A nonlinear control is proposed for the exponential stabilization and synchronization of chaotic behaviour in a model of Bose-Einstein condensate (BEC). The active control technique is designed based on Lyapunov stability theory and Routh-Hurwitz criteria. The control design approach in both cases guarantees the stability of the controlled states. Whereas the synchronization of two identical BEC in their chaotic states can be realized using the scheme; a suitable controller is also capable of driving the otherwise chaotic oscillation to a stable state which could be expected in practice. The effectiveness of this technique is theoretically and numerically demonstrated.
1. Introduction
Chaotic behaviour is a well-known phenomenon in physics, engineering, biology, and many other scientific disciplines. Recently, it has received much attention [1, 2], following several efforts to analyse and understand the complexity of nonlinear dynamical systems as well as the studies of feedback control systems design.
The control and synchronization of chaotic systems are challenging problems, since a chaotic system is extremely sensitive to small perturbations. Notwithstanding, the possibility of control and synchronization of chaotic systems under certain conditions has been established [1–8]. Due to the connection between control and synchronization, recent studies cast the problem of synchronization in the framework of control theory. In general, the control problem is aimed at stabilizing a chaotic attractor to either a periodic orbit or an equilibrium point, while for synchronization, two dynamical systems cooperate with each other such that the dynamics of the difference varnishes in the course of time. Many potential applications of these phenomena have come true and remarkable experiments have been done in secure communication, laser, and biological systems and other areas [1, 2].
Progress in the research activities on chaos synchronization and stabilization has given birth to various methods such as OGY method [3], PC method [7], active control method [9–13], adaptive control method [14–18], impulsive synchronization technique [19, 20], auxiliary system approach [21], backstepping method [22, 23], hybrid control [24], and sliding mode.
In this light, various linear and nonlinear techniques have been proposed for achieving control and synchronization of a wide variety of physical systems. Examples of these abounds. For instance, in ratchet systems where directed transports play prominent role, synchronization has been proposed as a mechanism for controlling directed transports [10, 11, 22, 25]; in lasers, amplified laser output intensity was reported [26] and in long Josephson junctions, the possibility of controlling rapid flux quantum device and phase-locking phenomenon among others have been demonstrated [14, 26–28].
In this paper, first, we modified the active control technique presented by Lei et al. [12] and extended to nonidentical systems [13, 29] such that the designed controller inputs based on Lyapunov stability theory and Routh-Hurwitz criteria will be used to achieve both the stabilization and synchronization in the Bose-Einstein condensates (BEC). Secondly, the method is applied such that a single control input is sufficient to achieve the stability of the controlled states. Our approach is in contrast to the usual active control technique in which the number of controllers is numerically equal to the dimension of the system. Thus, reducing considerably the controller complexity which is an important factor for practical applications [23, 30].
The Bose-Einstein condensates which we study here has been observed in trapped gases of rubidium, sodium, and lithium [31, 32]. BEC has attracted significant attention due to its potential applications. The creation of BEC has provided a platform for investigating many important phenomena in atomic physics, quantum optics, quantum computation, and so forth. The existence of BEC chaos has been proved and the chaotic properties have also been extensively researched in many previous works (see [33, 34] and references therein). Naturally, chaos, which plays a role in the regularity of the system, will cause instability of the condensate wave function. Despite the great interest in the investigation of BEC chaos, only few recent studies have been devoted to the investigation of chaos control in a BEC. For instance Chong et al. [34] employed the Ott-Grebogi-Yorke (OGY) scheme [3] to examine chaos control in a weakly coupled array of BEC. Similarly, a linear feedback obtained by adjusting the intensity of cold atoms with atomic mirrors was employed to drive the BEC system to regular state [35]. In [36], a parametric modulation of the laser intensity was employed to regulate the BEC dynamics and in particular achieve anticontrol of chaos, in which case periodic states are converted into chaotic states. It is well known that the periodic lattice systems in BEC contain many fantastic properties. For example, quantum computation with BEC atoms in a Mott insulating state is an interesting advancement in applications of the BEC. On the other hand, chaos is associated with quantum entanglement and quantum error correcting which are both the fundamental subjects in quantum computations. Thus, it is valuable to apply or control the chaos in the system.
