Abstract
This work aims to propose and design a class of networks of coupled linear and nonlinear oscillators, in which short bursts of exogenous excitation result in sustained endogenous network activity that returns to a quiescent state only after a characteristic time and along a different path than when originally excited. The desired hysteretic behavior is obtained through the coupling of self-excited oscillations with purposely designed rate laws for slowly varying nodal parameters, governed only by local interactions in the network. The proposed architecture and the sought dynamics take inspiration from complex biological systems that combine endogenous energy sources with a paradigm for distributed sensing and information processing. In this paper, the network design problem considers arbitrary topologies and investigates the dependence of the desired response on model parameters, as well as on the placement of a single nonlinear node in an otherwise linear network. Perturbation analysis in various asymptotic parameter limits is used to define the proposed internal dynamics. Parameter continuation techniques validate the asymptotic results numerically and demonstrate their robustness over finite ranges of parameter values. Both approaches suggest a nontrivial dependence of the optimal distribution of nonlinearity on the network topology.
Similar content being viewed by others
Data availability
Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.
References
Carroll, R.E., Little, J.A., Mann, B.P., Arnold, D.P.: Demonstration of tunable energy propagation using magneto-mechanical oscillator arrays. In: 2017 IEEE 30th International Conference on Micro Electro Mechanical Systems (MEMS), pp. 873–876 (2017)
Chen, J.Y., Pandurangan, G.: Almost-optimal gossip-based aggregate computation. SIAM J. Comput. 41(3), 455–483 (2012)
Crall, J.D., Gravish, N., Mountcastle, A.M., Kocher, S.D., Oppenheimer, R.L., Pierce, N.E., Combes, S.A.: Spatial fidelity of workers predicts collective response to disturbance in a social insect. Nat. Commun. 9(1), 1–13 (2018)
Dankowicz, H., Schilder, F.: Recipes for Continuation. Society for Industrial and Applied Mathematics, Philadelphia (2013)
Defontaines, A.D., Pomeau, Y., Rostand, B.: Chain of coupled bistable oscillators: a model. Physica D 46(2), 201–216 (1990)
Dombovari, Z., Munoa, J., Kuske, R., Stepan, G.: Milling stability for slowly varying parameters. Proc. CIRP 77, 110–113 (2018)
Duarte, A., Weissing, F.J., Pen, I., Keller, L.: An evolutionary perspective on self-organized division of labor in social insects. Annu. Rev. Ecol. Evol. Syst. 42, 91–110 (2011)
Engler, H., Kaper, H.G., Kaper, T.J., Vo, T.: Dynamical systems analysis of the Maasch–Saltzman model for glacial cycles. Physica D 359, 1–20 (2017)
Gallage, R.S.: Approximation of continuously distributed delay differential equations. Master’s thesis, Southern Illinois University at Carbondale (2017)
Huang, K., Hossein-Zadeh, M.: Detection and sensing using coupled oscillatory systems at the synchronization edge. IEEE Sens. J. 20(21), 12992–12998 (2020)
Johnson, B.R., Scott, S.K., Thompson, B.W.: Modelling complex transient oscillations for the BZ reaction in a batch reactor. Chaos Interdiscip. J. Nonlinear Sci. 7(2), 350–358 (1997)
Kasai, Y., Yabuno, H., Yamamoto, Y., Matsumoto, S.: Ultra-sensitive minute mass sensing using a microcantilever virtually coupled with a virtual cantilever. Sensors 20(7), 1823 (2020)
Khasawneh, F.A., Mann, B.P.: Stability of delay integro-differential equations using a spectral element method. Math. Comput. Model. 54(9–10), 2493–2503 (2011)
