Abstract
Starting from the piecewise-linear framework applied to networked FitzHugh-Nagumo oscillators, this paper aims at presenting a control strategy relying on phase portrait reshaping through the manipulation of their nullclines in order to fulfil both phase and time requirements. This has been achieved by relating the slopes of piecewise linearly approximated model’s nullclines in its limit cycle to its oscillation period. Additionally, the targeted issue addressed by this paper has been assessed by combining the former framework with an event-driven control strategy aimed at reducing its effects to specific time instants instead of continuous, instantaneous phases, which is much more computationally expensive. The strategy is therefore motivated by its simplicity and is supported by key applications to bioinspired locomotion control in legged robots, which suggest how a dynamics-preserving approximation of the phase portrait combined with a sampled control action can produce pre-defined phase topologies in directed, non-diffusive tree graphs.
Similar content being viewed by others
References
Ambe, Y., Nachstedt, T., Manoonpong, P., Wörgötter, F., Aoi, S., Matsuno, F.: Stability analysis of a hexapod robot driven by distributed nonlinear oscillators with a phase modulation mechanism. In: 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 5087–5092 (2013)
Aoi, S., Tsuchiya, K.: Locomotion control of a biped robot using nonlinear oscillators. Auton. Robots 19(3), 219–232 (2005)
Appleton, E., van der Pol, B.: XXV. on the form of free triode vibration. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science 42(248), 201–220 (1921)
Arena, P., Bonanzinga, A., Patané, L.: Emergence of locomotion gaits through sensory feedback in a quadruped robot, chap. 25 in From Parallel to Emergent Computing Taylor and Francis Group (2018)
Arena, P., Fortuna, L., Frasca, M., Lombardo, D., Patané, L., Crucitti, P.: Turing patterns in RD-CNNs for the emergence of perceptual states in roving robots. Int. J. Bifurcat. Chaos 17(01), 107–127 (2007)
Arena, P., Furia, P., Patané, L., Pollino, M.: Fly-inspired sensory feedback in a reaction-diffusion neural system for locomotion control in a hexapod robot. In: 2015 International Joint Conference on Neural Networks (IJCNN), pp. 1–8 (2015)
Arima, N.: A generation mechanism of canards in a piecewise linear system, pp. 447–453. IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences pp (1997)
Avishai, Y., Luck, J.M.: Tight-binding electronic spectra on graphs with spherical topology: II. The effect of spin–orbit interaction. J. Stat. Mech. 2008(06), P06008 (2008)
Balti, K., Elloumi, S.: Optimizing swarm robot’s exploration time using graph matrices. In: 2018 International Conference on Advanced Systems and Electric Technologies, pp. 484–489 (2018)
Barbosa, P.T., Saa, A.: Chaotic oscillations in singularly perturbed Fitzhugh–Nagumo systems. Chaos Solitons Fractals 59, 28–34 (2014)
Barca, J.C., Sekercioglu, A., Ford, A.: Controlling formations of robots with graph theory. In: Lee, S., Cho, H., Yoon, K.J., Lee, J. (eds.) Intelligent Autonomous Systems, vol. 12, pp. 563–574. Springer, Berlin (2013)
Biancardi, C.M., Minetti, A.E.: Biomechanical determinants of transverse and rotary gallop in cursorial mammals. J. Exp. Biol. 215(23), 4144–4156 (2012)
Blanchar, P., Krüger, T.: The “Cameo Principle” and the origin of scale-free graphs in social networks. J. Stat. Phys. 114(5), 1399–1416 (2004)
