillator, and $\epsilon$ is the coupling strength.
110: It is known that PS appears in the regime $\epsilon \geq \epsilon_c$
111: and that $2\pi$ phase jumps arise when $\epsilon < \epsilon_c$.
112: Lyapunov exponents play an essential role in the investigation of
113: the transition phenomenon with coupled chaotic oscillators
114: and as generally understood that PS transition
115: is closely related to the transition to the negative value
116: in one of the vanishing Lyapunov exponents \cite{SyncOrg1}.
117:
118:
119: Figure 1 shows two largest conditional Lyapunov
120: exponents from Eq. (1) according to the coupling strength $\epsilon$.
121: One can see the dip characterized by the local negativeness
122: in the vanishing Lyapunov exponent.
123: The reference points A and C indicate the borders of the dip and
124: B its center.
125: PS transition occurs in the right of C ($\epsilon=0.085$)
126: where the phase difference
127: of coupled chaotic oscillators is bounded within a constant.
128: The temporal behaviors at the three reference points are presented in Fig. 2.
129: The phase jumps at B
130: looks quite regular, compared to those of A and C.
131: Though this observation may seem rather intuitive,
132: we shall see that it is a valid one and that all the phenomenon
133: is deeply related to the dip, the local negativeness
134: in the vanishing Lyapunov exponent of Fig. 1.
135:
136:
137:
138: \begin{figure}
139: \begin{center}
140: \rotatebox[origin=c]{0}{\includegraphics[width=8.5cm]{./uni_lyap.eps}}
141: \caption{Two largest conditional Lyapunov exponents and
142: the dip in coupled R\"ossler oscillators when
143: $\omega_0=1.0$:
144: A ($\epsilon=0.058$), B($\epsilon=0.072$), and C ($\epsilon=0.083$) are the reference
145: points to explain different characteristics of phase dynamics.}
146: \end{center}
147: \end{figure}
148: A vanishing Lyapunov exponent corresponds to a phase variable of an oscillator
149: and it exhibits the neutrality of an oscillator in the phase direction.
150: Accordingly, the local negativeness of an exponent
151: indicates this neutrality is locally broken \cite{PhaseSync}.
152: It is important to define an appropriate phase variable
153: in order to study the TPL state more thoroughly.
154: In this regard, several methods have been proposed
155: methods of using linear interpolation at a Poincar\'e
156: section \cite{PhaseSync},
157: phase space projection \cite{PhaseSync,Lee},
158: tracing of the center of rotation in phase space \cite{Spline},
159: Hilbert transformation \cite{PhaseSync,Wavelet}, or wavelet transformation \cite{Wavelet}.
160: Among these we take the method of phase space projection
161: onto the $x_1-y_1$ and $x_2-y_2$ planes
162: with the geometrical relation $\theta_{1,2}=\arctan(y_{1,2}/x_{1,2})$,
163: and obtain phase difference $\phi=\theta_1-\theta_2$.
164:
165: The system of coupled oscillators is said to be in a TPL state
166: (or laminar state) when
167: $\langle \dot{\phi} \rangle < \Lambda_c$
168: where $\langle ... \rangle$ is the running average over appropriate short time scale and
169: $\Lambda_c$ is the cutoff
170: value to define a TPL state.
171: The locking length of the TPL state, $\tau$, is defined by time interval between
172: two adjacent peaks of $\langle \dot{\phi} \rangle$ (see Fig. 2).
173: \begin{figure}
174: \begin{center}
175: \rotatebox[origin=c]{0}{\includegraphics[width=8.5cm]{./uni_tx.eps}}
176: \caption{Temporal behaviors of the phase difference:
177: (a) $\phi$ at each reference point.
178: $\tau$ is the locking length of a TPL state.
179: (b) $\langle \dot{\phi} \rangle$ at reference point C
180: (running averaged over 10 seconds
181: to remove irrelevant fast fluctuations).