Lulec, S.Z., Adiyan, U., Yaralioglu, G.G., Leblebici, Y., Urey, H.: MEMS cantilever sensor array oscillators: theory and experiments. Sens. Actuat. A 237, 147–154 (2016)
Manav, M., Phani, A.S., Cretu, E.: Mode localization and sensitivity in weakly coupled resonators. IEEE Sens. J. 19(8), 2999–3007 (2019)
Mao, Y., Dankowicz, H.: Topology-dependent excitation response of networks of linear and nonlinear oscillators. J. Comput. Nonlinear Dyn. 16(4), 041001 (2021)
Marquez, S., Alvarez, M., Plaza, J.A., Villanueva, L.G., Dominguez, C., Lechuga, L.M.: Asymmetrically coupled resonators for mass sensing. Appl. Phys. Lett. 111(11), 113101 (2017)
Masuda, A., Inoue, M., Fujiwara, Y.: Self-powered oscillation of piezoelectric ultrasonic active sensors. In: Sensors and Smart Structures Technologies for Civil. Mechanical, and Aerospace Systems 2020, vol. 11379, pp. 136–145. International Society for Optics and Photonics, SPIE (2020)
Nayfeh, A.H., Mook, D.T.: Nonlinear Oscillations. Wiley, New York (1995)
Pyles, C.S., Bajaj, N., Rhoads, J.F., Weinstein, D., Quinn, D.D.: Threshold color sensing using coupled oscillator networks. In: 2019 IEEE Sensors, pp. 1–4 (2019)
Roongthumskul, Y., Fredrickson-Hemsing, L., Kao, A., Bozovic, D.: Multiple-timescale dynamics underlying spontaneous oscillations of saccular hair bundles. Biophys. J . 101(3), 603–610 (2011)
Saghafi, M., Dankowicz, H., Tabor, W.: Emergent task differentiation on network filters. SIAM J. Appl. Dyn. Syst. 16(3), 1686–1709 (2017)
Spletzer, M., Raman, A., Wu, A.Q., Xu, X., Reifenberger, R.: Ultrasensitive mass sensing using mode localization in coupled microcantilevers. Appl. Phys. Lett. 88(25), 254102 (2006)
Storti, D.W., Rand, R.H.: Dynamics of two strongly coupled van der pol oscillators. Int. J. Non-Linear Mech. 17(3), 143–152 (1982)
Urasaki, S., Yabuno, H., Yamamoto, Y., Matsumoto, S.: Sensorless self-excited vibrational viscometer with two hopf bifurcations based on a piezoelectric device. Sensors 21(4), 1127 (2021)
Wang, C.Q., Pumir, A., Garnier, N.B., Liu, Z.H.: Explosive synchronization enhances selectivity: example of the cochlea. Front. Phys. 12(5), 1–9 (2017)
Wood, G.S., Zhao, C., Pu, S.H., Boden, S.A., Sari, I., Kraft, M.: Mass sensor utilising the mode-localisation effect in an electrostatically-coupled MEMS resonator pair fabricated using an SOI process. Microelectron. Eng. 159, 169–173 (2016)
Yashin, V.V., Levitan, S.P., Balazs, A.C.: Coupling mechanics to chemical reactions to create “materials that compute” In: Mechanochemistry in Materials, pp. 167–193. The Royal Society of Chemistry (2018)
Zhang, L., Niknejad, A.M.: Design and analysis of a microwave-optical dual modality biomolecular sensing platform. IEEE J. Solid-State Circuits 55(3), 639–649 (2020)
Acknowledgements
This material is based upon work supported by the National Science Foundation under Grant No. BCS-1246920. Part of the editing of this paper was performed while the second author served at the National Science Foundation. Any opinion, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: Nomenclature
- \(\delta \) :
-
A small offset for damping dynamics
- \(\epsilon \) :
-
Scaling parameter of linear and nonlinear damping in the network model
- \(\eta \) :
-
Nonlinear damping parameter of the nonlinear oscillator in the network model
- \(\mu \) :
-
Damping parameter for linear oscillator at each node when \(\zeta _k=\mu \) for all k
- \(\nu \) :
-
Linear damping parameter of the nonlinear oscillator in the network model
- \(\omega _k\) :
-
Natural frequency of k-th mode in the network model
- \(\phi (\epsilon t)\) :
-
Phase of \(\mathcal {O}(1)\) solution in multiple-scale analysis
- \(\tau \) :
-
A large time scale for damping dynamics