Recently, vibrational state inversion of a Bose Einstein condensate with optimal control and state tomography was considered in [37], where they presented theoretical and experimental results on high-fidelity transfer of a trapped Bose-Einstein condensate into its first vibrationally excited eigenstate. The excitation is driven by mechanical motion of the trap along a trajectory obtained from optimal control theory. Excellent agreement between theory and experiment is found over a large range of parameters. Similarly, in [38], local exact controllability of a 1D Bose-Einstein condensate in a time-varying infinite square well (box) potential was considered. This is a nonlinear control system in which the state is the wave function of the Bose-Einstein condensate and the control is the length of the box. Here they proved that local exact controllability around the ground state (associated with a fixed length of the box) holds generically with respect to the chemical potential. Vortex synchronization in Bose-Einstein condensates vide a time-dependent Gross-Pitaevskii equation approach was considered in [39]. In the work, they consider vortex lattices in rotating Bose-Einstein condensates composed of two species of bosons having different masses. They solved the full two-component Gross-Pitaevskii equations numerically to verify the stability of the putative locked state that is found to exist within a disc centered on the axis of rotation and that depends on the mass ratio of the two bosons.
In this paper, we propose a nonlinear control scheme to realize both chaotic synchronization and exponential stabilization of BEC. In the next section, we formulate the control theories and in Section 3 give a brief description of the system under consideration. Sections 4 and 5 are concerned with the applications of the control techniques for synchronization and stabilization of BEC, respectively. We conclude the paper in Section 6.
2. Theory2.1. Synchronization Principle
Let us consider a driver chaotic system having the same form as in our previous paper [13] and described by
(1)x˙=Ax+f(x),
and a responding system of different form given by
(2)y˙=By+f(y)+u(t),
where in (1), x(t)∈Rn is ann-dimensional state vector of the system, A∈Rn is a time-periodic matrix for the system parameter, and f:Rn→Rn is the nonlinear part of the system. Similarly in (2), y(t)∈Rn denotes the responding state vector, B∈Rn is a time-periodic matrix for the response system parameter, and u(t)∈Rn is a controller which is to be designed. If A=B, then systems (1) and (2) are identical, and the problem reduces to that of [12]; otherwise they are nonidentical.
The dynamic equation of synchronization error, e=y-x can be expressed as
(3)e˙=Ce+f(y)-f(x)+u(t),
where C=B-A. Thus, the objective of synchronization is to make limt→+∞∥e(t)∥=0.
Our goal is to design a suitable control input u(t) to make the dynamical system (3) asymptotically stable at the origin. The active control function u(t) can be redefined as
(4)u(t)=-De-f(y)+f(x),
where D∈Rn is a constant feedback gain matrix.
Then, the error dynamical system (3) can be rewritten as
(5)e˙=Me,
where M=C-D and M∈Rn.
Theorem 1.
If there exists the feedback matrixDsuch that the eigenvalues of the matrixMare negative real or complex with negative real parts, then the error dynamical system (5) is globally asymptotically stable at the origin, thus implying that the two nonidentical systems (1) and (2) are globally asymptotically synchronized.
Remark 2.
The proof of Theorem 1 follows immediately as in [12].
2.2. Stabilization Principle
Consider the controlled chaotic system given by (1) and written as
(6)x˙=Ax+f(x)+u(t),
where u(t) is a control input. Without loss of generality, let P⊂Rnbe a chaotic bounded set of (6) which is globally attractive, and suppose that x*=0 is an equilibrium point embedded in P. Stabilizing the chaotic orbits in (6) to its equilibrium point x* implies that the limt→+∞∥e(t)∥=0, where e(t)=x-x* or limt→∞∥x(t)∥=0, since x*=0, so that e(t)=x(t). For any n-dimensional chaotic system, we can follow a recursive procedure and obtain the error dynamics
(7)e˙=αx+f(x)+u(t),
where α is the matrix of the error dynamics. The control input u(t) can be redefined as
(8)u(t)=-Dx+f(x),
where D∈Rn is a constant feedback gain matrix. Then, the error dynamical system (7) can be rewritten in the form of (5); that is,
(9)e˙=Nx,
where N=α-D and N∈Rn.