Boccaletti, S., Pecora, L.M., Pelaez, A.: Unifying framework for synchronization of coupled dynamical systems. Phys. Rev. E 63(6), 066219 (2001)
Bolzano, B.: Rein analytischer Beweis des Lehrsatzes, daß zwischen je zwei Werthen, die ein entgegengesetztes Resultat gewähren, wenigstens eine reelle Wurzel der Gleichung liege. Haase (1817)
Brown, R., Kocarev, L.: A unifying definition of synchronization for dynamical systems. Chaos 2(2) (2000)
Brunel, N., van Rossum, M.: Lapicque’s 1907 paper: from frogs to integrate-and-fire. Biol. Cybern. 97, 337–9 (2008). https://doi.org/10.1007/s00422-007-0190-0
Bucher, D., Haspel, G., Golowasch, J., Nadim, F.: Central Pattern Generators, pp. 1–12. American Cancer Society (2015)
Chen, C.-N., Kung, S.-Y.: Turing patterns and standing waves in Fitzhugh-Nagumo type systems. In: Progress in Variational Methods, pp. 30–39. World Scientific, Singapore (2011). https://doi.org/10.1142/9789814327848_0003
Chowdhury, A.R., Soh, G.S., Foong, S., Wood, K.L.: Implementation of caterpillar inspired rolling gait and nonlinear control strategy in a spherical robot. J. Bionic Eng. 15(2), 313–328 (2018)
Chung, F.: Laplacians and the Cheeger inequality for directed graphs. Ann. Comb. 9(1), 1–19 (2005)
Ciesielski, K.: The Poincaré-Bendixson theorem: from poincaré to the XXIst century. Cent. Eur. J. Math. 10(6), 2110–2128 (2012)
Coombes, S.: Neuronal networks with gap junctions: a study of piecewise linear planar neuron models. SIAM J. Appl. Dyn. Syst. 7(3), 1101–1129 (2008)
van der Schaft, A., Schumacher, H.: Introduction to Hybrid Dynamical Systems. Springer, Berlin (1999)
Dietz, V., Zijlstra, W., Duysens, J.: Human neuronal interlimb coordination during split-belt locomotion. Exp. Brain Res. 101(3), 513–520 (1994)
Dong, H., Duan, S., Zhao, Y.: Delaunay graph based path planning method for mobile robot. In: 2010 International Conference on Communications and Mobile Computing, vol. 3, pp. 528–531 (2010)
Dutra, M.S., de Pina Filho, A.C., Romano, V.F.: Modeling of a bipedal locomotor using coupled nonlinear oscillators of van der pol. Biol. Cybern. 88(4), 286–292 (2003)
Ermentrout, B.: An introduction to neural oscillators. In: Ventriglia, F. (ed.) Neural Modeling and Neural Networks, Pergamon Studies in Neuroscience, pp. 79–110. Pergamon, Amsterdam (1994)
Fan, D., Song, X., Liao, F.: Synchronization of coupled Fitzhugh–Nagumo neurons using self-feedback time delay. Int. J. Bifurcat. Chaos 28(02), 1850031 (2018)
Fernández-García, S., Desroches, M., Krupa, M., Clément, F.: A multiple time scale coupling of piecewise linear oscillators. Application to a neuroendocrine system. SIAM J. Appl. Dyn. Syst. 14(2), 643–673 (2015)
Fukuoka, Y., Habu, Y., Fukui, T.A.: A simple rule for quadrupedal gait generation determined by leg loading feedback: a modeling study. Sci. Rep. 5, 8169 (2015)
Turin, G.: An introduction to matched filters. IRE Trans. Inf. Theory 6(3), 311–329 (1960)
Gerstner, W., Kistler, W.M., Naud, R., Paninski, L.: Neuronal Dynamics: From Single Neurons to Networks and Models of Cognition. Cambridge University Press, Cambridge (2014)
Ginoux, J.M., Llibre, J.: Canards existence in Fitzhugh–Nagumo and Hodgkin–Huxley neuronal models. Mathematical Problems in Engineering 2015 (2015)
Golubitsky, M., Stewart, I., Buono, P.L., Collins, J.: A modular network for legged locomotion. Physica D 115(1), 56–72 (1998)