- \(\tilde{\zeta }\) :
-
A damping parameter in bifurcation analysis, \(\tilde{\zeta }=\sum _{k\ne Q}\frac{K_{Q,k}K_{k,Q}}{\epsilon ^2\zeta _k K_{Q,Q}}\)
- \(\tilde{C}(x)\) :
-
Modal damping matrix
- \(\tilde{K}\) :
-
Modal stiffness matrix
- \(\zeta \) :
-
A damping parameter in bifurcation analysis, \(\zeta =\sum _{k\ne Q}P^2_{k,I}\zeta _k\)
- \(\zeta _k\) :
-
Damping parameter for linear oscillator at node k in the network model
- \(\zeta _{\text {HB}}\), \(\tilde{\zeta }_{\text {HB}}\), \(\mu _{\text {HB}}\):
-
Damping parameter at Hopf bifurcation
- \(\zeta _{\text {SN}}\), \(\tilde{\zeta }_{\text {SN}}\), \(\mu _{\text {SN}}\):
-
Damping parameter at saddle-node bifurcation
- \(A(\epsilon t)\) :
-
Amplitude of \(\mathcal {O}(1)\) solution in multiple-scale analysis
- \(A_k(t)\) :
-
Amplitude of node k in the network
- C(u):
-
Damping matrix of the network model
- f :
-
Amplitude vector for exogenous excitation
- F(t):
-
Time-dependent exogenous excitation vector
- \(I_N\) :
-
\(N\times N\) identity matrix
- K :
-
Stiffness matrix of the network model, determined by network topology
- L :
-
The Laplacian matrix of network
- P :
-
An orthogonal matrix whose columns span the eigenspace of the network Laplacian
- Q :
-
The index of the nonlinear oscillator node in the network model
- u(t):
-
Nodal displacement vector in the network model
- x :
-
Modal coordinate vector
Appendix B: Multiple-scales perturbation analysis for the 4-node network with small linear damping
In this appendix, we present a detailed derivation of the slow-flow dynamics in Eqs. (7) and (8) in the limit of \(\epsilon \ll 1\) using the method of multiple scales [19]. For the case of constant \(\zeta _2,\zeta _3,\zeta _4\sim \mathcal {O}(1)\), we seek to arrive at a zeroth-order (in \(\epsilon \)) description of the displacement vector u that mirrors the form of the corresponding linear normal-mode, albeit with slowly varying amplitude and phase. To this end, we assume two time scales \(t_0=t\) and \(t_1=\epsilon t\) and, per Eq. (6), a response of the form
and compute derivatives using the relationship
Using this notation, we obtain
and so on. Substitution into the governing equation (1) with \(F=0\) results in the perfect cancellation of all \(\mathcal {O}(1)\) terms. A similar cancellation of all terms of \(\mathcal {O}(\epsilon )\) requires that
where
where we use the shorthand \(c_k=\cos \left( 2kt_0+k\phi (t_1)\right) \) and \(s_k=\sin \left( 2kt_0+k\phi (t_1)\right) \). It follows that
where \(w=(v_1-v_4)/\sqrt{2}\). For a consistent approximation, the secular terms on the right-hand side proportional to \(\cos (2t_0+\phi (t_1))\) and \(\sin (2t_0+\phi (t_1))\), respectively, must cancel, resulting in the conditions
and
Appendix C: 4-node network with small linear damping and \(Q=2\), 3, and 4, respectively
The existence of a subcritical Hopf bifurcation and an associated branch of limit cycles found in the case of \(Q=1\) carries over also to other configurations of the network with the nonlinear oscillator located at node 2, 3, or 4. Indeed, when the nonlinear oscillator is located at node 4 with \(\zeta _1,\zeta _2,\zeta _3\sim \mathcal {O}(1)\), the network symmetry results in an identical set of results with \(\zeta _4\) replaced by \(\zeta _1\). In contrast, if the nonlinear oscillator is located at node 2 with \(\zeta _1,\zeta _3,\zeta _4\sim \mathcal {O}(1)\), the complex exponential rates in Eq. (6) become
In this case, with \(\zeta :=\zeta _1+\zeta _3+\zeta _4\), \(u=0\) is asymptotically stable for \(\zeta >9\) and unstable for \(\zeta <9\). Substitution of the ansatz