182: The cutoff value for TPL a state is $\Lambda_{c}=1.0\times 10^{-4}$.}
183: \end{center}
184: \end{figure}
185: In order to study the characteristics
186: of the locking length $\tau$,
187: we introduce a measure:
188: \begin{eqnarray}
189: P(\epsilon)= \frac{\sqrt{\mbox{var}(\tau)}}{\langle \tau \rangle},
190: \end{eqnarray}
191: which is the ratio between the average value of
192: time lengths of TPL states and their standard deviation.
193: In terminology of stochastic resonance,
194: it can be interpreted as noise-to-signal ratio \cite{Coherence,SR}.
195: The measure would be minimized
196: where the periodicity is maximized in TPL states.
197: \begin{figure}
198: \begin{center}
199: \rotatebox[origin=c]{0}{\includegraphics[width=8.5cm]{./uni_p.eps}}
200: \caption{Measure P as a function of $\epsilon$. Big dots
201: are the reference points of Fig.1 and
202: each point is the average value of 2500 TPL states.}
203: \end{center}
204: \end{figure}
205:
206: Measure $P$ as a function of $\epsilon$ is presented in Fig. 3.
207: The big dots indicate the reference points of Fig. 1.
208: We see that the value of P begins to drop rapidly from reference
209: point A which corresponds to the left border of the dip in Fig. 1.
210: And the value of $P$ is minimized in a broad region around B, the center of the dip.
211: The value rapidly increases
212: after passing reference point C, the right border of the dip.
213: What is interesting here is that
214: in the region from $\epsilon=0.062$ to $0.078$ the periodicity
215: is maximized and corresponds to the central part of the dip.
216: Eventually, the coupled chaotic oscillators develop to PS near $\epsilon=0.085$.
217: The result presented in Fig. 3 leads us to argue that
218: the dip in a vanishing Lyapunov exponent
219: shows the maximal periodicity of TPL states around the minimum of $P$.
220: We call the TPL state inside the dip a PPS state in
221: the sense that the TPL state appears rather periodically
222: than outside the dip.
223:
224: To validate the argument, we explain the phenomenon in simplified
225: dynamics. From Eq. (1), we obtain the equation of motion in terms
226: of phase difference:
227: \begin{eqnarray}
228: \frac{d \phi}{dt}&=&\Delta \omega +A(\theta_1, \theta_2, \epsilon)
229: \sin\phi + \xi(\theta_1, \theta_2, \epsilon),
230: \end{eqnarray}
231: where,
232: \begin{eqnarray}
233: A(\theta_1, \theta_2, \epsilon) &=& (\epsilon+0.15)\cos(\theta_1+\theta_2)
234: -\frac{\epsilon}{2}(\frac{R_1}{R_2}), \nonumber\\
235: \xi(\theta_1, \theta_2, \epsilon) &=&
236: \frac{\epsilon}{2}\frac{R_1}{R_2}\sin(\theta_1+\theta_2)
237: + \frac{z_1}{R_1}\sin(\theta_1)-\frac{z_2}{R_2}\sin(\theta_2) \nonumber \\
238: & & +(\epsilon + 0.015) \cos(\theta_2) \sin(\theta_2).\nonumber
239: \end{eqnarray}
240: Here $\Delta \omega=\omega_1-\omega_2$ and $R_{1,2}= \sqrt{x_{1,2}^2 + y_{1,2}^2}$.
241: And from Eq. (3) we obtain the simplified equation to describe the phase dynamics:
242: ${d \phi }/{dt} = \Delta \omega +\langle A \rangle \sin(\phi) + \xi$,
243: where $\langle A \rangle$ is the time average of $A(\theta_1, \theta_2, \epsilon)$.
244: This is a nonuniform oscillator in the presence of noise
245: where $\xi$ plays a role of effective noise \cite{Strogatz} and
246: the value of $\langle A \rangle$ controls the width of bottleneck (i.e, nonuniformity of the flow).