Gottschall, J.S., Nichols, T.R.: Head pitch affects muscle activity in the decerebrate cat hindlimb during walking. Exp. Brain Res. 182(1), 131–135 (2007)
Guertin, P.: Central pattern generator for locomotion: anatomical, physiological, and pathophysiological considerations. Front. Neurol. 3, 183 (2013)
Harris-Warrick, R.M.: Locomotor pattern generation in the rodent spinal cord. In: Encyclopedia of Computational Neuroscience (2014)
Hodgkin, A., Huxley, A.: A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. 117(4), 500–544 (1952)
Iqbal, M., Rehan, M., Hong, K.S.: Robust adaptive synchronization of ring configured uncertain chaotic Fitzhugh–Nagumo neurons under direction-dependent coupling. Front. Neurorobot. 12, 6 (2018)
Iqbal, N., Wu, R., Liu, B.: Pattern formation by super-diffusion in Fitzhugh–Nagumo model. Appl. Math. Comput. 313, 245–258 (2017)
Izhikevich, E.M.: Dynamical Systems in Neuroscience: The Geometry of Excitability and Bursting, vol. 25 (2007)
Kaveh, A.: Introduction to Graph Theory and Algebraic Graph Theory, pp. 15–35. Springer, Vienna (2013)
Keener, J., Sneyd, J.: Mathematical Physiology. Springer, New York (2009)
Kuznetsov, M., Kolobov, A., Polezhaev, A.: Pattern formation in a reaction-diffusion system of Fitzhugh–Nagumo type before the onset of subcritical Turing bifurcation. Phys. Rev. E 95, 052208 (2017)
Li, G., Zhang, H., Zhang, J., Bye, R.T.: Development of adaptive locomotion of a caterpillar-like robot based on a sensory feedback CPG model. Adv. Robot. 28(6), 389–401 (2014)
Liu, G.L., Habib, M.K., Watanabe, K., Izumi, K.: Central pattern generators based on Matsuoka oscillators for the locomotion of biped robots. Artif. Life Robot. 12(1), 264–269 (2008)
Liu, H., Jia, W., Bi, L.: Hopf oscillator based adaptive locomotion control for a bionic quadruped robot. In: 2017 IEEE International Conference on Mechatronics and Automation (ICMA), pp. 949–954 (2017)
Lunze, J., Lamnabhi-Lagarrigue, F.: Handbook of Hybrid Systems Control: Theory, Tools, Applications. Cambridge University Press, Cambridge (2009)
Manganaro, G., Fortuna, L., Arena, P.: Cellular Neural Networks, 1st edn. Springer, Berlin (1999)
Marder, E., Bucher, D.: Central pattern generators and the control of rhythmic movements. Curr. Biol. 11(23), R986–R996 (2001)
McCulloch, W.S., Pitts, W.: A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys. 5(4), 115–133 (1943)
McKean, H.: Nagumo’s equation. Adv. Math. 4(3), 209–223 (1970)
Minassian, K., Hofstoetter, U.S., Dzeladini, F., Guertin, P.A., Ijspeert, A.: The human central pattern generator for locomotion: Does it exist and contribute to walking? Neuroscientist 23(6), 649–663 (2017). PMID: 28351197
Minati, L., Frasca, M., Yoshimura, N., Koike, Y.: Versatile locomotion control of a hexapod robot using a hierarchical network of nonlinear oscillator circuits. IEEE Access 6, 8042–8065 (2018)
Mishra, D., Yadav, A., Kalra, P.K.: Chaotic behavior in neural networks and Fitzhugh–Nagumo neuronal model. In: Pal, N.R., Kasabov, N., Mudi, R.K., Pal, S., Parui, S.K. (eds.) Neural Information Processing, pp. 868–873. Springer, Berlin (2004)
Muñoz-Pacheco, J.M., Zambrano-Serrano, E., Félix-Beltrán, O., Gómez-Pavón, L.C., Luis-Ramos, A.: Synchronization of pwl function-based 2d and 3d multi-scroll chaotic systems. Nonlinear Dyn. 70(2), 1633–1643 (2012)
Nguyen, L.H., Hong, K.S.: Synchronization of coupled chaotic Fitzhugh–Nagumo neurons via Lyapunov functions. Math. Comput. Simul. 82(4), 590–603 (2011)