into the fully nonlinear governing equations then yields the differential equation
governing the slow dynamics of the amplitude A with nonzero equilibria obtained from
i.e., for \(\zeta \in [0,81/4]\) with two co-existing solutions on the interval \(\zeta \in [9,81/4)\). We obtain the equivalent of Eq. (10), for example, by letting
since these imply that
Finally, with the nonlinear oscillator located at node 3 with \(\zeta _1,\zeta _2,\zeta _4\sim \mathcal {O}(1)\), the complex exponential rates in Eq. (6) become
In this case, with \(\zeta :=\zeta _1+\zeta _4\), \(u=0\) is asymptotically stable for \(\zeta >4\) and unstable for \(\zeta <4\). Substitution of the ansatz
into the fully nonlinear governing equations then yields the differential equation
governing the slow dynamics of the amplitude A with nonzero equilibria obtained from
i.e., for \(\zeta \in [0,9]\) with two co-existing solutions on the interval \(\zeta \in [4,9)\). We obtain the equivalent to Eq. (10), for example, by letting
since these imply that
Appendix D: 4-node network with large linear damping and \(Q=2\), 3, and 4, respectively
When the nonlinear oscillator is located at node 4 with \(\zeta _1,\zeta _2,\zeta _3\sim \mathcal {O}(1/\epsilon ^2)\), the analysis from Sect. 2.2 still applies, albeit with \(\zeta _4\) and \(\tilde{\zeta }_4\) replaced by \(\zeta _1\) and \(\tilde{\zeta }_1\), respectively. If, instead, the nonlinear oscillator is located at node 2 with \(\zeta _1,\zeta _3,\zeta _4\sim \mathcal {O}(1/\epsilon ^2)\), we obtain the complex exponential rates
where \(\tilde{\zeta }_1:=1/\epsilon ^2\zeta _1\), \(\tilde{\zeta }_3:=1/\epsilon ^2\zeta _3\), and \(\tilde{\zeta }_4:=1/\epsilon ^2\zeta _4\). With \(\tilde{\zeta }:=\tilde{\zeta }_1+\tilde{\zeta }_3+\tilde{\zeta }_4\), it follows that the trivial equilibrium is asymptotically stable for \(\tilde{\zeta }>4\) and unstable for \(\tilde{\zeta }<4\), with \(\tilde{\zeta }=4\) corresponding to a Hopf bifurcation out of which emanates a branch of periodic orbits approximated in the small-amplitude limit by the normal-mode oscillations \(u_2=A\cos (2t+\phi )\), \(u_1(t)=u_3(t)=u_4(t)=0\) for constant amplitude A and phase \(\phi \). A consistent multiple-scale ansatz now yields the amplitude equation
with nontrivial equilibria at
i.e., for \(\tilde{\zeta }\in (0,9]\) with two co-existing solutions on the interval \(\tilde{\zeta }\in [4,9)\). We obtain the equivalent of Eq. (10), for example, by letting
since these imply that
Finally, when the nonlinear oscillator is located at node 3 with \(\zeta _1,\zeta _2,\zeta _4\sim \mathcal {O}(1/\epsilon ^2)\), we obtain the complex exponential rates
where \(\tilde{\zeta }_2:=1/\epsilon ^2\zeta _2\). It follows that the trivial equilibrium is asymptotically stable for \(\tilde{\zeta }_2>2\) and unstable for \(\tilde{\zeta }_2<2\), with \(\tilde{\zeta }_2=2\) corresponding to a Hopf bifurcation out of which emanates a branch of periodic orbits approximated in the small-amplitude limit by the normal-mode oscillations \(u_3=A\cos (\sqrt{2}t+\phi )\), \(u_1(t)=u_2(t)=u_4(t)=0\) for constant amplitude A and phase \(\phi \). A consistent multiple-scale ansatz now yields the amplitude equation
with nontrivial equilibria at
i.e., for \(\tilde{\zeta }_2\in (0,9/2]\) with two co-existing solutions on the interval \(\tilde{\zeta }_2\in [2,9/2)\). We obtain the equivalent of Eq. (10), for example, by letting
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Mao, Y., Dankowicz, H. Design of active network filters as hysteretic sensors. Nonlinear Dyn 111, 1763–1789 (2023). https://doi.org/10.1007/s11071-022-07889-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11071-022-07889-3