247: If the bottleneck is wide enough,
248: (i.e., faraway from the saddle-node bifurcation point: $\Delta \omega \gg -\langle A \rangle$),
249: the effective noise hardly contributes to the phase dynamics of the system.
250: So the passage time is wholly governed by the width of the bottleneck as follows:
251: $\langle \tau \rangle \sim 1/\sqrt{\Delta \omega ^2 -\langle A \rangle ^2}
252: \sim 1/\sqrt{\Delta\omega^2 - \epsilon^2/4}$,
253: which is a slowly increasing function of $\epsilon$.
254: In this region while the standard deviation of TPL states is nearly constant
255: (because the widely opened bottlenecks periodically appears and those lead to
256: small standard deviation), the average value of locking length of TPL states is relatively short
257: and the ratio between them is still large.
258: Accordingly, the value of $P(\epsilon)$ slowly decreases
259: in the regime before reference point B in Fig. 3.
260:
261: On the contrary as the bottleneck becomes narrower
262: (i.e., near the saddle-node bifurcation point: $\Delta \omega \geq -\langle A \rangle$)
263: the effective noise begins to perturb the process of bottleneck passage
264: and regular TPL states develop into intermittent ones (see C in Fig. 2 (a)) \cite{Lee,Kye}.
265: It makes the standard deviation increase very rapidly and
266: this trend overpowers that of the average value
267: of locking lengths of the TPL states.
268: For that reason, the value of $P(\epsilon)$ rapidly increases
269: passing the PPS regime in Fig. 3.
270: Thus we understand that the competition between width of bottleneck
271: and amplitude of effective noise produces the crossover
272: at the minimum point of $P(\epsilon)$
273: which shows the maximal periodicity of TPL states.
274:
275: \begin{figure}
276: \begin{center}
277: \rotatebox[origin=c]{0}{\includegraphics[width=8.5cm]{./mut_lyapNP.eps}}
278: \caption{ Lyapunov exponents and measure P in mutually coupled R\"ossler oscillator:
279: $\dot{x}_{1,2}=-\omega_{1,2} y_{1,2} -z_{1,2}+\epsilon(x_{2,1}-x_{1,2}),
280: \dot{y}_{1,2}= \omega_{1,2} x_{1,2} +0.15 y_{1,2},
281: \dot{z}_{1,2}= 0.2 +z_{1,2}(x_{1,2} -10.0)$, where $\omega_{1,2}=1.0\pm0.015$.
282: (a) One can see the dip in one of the vanishing Lyapunov exponents
283: and the reference point D
284: indicates the center of it.
285: (b) The big dot shows the minimum point and it coincides with the point of
286: the center of the dip. Each point is the average value of 12000 TPL states.}
287: \end{center}
288: \end{figure}
289:
290: Rosenblum et al. firstly observed the dip in mutually coupled chaotic oscillators \cite{PhaseSync}.
291: However the origin and the dynamical characteristics of the dip
292: have been left unclarified. We argue that the dip observed in mutually coupled
293: chaotic oscillators has the same origin as observed above in unidirectionally coupled systems.
294: Figure 4 shows the four largest Lyapunov exponents and the value of $P$ according
295: to the coupling strength $\epsilon$.
296: We can see that the minimum region of $P$ from $\epsilon=0.022$ to $0.027$
297: coincides with the central part of the dip (D in the figure)
298: even though the valley in Fig. 4 (b) is not deeper than that of the unidirectionally coupled systems.
299: Thus it is reasonably summed up that a PPS state exists just before the transition
300: to PS and that the dip observed by Rosenblum et al. is the very evidence of a PPS state
301: in mutually coupled chaotic oscillators.
302:
303: Common apprehension is that
304: near the border of synchronization the phase difference in coupled regular
305: oscillators is periodic \cite{PhaseSync}
306: whereas in coupled chaotic oscillators
307: it is irregular \cite{Lee}.