Peng, B., Zhang, L., Zhang, D.: A survey of graph theoretical approaches to image segmentation. Pattern Recognit. 46(3), 1020–1038 (2013)
Plotnikov, S.A., Lehnert, J., Fradkov, A.L., Schöll, E.: Synchronization in heterogeneous Fitzhugh–Nagumo networks with hierarchical architecture. Phys. Rev. E 94, 012203 (2016)
Rajasekar, S., Lakshmanan, M.: Bifurcation, chaos and suppression of chaos in Fitzhugh–Nagumo nerve conduction model equation. J. Theor. Biol. 166(3), 275–288 (1994)
Rigatos, G.: Robust synchronization of coupled neural oscillators using the derivative-free nonlinear Kalman filter. Cogn. Neurodyn. 8(6), 465–478 (2014)
Robilliard, J.J., Pfau, T., Wilson, A.M.: Gait characterisation and classification in horses. J. Exp. Biol. 210(2), 187–197 (2007)
Rosenblum, M.G., Pikovsky, A.S., Kurths, J.: Phase synchronization of chaotic oscillators. Phys. Rev. Lett. 76, 1804–1807 (1996)
Rotstein, H., Coombes, S., Gheorghe, A.: Canard-like explosion of limit cycles in two-dimensional piecewise-linear models of Fitzhugh–Nagumo type. SIAM J. Appl. Dyn. Syst. 11(1), 135–180 (2012)
Rubensson, M., Lennartson, B., Pettersson, S.: Convergence to limit cycles in hybrid systems—an example. IFAC Proc. Vol. 31(20), 683–688 (1998)
Selverston, A.I.: Invertebrate central pattern generator circuits. Philos. Trans. R. Soc. Lond. B Biol. Sci. 365(1551), 2329–2345 (2010)
Shishkin, A., Postnov, D.: Stochastic dynamics of Fitzhugh–Nagumo model near the canard explosion. Phys. Control 2, 649–653 (2003)
Sigvardt, K.A., Williams, T.L.: Models of central pattern generators as oscillators: the lamprey locomotor CPG. Semin. Neurosci. 4(1), 37–46 (1992). The Use of Models in the Neurosciences
Solé, R.V., Munteanu, A.: The large-scale organization of chemical reaction networks in astrophysics. Europhys. Lett. (EPL) 68(2), 170–176 (2004)
Soleimani, H., Ahmadi, A., Bavandpour, M.: Biologically inspired spiking neurons: piecewise linear models and digital implementation. IEEE Trans. Circuits Syst. I Regul. Pap. 59(12), 2991–3004 (2012)
Song, Y.: Networks: an introduction by M. E. J. Newman. J. Math. Sociol. 37(4), 250–251 (2013)
Storace, M., Feo, O.D.: Pwl approximation of dynamical systems: an example. In: Proceedings of the 2003 International Symposium on Circuits and Systems, 2003. ISCAS ’03 (2003)
Tonnelier, A.: The McKean’s caricature of the Fitzhugh–Nagumo model I. The space-clamped system. SIAM J. Appl. Math. 63(2), 459–484 (2003)
Tonnelier, A., Gerstner, W.: Piecewise linear differential equations and integrate-and-fire neurons: insights from two-dimensional membrane models. Phys. Rev. E 67, 021908 (2003)
Tsujita, K., Toui, H., Tsuchiya, K.: Dynamic turning control of a quadruped robot using oscillator network. In: Robotics and Automation, 2004. Proceedings. ICRA ’04. 2004 IEEE International Conference on, vol. 3, pp. 2613–2618 Vol. 3 (2004)
Verhulst, F.: Singular perturbation methods for slow-fast dynamics. Nonlinear Dyn. 50(4), 747–753 (2007)
Vilela, M., Vinga, S., Maia, M.A.G.M., Voit, E.O., Almeida, J.S.: Identification of neutral biochemical network models from time series data. BMC Syst. Biol. 3(1), 47 (2009)
Wang, M., Yu, J., Tan, M.: Cpg-based sensory feedback control for bio-inspired multimodal swimming. Int. J. Adv. Robot. Syst. 11(10), 170 (2014)
Wang, W., Wang, K., Zhang, H.: Crawling gait realization of the mini-modular climbing caterpillar robot. Prog. Nat. Sci. 19(12), 1821–1829 (2009)