Theorem 3.
If there exists the feedback matrixDsuch that the eigenvalues of the matrixNare negative real or complex with negative real parts, then the error dynamical system (9) is globally asymptotically stable at the origin, and the state variable of (6); that is, x(t)→x* as t→∞.
3. System Description
Here, we consider the BEC system described in [35, 36], which is created in a harmonically trapped 3D potential and then loaded into a moving lattice. In principle, near the center of the magnetic trap, the magnetic potential Pm is much weaker and when the BEC is formed in this region, Pm can be neglected. The quasi-1D potential which plays the main role in the system is valid in the potential region, where kx2+(y2/2)+(z2/4)≤100π, k being the laser wave vector which fixes the velocity of the travelling lattice as VL=δ/2k. By taking into account the effect of damping caused by the incoherent exchange of normal atoms, the system is simply governed by the following Gross-Pitaevskii (GP) equation:
(10)ih(1-iλ)∂ψ∂t1=-ℏ22m∂2ψ∂x2+g0|ψ|2ψ+v1cos2(kξ)ψ,
where m is the atomic mass and g0=4πℏ2a/m represents the interatomic interaction; a being the s-wave scattering length and ψ is the macroscopic quantum equation. Assuming the travelling wave solution ψ=φ(ξ)e(αx+βt1) such that the matter wave is a Bloch-like wave (see [35, 36]).
Inserting ψ into (10) we can turn the partial differential equation (10) into an ordinary differential one as follows (see [35, 36] for full details):
(11)ℏ22md2φdξ2+i(ℏαm+ℏvL-iℏγvL)dφdξ-(ℏβ+ℏ2α2m-iℏβγ)φ-g0|φ|2φ=v1cos2(kξ)φ,
where α and β are undetermined real constants and the term proportional to γ represents the damping effect.
It is convenient to recast equation (11) in dimensionless variables. Now redefining the variables as t=kξ, v=2mvL/ℏk, β1=ℏβ/Er, α1=α/k, and I0=v1/Er, and letting φ=R(t)eiθ(t) and dθ/dt=-β1/v=-(v/2)+α1; then (11) in dimensionless form now reads
(12)R¨=-μvR˙+14v2R+gR13+Iocos2(t)R,
where the over-dots denote the derivative with respect to the re-scaled time, t, Io is optical intensity, and v=2mvl/ℏk. The square of the amplitude R is just the particle number density because |R|=|φ|=|ψ| and θ is the phase of φ.
According to the general theory of the Duffing equation, underlying equation (12), has a monoclinic solution only when the coefficients of the linear (R) and nonlinear (R3) terms on the left-hand side of (12) have opposite signs. Therefore, in order to study the chaos for the negative R term we must consider the case of attractive atom-atom interactions; that is, g<0; (12) is just the parametrically driven Duffing equation with a damping term [36].
Solving (12) numerically using the 4th-order Runge-Kutta algorithms with the following initial conditions: (R,R˙)=(1,0) and the parameters: Io=0.5,μ=0.05,v=2.03, and g=-0.75, the period-1 attractor (single well) is found as shown in Figure 1 and the period-1 attractor (double well) is displayed in Figure 2, where Io=0.7. When the parameters is Io=5.5,μ=0.05,v=2.03, and g=-0.75, the strange chaotic attractor projected in the (R,R˙) phase plane as shown in Figure 3 is obtained and its time series is shown in Figure 4. Notice that Figure 3 is similar to the plots in [35, 36], showing the chaotic state of the BEC. Figure 4 shows the time series of y1 and one can find that the value seems random, but it is different from those noise signals without rules and seems to change following some irregularity. If a different initial condition is selected, the trajectories of system (2) would diverge since the system is deterministic and, thus, unpredictable.
The period-1 attractor (single well) of the BEC with Io=0.5.
The period-1 attractor (double well) of the BEC with Io=0.7.
The chaotic attractor of the BEC.
The time series of the chaotic attractor of the BEC.