308: On the contrary, we report that the special
309: locking regime exhibiting the maximal periodicity of a TPL
310: state also exists in the case of coupled chaotic oscillators.
311: In general, the phase difference of coupled chaotic oscillators is
312: described by the one-dimensional Langevin equation:
313: $\dot{\phi}=F(\phi) + \xi$ where $\xi$ is the effective noise with finite amplitude.
314: The investigation with regard to PS transition is the study of scaling of the laminar
315: length around the virtual fixed point $\phi^*$ where $F(\phi^*)=0$ \cite{Kye,Virtual} and
316: PS transition is established when $|\int^{\phi^*}_{\phi} F(\phi) d\phi| > \max|\xi|$.
317: Consequently, the crossover region, from which the value of $P$ grows exponentially
318: (as shown in Fig. 3-4), exists
319: because intermittent series of TPL states with longer locking length $\tau$
320: appears as PS transition is nearer.
321: Eventually it leads to an exponential growth of
322: the standard deviation of the locking length.
323: Thus we argue that PPS is the generic phenomenon
324: mostly observed in coupled chaotic
325: oscillators prior to PS transition.
326:
327: In conclusion, analyzing the dynamic behaviors in coupled chaotic oscillators with
328: slight parameter mismatch we have completed the whole transition route to PS.
329: We find that there exists a special locking regime called PPS in which
330: a TPL state shows maximal periodicity and
331: that the periodicity leads to local negativeness in
332: one of the vanishing Lyapunov exponents.
333: We have also made a qualitative description of this phenomenon
334: with the nonuniform oscillator model in the presence of noise.
335: Investigating the characteristics of TPL states between nonsynchronization and PS,
336: we have clarified the transition route before PS.
337: Since PPS appears in the intermediate regime
338: between non-synchronization and PS,
339: we expect that the concept of PPS can be used as a tool for analyzing
340: weak interdependences, i.e. those not strong enough to develop to PS,
341: between nonstationary bivariate data coming from biological systems, for instance.
342: Moreover PPS could be a possible mechanism of the
343: chaos regularization phenomenon \cite{Regular,Regular1}
344: observed in neurobiological experiments.
345:
346: {\it Note added.} - Recently, we were informed by S. Boccaletti that
347: the phenomenon observed by us was confirmed
348: in CO$_2$ laser systems, experimentally \cite{Bocca}.
349:
350: The authors acknowledge the correspondence with S. Boccaletti, E. Allaria, R. Meucci, and F.T. Arecchi
351: about experimental confirmation of PPS and thank
352: A. Pikovsky, Y. -C. Lai, K. Josi\'c, and M. Choi for helpful comments.
353: This work is supported by Creative Research Initiatives of
354: the Korean Ministry of Science and Technology.
355: \begin{thebibliography}{150}
356: \bibitem{SyncOrg0} H. Fujisaka and T. Yamada, Prog. Theor. Phys. {\bf 69}, 32 (1983);
357: V. S. Afraimovich, N. N. Verichev, and M. I, Rabinovich, Radiophys.
358: Quantum Electron. {\bf 29}, 747 (1986).
359: \bibitem{SyncOrg1} L. M. Pecora and T. L. Carroll, Phys. Rev. Lett. {\bf 64}, 821 (1990)
360: \bibitem{PSExp} D. J. DeShazer, R. Breban, E. Ott, and R. Roy, Phys. Rev. Lett. {\bf 87}, 044101(2001);
361: K. V. Volodchenko, V. N. Ivanov, and S.-H. Gong et al., Optics Letters {\bf 26}, 1406 (2001).
362: \bibitem{Circuit} C. M. Kim, G. S. Yim, J. W. Ryu, and Y. J. Park, Phys. Rev. Lett.
363: {\bf 80}; L. Zhu and Y.-C. Lai, Phys. Rev. E {\bf 64}, 045205 (2001).