Yang, C.C., Lin, C.L.: Robust adaptive sliding mode control for synchronization of space-clamped Fitzhugh–Nagumo neurons. Nonlinear Dyn. 69(4), 2089–2096 (2012)
Yoder, J.G.: Chapter 3—Christiaan huygens, book on the pendulum clock (1673). In: Grattan-Guinness, I., Cooke, R., Corry, L., Crépel, P., Guicciardini, N. (eds.) Landmark Writings in Western Mathematics 1640–1940, pp. 33–45. Elsevier Science, Amsterdam (2005)
Zakharov, A., Barinov, A., Zhiznyakov, A.: Faces selection in images using the spectral graph theory and constraints. In: 2017 International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM), pp. 1–5 (2017)
Zavlanos, M.M., Egerstedt, M.B., Pappas, G.J.: Graph-theoretic connectivity control of mobile robot networks. Proc. IEEE 99(9), 1525–1540 (2011)
Zeng, Y., Li, J., Yang, S.X., Ren, E.: A bio-inspired control strategy for locomotion of a quadruped robot. Appl. Sci. (2018). https://doi.org/10.3390/app8010056
Zheng, Q., Shen, J.: Pattern formation in the Fitzhugh–Nagumo model. Comput. Math. Appl. 70(5), 1082–1097 (2015)
Zhu, S., Zhu, X., Luo, Q.: Graph theory based image segmentation. In: 2013 6th International Congress on Image and Signal Processing (CISP), vol. 2, pp. 593–598 (2013)
Funding
This study was funded by MIUR project CLARA - Cloud platform for LAndslide Risk Assessment (Grant Number SNC\(\_\)00451).
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
P-FHNs and stability issues
Because of the nature of P-FHNs, we want to prove that asymptotic stability of this kind of systems does not depend on the particular fixed point as in the case of linear systems, but before doing this, we have to prove the following statement.
Lemma 1
Given a PWL function as described by Eq. 7, let \(s_j\doteq \text {sgn}\left( x-E_j\right) \), \(\bar{{\mathbf {m}}}\doteq \left[ m_1,\ldots ,m_{n-1}\right] \) and \(\bar{{\mathbf {s}}}\doteq \begin{bmatrix} s_1-s_2\\ \vdots \\ s_{n-1}-s_{n} \end{bmatrix}\), then its derivative with respect to x is:
Proof
Every PWL function in the form of Eq. 7 admits the following derivative:
If \({\mathbf {s}}\doteq \begin{bmatrix} s_1\\ \vdots \\ s_{n} \end{bmatrix}\) and \({\mathbf {m}}\doteq \left[ m_0,\ldots ,m_n\right] \), then the summation can be rewritten as follows:
because every \(b_j\) depends on both \(m_j\) and \(m_{j-1}\). Thus:
By grouping together all the common terms:
\(\square \)
This lemma is valid \(\forall E_j\), therefore it still holds if \(E_{j-1}<E_j<E_{j+1}\). However, if the former ordering criterion holds too, the aforementioned expression given by the previous lemma simply states that the derivative of the PWL function is given by the slope of the branch where \(x_1\) is located in. For example, if \(s_1=-1\) (therefore, \(s_j=-1\), with \(j\ge 2\)) then the selected branch is the leftmost one, thus the derivative is equal to \(m_0\). Similarly, if \(s_n=1\) (therefore, \(s_j=1\), with \(j<n\)) then the selected branch is the rightmost one, thus the derivative is equal to \(m_n\). In all the other cases, \(\exists !j:s_j=-s_{j+1}\): in this region, the derivative is equal to \(m_j\). We can summarise these statements as follows:
Corollary 1
If the hypotheses of Lemma 1 hold, then:
Now, we are ready to enunciate a simple stability condition which employs the previous results:
Lemma 2
(Stability of fixed points) Given the system 8 with n break-points, asymptotic stability within the \(j+1\)th sub-region is achieved iff \(b>\frac{m_j}{\varepsilon }\) and \(bm_j<1, j\in \left\{ 0,1,\ldots ,n\right\} \).
Proof
The Jacobian matrix of such systems is given by:
where \(\frac{\text {d}\varPi (x_1)}{\text {d}x_1}\) is given by Eq. 27. Due to the nature of the system, the characteristic polynomial is simply given by:
By applying the Descartes’ rule of signs, we can infer that:
where \(\text {LHS}\doteq \left\{ \mu \in {\mathbb {C}}|\text {Re}\left\{ \mu \right\} <0\right\} \) stands for Left-Half Semiplane of the Argand–Gauss plane and \(0\le j\le n\). \(\square \)
To apply the previous lemma, it is sufficient to check if Condition 30 is verified within the region of interest, according to the desired branch of the phase portrait. It derives that P-FHNs constitute a simpler case study with \(n=2\) and therefore we have only 3 slopes. Thus, if these slopes were chosen according to Table 2, then the central region of the phase portrait would be always unstable.
Real and virtual equilibrium points in P-FHNs
The calculation of fixed points of a P-FHN consists on finding those points which nullify both the nullclines. In order to calculate P-FHN’s fixed points, we can compute the intersection points between the \(x_2\)-nullcline and each straight line the phase portrait is made of. As stated by Eq. 18, we can write the generic line with pre-defined slope \(m\in \left\{ m_0,m_1,m_2\right\} \) and selected break-point \(E\in \left\{ E_1,E_2\right\} \) alternatively as \(x_2=mx_1+(m_1-m)E+a_0\). Thus, the intersection between the former line and the \(x_2\)-nullcline gives equilibrium points in the form \(\left[ x_1^{\mathrm{(eq)}},\frac{1}{b}x_1^{\mathrm{(eq)}}+\frac{a}{b}\right] \), where:
Equation 31 gives the abscissa of several fixed points, depending on the number of slopes. However, some of them may not belong to the PWL function. To make it clearer, suppose a P-FHN system with \(m_0=-m_1=m_2=-1\), \(E_1=-E_2=-1\), \(a=a_0=0\) and \(b=0.5\). Thus, fixed points are \(\left[ 0,0\right] \), \(\left[ -2/3,-4/3\right] \) and \(\left[ 2/3,4/3\right] \). However, the last two points do not belong to the N-shaped nullcline, even though they are drawn from the computation we have introduced before. This can be easily proven by calculating \(\varPi (\pm 2/3)=\pm 2/3\ne \pm 4/3\), according to Eq. 5. Despite they can be considered fixed points, they are virtual fixed points, whilst the first one is effectively a real fixed point.
The presence of virtual fixed points can be expressed differently as follows. The PWL framework implies a state space fragmentation where each single “slice” is adequately characterised, as stated by Lemma 2. However, the former lemma does not consider the existence of fixed points belonging to regions they are not assigned to. This fact has a great relevance when demonstrating the emergence of limit cycles. Thereby, what follows is an immediate consequence of the previous results which guarantee the existence of limit cycles in P-FHNs:
Lemma 3
(Stable limit cycles) Given a P-FHN system as stated by Eq. 8, if Eq. 7 are verified and the fixed points associated to the most external slopes of its phase portrait are virtual and asymptotically stable, then the system admits a limit cycle.
Proof
If Eq. 7 are satisfied, then Lemma 2 assures the existence of two asymptotically stable sub-regions (i.e., the most external ones) and one unstable sub-region between them. For the sake of simplicity, we refer to them as \({\mathcal {R}}^{\mathrm{Left}}\), \({\mathcal {R}}^{\mathrm{Centre}}\) and \({\mathcal {R}}^{\mathrm{Right}}\). On account of the aforementioned results, each region comprises a fixed point, which could be either real or virtual. We refer to these solutions as \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Left}}\), \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Centre}}\) and \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Right}}\). Let \(\varsigma ({\mathbf {x}})\) be the trajectory of a P-FHN starting from \({\mathbf {x}}\). Because of the stable and virtual nature of both \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Left}}\) and \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Right}}\), \(\forall \bar{{\mathbf {x}}}\in \left\{ {\mathcal {R}}^{\mathrm{Left}},{\mathcal {R}}^{\mathrm{Right}}\right\} \)\(\varsigma (\bar{{\mathbf {x}}})\) will start to converge to virtual points, which are fixed solutions placed in a region they do not actually belong to. In fact, \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Left}}\notin {\mathcal {R}}^{\mathrm{Left}}\) and \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Right}}\notin {\mathcal {R}}^{\mathrm{Right}}\) by definition. Conversely, \(\forall \bar{{\mathbf {x}}}\in {\mathcal {R}}^{\mathrm{Centre}}\) the trajectory will escape from \({\mathbf {x}}^{\mathrm{(eq)}}_{\mathrm{Centre}}\) naturally, reaching either \({\mathcal {R}}^{\mathrm{Left}}\) or \({\mathcal {R}}^{\mathrm{Right}}\). Once the trajectory has reached a stable region having a virtual fixed point, it will be directed towards it, but having placed it within a different region the trajectory will not be able to reach it. Thus, this moving trajectory will create a limit cycle due to the unreachability of its stable fixed points. That having been said, we have to prove that if \(x_1\ll E_1\) (\(x_1\gg E_2\)) the system reduces to a stable one. This is true because in these cases it results taht \(\varPi (x_1)\approx a_1x_1\), where \(a_1<0\) by construction. Therefore, the asymptotic state matrix is given by:
which satisfies Lemma 2, thus it is a Hurwitz matrix and therefore the Poincaré–Bendixson theorem [22] about stability of limit cycles in continuous dynamical systems is verified. \(\square \)
Lemma 3 has an important consequence for further applications and considerations. In fact, it is possible to switch from an oscillatory regime to a steady stable by turning a virtual but stable fixed point into a real one.
P-FHNs as switching systems
When \(\varepsilon \ll 1\), systems in a PWL framework can be considered examples of switching systems, whose dynamics can be linearised through the employment of switching functions which allow to change among diverse working conditions, or modalities.
Let \({\mathbb {M}}\doteq \left\{ M_1,M_2,M_3\right\} \) be the set of working modalities; each modality determines a proper linearised version of the original system through some matrices \({\mathbf {A}}(\text {M})\) and \({\mathbf {B}}(\text {M})\), with \(M\in {\mathbb {M}}\), such that the overall dynamics can be expressed as
In particular, \(\exists {\mathcal {S}}:\varOmega \rightarrow {\mathbb {M}}\) which gives the next operating modality according to the current values of the state variables. For P-FHNs, this switching function is given by:
which in turns give:
where \(j\in \left\{ 1,2,3\right\} \), \({\mathbf {J}}\) is the Jacobian matrix. This notation is perfectly coherent to the works in [66] about the convergence of hybrid systems to limit cycles. We have reported Fig. 12 to show how modalities are linked together and how a P-FHN switches among them, according to Table 9.
Rights and permissions
About this article
Cite this article
Arena, P., Patané, L. & Spinosa, A.G. A nullcline-based control strategy for PWL-shaped oscillators. Nonlinear Dyn 97, 1011–1033 (2019). https://doi.org/10.1007/s11071-019-05028-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11071-019-05028-z