4. Synchronization of Chaos in BEC
In this section, the method proposed in Section 2 is applied to synchronize the BEC system described in Section 3. Some global asymptotic conditions are first obtained and then further verified by numerical simulations. The synchronization control technique consists of controlling a condensate in the presence of another condensate by means of suitably designed control inputs that links both systems in a drive-response configuration. Let R=x and R˙=y; then we expressed the BEC system in drive-response configuration as follows:
(13)x˙1=y1,y˙1=14v2x1+gx13+Iocos2(t)x1-μvy1,
for the drive
(14)x˙2=y2+u1(t),y˙2=14v2x2+gx23+Iocos2(t)x2-μvy2+u2(t),
for the response, where ui(i=1,2) are active control inputs to be determined. In order to determine the control functions for system (14), that are required to synchronize it with system (13), we first define the states between the drive-response system as e1=x2-x1 and e2=y2-y1 and then subtract (13) from (14) to obtain the error dynamics system given by
(15)e˙1=e2+u1(t),e˙2=14v2e1+g(x23-x13)Iocos2(t)e1-μve2+u2(t).
For convenience, we redefine the active control functions u1(t) and u2(t) as follows:
(16)u1=v1(t),u2=-g(x23-x13)-I0cos2(t)e1+v2(t).
Hence, the error system (15) becomes
(17)e˙1=e2+v1(t),e˙2=14v2e1-μve2+v2(t).
Equation (17) is the error dynamics, which can be interpreted as a control problem, where the system to be controlled is a linear system with control inputs v1(t) and v2(t) as functions of e1 and e2. As long as these feedbacks stabilize the system, limt→∞∥ei(i=1,2)∥=0. This simply implies that the two BEC systems are synchronized. We can choose the control inputs v1(t) and v2(t) as
(18)(v1(t)v2(t))=D(e1(t)e2(t)),
where D=(abcd) is a 2 × 2 constant feedback matrix. So that the error system (17) can be rewritten as
(19)(e˙1(t)e˙2(t))=C(e1(t)e2(t)),
where C=(a1+bc+(1/4)v2d-μv) is the coefficient matrix.
According to Lyapunov stability theory and Routh-Hurwitz criteria, if
(20)a+d-μv<0,(1+b)(c+14v2)-a(d-μv)<0,
then the eigenvalues of the error system (15) must be real or complex with negative real parts and stable synchronized dynamics between the two identical BEC systems is readily achievable.
The 4th-order Runge-Kutta algorithm was used to obtain the numerical solutions of (13) and (14) with a time grid of 0.001. The parameters of the BEC system were fixed as follows: Io=5.5,μ=0.05,v=2.03, and g=-0.75, which place the system in the chaotic state. The initial conditions are (x1,y1)=(1,0) and (x2,y2)=(0.3,0.6), respectively. For this particular choice of system parameters, it can be readily shown that the matrix D given by D=(00-2.03-0.8985) satisfy the conditions (18) and the eigenvalues of the error system is negative real or complex with negative real parts, thus leading to the synchronization of systems (13) and (14). Finally the control inputs are
(21)u1=0,u2=k0+k1e1+k2e2,
where k0=-g(x23-x13), k1=-2.03 and k2=-I0cos2(t)-0.8985.
With the form of (21) satisfying the stability condition (20), we do not need to compute the conditional Lyapunov exponent (CLE) of the systems: more so the negativity of the CLE is only a necessary but not a sufficient condition for stable synchronization [40]. Note that other choices of the feedback gains (k1,k2) could equally satisfies the conditions (18) and thus lead to synchrony.
The simulation results when the controller (21) is activated are shown in Figure 5. It is obvious that the synchronization error converges to zero asymptotically and the two identical BEC systems indeed achieve chaos synchronization as proposed. In Figures 6 and 7, we display the behaviour of the state variables of the master and the slave systems when the control is activated at t≥30. We see clearly that the slave BEC traces the dynamics of the master BEC. This in particular suggests that the motion of the condensate with a specific dynamics can be driven to follow a desired dynamic state. The idea could be employed to drive a chaotic BEC state to a periodic one and also to achieve the anticontrol (or chaotification) [35].
The asymptotic convergence of the synchronization error system to zero for identical BEC system.
The x1,x2 variables of the synchronized BEC master and slave system.
The y1 and y2 variables of the synchronized BEC master and slave system.
5. Stabilization of Chaos in BEC
Here we propose a recursive active control-to-control the evolution of chaotic state in the BEC system. The goal of the control technique is to drive the BEC system from its current chaotic state (x,y) to a stable regular state (x*,y*), where a desired dynamic is obtained, since the chaotic property in the spatial distribution of BEC could constitute a disadvantage in many applications. The method proposed by [35, 36] relies on the knowledge of the Lyapunov exponent (LE) of the system, which is often time consuming to compute or even difficult in some cases. In our proposed method, the computation of LE is ruled out, while the stability is fully guaranteed using the Lyapunov stability theory and Routh Hurwitz criteria. Now considering the system (22) to be controlled given below
(22)x˙=y+u1(t),y˙=14v2x+gx3+Iocos2(t)x-μvy+u2(t),
where u1 and u2 are control functions to be determined. Starting in a recursive manner, suppose x*=0 is an arbitrarily chosen scalar input corresponding to a stable fixed point or an equilibrium point of the system (22) and y*=c1e1 is recursively introduced; c1 is an arbitrary control parameter. Let the error signal between the current chaotic state and the desired state be given as e1=x-x* and e2=y-y*, respectively. Since (x,y)=(0,0) is a stable equilibrium of the system (22), it is convenient to set the arbitrary control parameter c1=0 so that y*=0, and some simple possible control inputs would drive the system to the state (x*,y*)=(0,0) that could be obtained. This leads to the following error dynamics system:
(23)e˙1=e2+u1(t),e˙2=14v2e1+ge13+Iocos2(t)e1-μve2+u2(t).
Again, we find that the control problem is reduced to that of achieving the asymptotic stabilization of the zero solution of the error dynamics (23). To achieve this, we redefine the active control functions u1(t) and u2(t) as follows:
(24)u1=v1(t),u2=-ge13-Iocos2(t)e1+v2(t).
Hence, the error system (23) becomes
(25)e˙1=e2+v1(t),e˙2=14v2e1-μve2+v2(t).
The error dynamics of (25) now takes the form of (17) derived in the synchronization process of the two BEC systems. As a result we can apply Lyapunov stability theory and Routh-Hurwitz criteria to system (25) to obtain the control functions. Proceeding as in the previous section and using the same system parameters, one can show that the matrix D given by D(t)=(00-2.03-0.8985) satisfies the stability criteria. Finally we obtain the following control inputs:
(26)u1=0,u2=k0+k1x+k2y,
where k0=-gx3, k1=-2.03, and k2=-I0cos2(t)-0.8985. Notice the nonlinear form of the controller (26) which is in contrast to the proposed controller in [36]. This is expected because in practice, due to some physical limitations, the control input usually contains nonlinear term. Significantly, nonlinearity in systems like BEC could lead to serious degradation of the system performance, decrease in speed of response, and possibly may cause chaotic perturbations to original regular behaviour if in the design process of the controller their effects are ignored [41–43]. Numerically, we implement the controller of (26) in (22), using the 4th order Runge-Kutta algorithm. The result is displayed in Figure 8. Clearly, we see that the chaotic dynamic has been driven to the stable equilibrium point, implying that the BEC chaos is globally controlled.
The controlled state of the BEC system.
6. Conclusions
In this paper, the synchronization between two-coupled Bose-Einstein Condensates in their chaotic states have been investigated based on a technique derived from nonlinear control theory. The nonlinear control is obtained using the active control and the controller is chosen such that a single control input is sufficient to guarantee global stability of the synchronized state. The systematic design approach of the control input does not require the calculation of the Lyapunov exponents; hence, it is simple and convenient. Besides the synchronization of two chaotic BEC systems, the method was also used to obtain a nonlinear control that is capable of driving the otherwise chaotic oscillation to a stable equilibrium state which could be expected in applications. The effectiveness and feasibility of this technique are theoretically and numerically demonstrated. We believe that such nonlinear control would find applications in the experimental control of spatial chaos in BEC. It is likely that nonlinear interactions are necessary for the stability of continuously pumped atom lasers as described in [44]. From the physical point of view, the underlying mechanism is based on the Josephson-type coupling and is well known in the solid state physics as well as in the ultracold quantum superfluid atoms.
According to Haine et al. [45] in the case of a Bose-Einstein condensate with a large nonlinear interaction, there is already coupling between different modes of oscillations. This means that each mode of feedback can remove energy from more than one mode of oscillation. This method of extracting energy from the higher modes is quite inefficient. Adding a nonlinear control improves the efficiency of the feedback because it directly removes energy from a larger range of modes.
Acknowledgments
The research work of U. E. Vincent is supported by the Royal Society of London through the Newton International Fellowship Alumni Scheme. The author acknowledges useful discussions with Peter McClintock.
PikovskyA. S.RosenblumM. G.KurthsJ.2001Cambridge, UKCambridge University PressLakshmananM.MuraliK.1996SingaporeWorld ScientificOttE.GrebogiC.YorkeJ. A.Controlling chaos19906411119611992-s2.0-424348955210.1103/PhysRevLett.64.1196FujisakaH.YamadaT.Stability theory of synchronized motion in coupled-oscillator systems1983691324710.1143/PTP.69.32ChenG.DongX.1998SingaporeWorld ScientificLüJ.LuJ.ChenS.2002Wuhan, ChinaWuhan University PressPecoraL. M.CarrollT. L.Synchronization in chaotic systems19906488218242-s2.0-034368990410.1103/PhysRevLett.64.821KocarevL.ParlitzU.Generalized synchronization, predictability, and equivalence of unidirectionally coupled dynamical systems19967611181618192-s2.0-0000929946HoM. C.HungY. C.Synchronization of two different chaotic systems using generalized active network20023015-642442810.1016/S0375-9601(02)00987-8VincentU. E.LaoyeJ. A.Synchronization and control of directed transport in chaotic ratchets via active control20073631-291952-s2.0-3384660374310.1016/j.physleta.2006.10.086VincentU. E.LaoyeJ. A.Synchronization, anti-synchronization and current transports in non-identical chaotic ratchets200738422302402-s2.0-3454803021310.1016/j.physa.2007.05.044LeiY.XuW.ShenJ.FangT.Global synchronization of two parametrically excited systems using active control20062824284362-s2.0-2714444610610.1016/j.chaos.2005.05.043IdowuB. A.VincentU. E.NjahA. N.Synchronization of chaos in non-identical parametrically excited systems2009395232223312-s2.0-6404911048510.1016/j.chaos.2007.06.128GuoR.VincentU. E.IdowuB. A.Synchronization of chaos in RCL-shunted Josephson junction using a simple adaptive controller20097932-s2.0-6614913652010.1088/0031-8949/79/03/035801035801LuJ.WuX.HanX.LüJ.Adaptive feedback synchronization of a unified chaotic system20043294-53273332-s2.0-424310371810.1016/j.physleta.2004.07.024SundarapandianV.Adaptive control and synchronization of hyperchaotic Lu system2011122940SundarapandianV.Adaptive control and synchronization of hyperchaotic Newton-Leipnik system201113223310.5121/ijait.2011.1302LuJ.CaoJ.HoD. W. C.Adaptive stabilization and synchronization for chaotic Lur'e systems with time-varying delay2008555134713562-s2.0-4554909903810.1109/TCSI.2008.916462HeW.QianF.CaoJ.HanQ.Impulsive synchronization of two nonidentical chaotic systems with time-varying delay201137534985042-s2.0-7865086143110.1016/j.physleta.2010.11.052CaoJ.HoD. W. C.YangY.Projective synchronization of a class of delayed chaotic systems via impulsive control200937335312831332-s2.0-6765121166210.1016/j.physleta.2009.06.056HeW.CaoJ.Generalized synchronization of chaotic systems: an auxiliary system approach via matrix measure20091912-s2.0-6384926078810.1063/1.3076397013118VincentU. E.Controlling directed transport in inertia ratchets via adaptive backstepping control20073824592469TanX.ZhangJ.YangY.Synchronizing chaotic systems using backstepping design200316137452-s2.0-003733281110.1016/S0960-0779(02)00153-4LiP.CaoJ.Stabilisation and synchronisation of chaotic systems via hybrid control2007137958012-s2.0-3424837869910.1049/iet-cta:20050528VincentU. E.KenfackA.SenthilkumarD. V.MayerD.KurthsJ.Current reversals and synchronization in coupled ratchets201082452-s2.0-7865129179010.1103/PhysRevE.82.046208046208ParvathiM. R.KrishnaB. M.RajeshS.JohnM. P.NandakumaranV. M.Synchronization and control of chaos in coupled chaotic multimode Nd:YAG lasers20083731961012-s2.0-5644910496510.1016/j.physleta.2008.10.072DanaS. K.RoyP. K.SethiaG. C.SenA.SenguptaD. C.Taming of chaos and synchronisation in RCL-shunted Josephson junctions by external forcing200615354534602-s2.0-3375046271510.1049/ip-cds:20045141FengY. L.ShenK.Chaos synchronization in RCL-shunted Josephson junctions via a common chaos driving20086111051102-s2.0-3884911691510.1140/epjb/e2008-00037-9NjahA. N.Synchronization via active control of identical and non-identical ϕ6 chaotic oscillators with external excitation20093273–53223322-s2.0-6904909692710.1016/j.jsv.2009.07.015CorronN. J.PethelS. D.HopperB. A.Controlling chaos with simple limiters20008417383538382-s2.0-0001553819AndersonM. H.EnsherJ. R.MatthewsM. R.WiemanC. E.CornellE. A.Observation of Bose-Einstein condensation in a dilute atomic vapor199526952211982012-s2.0-11944274056BradleyC. C.SackeuC. A.TolleuJ. J.HuletR. G.Evidence of Bose-Einstein condensation in an atomic gas with attractive interactions19957591687169010.1103/PhysRevLett.75.1687ChongG. S.HaiW. H.XieQ. T.Spatial chaos of trapped Bose-Einstein condensate in one-dimensional weak optical lattice potential20041422172232-s2.0-304255949410.1063/1.1648243ChongG. S.HaiW. H.XieQ. T.Controlling chaos in a weakly coupled array of Bose-Einstein condensates20057112-s2.0-4134910227110.1103/PhysRevE.71.016202016202WangZ. X.ShenK.Anti-control of chaos in Bose-Einstein condensates2008634024062-s2.0-4834910981610.2478/s11534-008-0063-8WangZ. X.ZhangX. H.ShenK.Controlling chaotic behavior in a Bose-Einstein condensate with linear feedback20085012152192-s2.0-5134915211510.1088/0253-6102/50/1/41BuckerR.BerradaT.van FrankS.Vibrational state inversion of a Bose-Einstein condensate: optimal control and state tomography2013461010401210.1088/0953-4075/46/10/104012BeauchardK.LangeH.TeismannH.Local exact controllability of a 1D Bose-Einstein condensate in a time-varying boxhttp://arxiv.org/abs/1303.2713BarnettR.ChenE.RefaelG.Vortex synchronization in Bose-Einstein condensates: a time-dependent Gross-Pitaevskii equation approach2010122-s2.0-7795162465710.1088/1367-2630/12/4/043004043004HeagyJ. F.CarrollT. L.PecoraL. M.Desynchronization by periodic orbits1995522R1253R12562-s2.0-2174444212210.1103/PhysRevE.52.R1253GaoT.ChenZ.YuanZ.Control for the synchronization of Chen system via a single nonlinear input20064329730110.1007/s11768-006-4174-8YanJ. J.Design of robust controllers for uncertain chaotic systems with nonlinear inputs20041935415472-s2.0-004162935410.1016/S0960-0779(03)00123-1HongY.HuangJ.XuY. S.On an output feedback finite-time stabilization problem20014623053092-s2.0-003524914510.1109/9.905699HaineS. A.HopeJ. J.Mode selectivity and stability of continuously pumped atom lasers200368270236072-s2.0-18844468148HaineS. A.FerrisA. J.CloseJ. D.HopeJ. J.Control of an atom laser using feedback200469160136052-s2.0-864429148210.1103/PhysRevA.70.039902