364: \bibitem{NeuronSync} R. C. Elson, A. I. Selverston, and R. Huerta et al., Phys. Rev. Lett. 81, 5692 (1998).
365: \bibitem{Secure} L. Kocarev and U. Parlitz, Phys. Rev. Lett. {\bf 74}, 5028 (1995) and references therein.
366: \bibitem{PhaseSync}M. Rosenblum, A. Pikovsky, and J. Kurths, Phys. Rev. Lett. {\bf 76}, 1804 (1996);
367: A. Pikovsky, M. Rosenblum, and J. Kurths, {\it Synchronization A universal concept in
368: nonlinear sciences}, Cambridge University Press 2001.
369: \bibitem{Lee} K.J. Lee, Y. Kwak, and T.K. Lim, Phys. Rev. Lett. {\bf 81}, 321 (1998);
370: E. Rosa, E. Ott, and M.H. Hess, Phys. Rev. Lett. {\bf 80}, 1642 (1998);
371: K. Josi\'c and D.J. Mar, Phys. Rev. E{\bf 64}, 056234 (2001).
372: \bibitem{Lag} M. Rosenblum, A. Pikovsky and J. Kurths, Phys. Rev. Lett. {\bf 78}, 4193 (1997).
373: \bibitem{GenSync} L. Kocarev and U. Parlitz, Phys. Rev. Lett. {\bf 76}, 1816 (1996).
374: \bibitem{HeartBeat} C. Sch\"afer, M.G. Rosenblum, J. Kurths, and H.-H Abel,
375: Nature {\bf 392} 239 (1998);
376: C. Sch\"afer, M.G. Rosenblum, H.-H Abel, and J. Kurths, Phys. Rev. E 60, 857 (1999).
377: \bibitem{BrainMuscle} P. Tass, M. G. Rosenblum, J, Weule, J. Kurths et al., Phys. Rev. Lett. 81, 3291 (1998).
378: \bibitem{Visual}E. Rodriguez, N. George, J. Lachaux, J. Martinerie, B. Renault and F. Varela, Nature, {\bf 397}, 430 (1999).
379: \bibitem{Wavelet} R. Q. Quiroga, A. Kraskov, T. Kreuz, and P. Grassberger, arXiv:nln.CD/0109023.
380: \bibitem{Spline} T. Yalcinkaya and Y. -C. Lai, Phys. Rev. Lett. {\bf 79}, 3885 (1997).
381: \bibitem{Strogatz} S. H. Strogatz, {\it NONLINEAR DYNAMICS AND CHAOS}, PERSEUS BOOKS, 1998.
382: \bibitem{Coherence} A. Pikovsky and J. Kurth, Phys. Rev. Lett. {\bf 78}, 775 (1997).
383: \bibitem{SR} P. Jung, Phys. Reports {\bf 234}, 175 (1993).
384: \bibitem{Kye} W.-H. Kye and C.-M. Kim, Phys. Rev. {\bf E 62}, 6304 (2000).
385: \bibitem{Virtual} W.-H. Kye and D. Topaj, Phys. Rev. {\bf E 63}, 045202(R) (2001).
386: \bibitem{Regular} {\it The Stomatogastric Nervous System}, edited by R.M. Harris-Warrick et al. (MIT Press, Cambridge, MA, 1992);
387: M.I. Rabinovich et al., IEEE Trans. Circuits Syst.-I {\bf 44}, 997 (1997).
388: \bibitem{Regular1} N. F. Rulkov, Phys. Rev. Lett. {\bf 86}, 183 (2001) and references therein.
389: \bibitem{Bocca} S. Boccaletti, E. Allaria, R. Meucci and F. T. Arrecchi,
390: Phys. Rev. Lett. {\bf 89}, 194101 (2002); private communication.
391: \end{thebibliography}
392: \end{document}
393: