past and on-going research projects,
involved as main investigator (cumulated amount of ca. 3M€)
European projects
- 2010-2013: PHOCUS (towards a PHOtonic liquid state machine based on delay-CoUpled Systems)
FET project FP7-ICT-2009-C 240763, funded by the European Commission.
This currently running project (started in January 2010) is aimed at demonstrating the experimental possibility to perform a photonic computer, according to brain-inspired computational principles recently proposed by W. Maass in 2002, and also proposed independently by the neural network computing community in 2001, by H. Jaeger. This emerging field of research is still rather poorly popular within the broad scientific community (i.e. in Physics, Optics or Nonlinear dynamics). Our original approach is not only to propose technologically one of the first implementation of such LSM, ESN or RC principles. It also consists of a fundamentally original approach intended to replace the traditional spatially extended network of neuron devoted the complex processing and computing, by a nonlinear delay dynamics well known to possess similar properties compared to spatially extended dynamical network (such as a infinite dimensional phase space). Our contribution to that project is to propose a photonic implementation of such LSM via our experience in electro-optic delay dynamical systems.
- 2006-2009: PICASSO (Photonic Integrated Components Applied to Secure ChaoS encoded Optical communication systems)
STREP project FP6-2006-IST-2.5.1, funded by the European Commission.
The aim of the PICASSO project was to investigate and demonstrate efficient and operational devices and systems for practical optical chaos communications.
Our contribution was to propose a novel electro-optic architecture, the EO phase chaos generator. This architecture has introduced a temporally nonlocal non linearity in the dynamics, via a second short delay additionally to the usual long one in delay dynamics) providing a higher stability together with a higher bandwidth (>10GHz) and a better chaos synchronization quality (ca. 20dB over the full bandwidth).
Record bit rate at 10Gb/s has been demonstrated with 10-9 Bit Error Rate in back-to-back transmission. A first field experiment at 10Gb/s was conducted with this phase chaos setup, over the fiber network of our city of Besançon (the Lumière brothers ring network of 22km), and a second demonstration with the same setup also successfully conducted over more than 110km of the Athens metropolitan network, together with our colleagues from the university of Athens. The Athens group also performed the field experiment, but with the more conventional external cavity semiconductor laser setup which was however proposed in a fully integrated and compact packaging fabricated by the Heinrich Herz Institute in Berlin.
Security issues have been investigated as well in terms of chaos generator architecture, device dependent eavesdropping (customized imbalanced passive interferometer have been developped with the componies Phoenix Photonics and Kylia). A very promising information / chaos mixing protocol has been also proposed in order to enhance the security, more precisely we have used a mixing algebra principle resulting in entropy amplification between the chaotic motion and a pseudo-random bit sequence provided by an additional digital secret key.
- 2001-2004: OCCULT (Optical Chaos Communications Using Laser-diode Transmitters)
FP5 project FET IST-2000-29683, funded by the European Commission.
This project was a major European initiative in the exploration of the potential of optical chaos for securing at the physical layer standard optical fiber communications. It started one year after a similar US MURI project, but it succeeded to strongly improve the results obtained by the US consortium. Among numerous results, the EC OCCULT project succeeded in the first field experiment of optical chaos communications. The field experiment was conducted over the Metrolpolitan Area Network of the city of Athens, a fiber network specifically installed for the 2005 Olympic Games. Record bit rate (up to 3G/s) and record signal transmission quality (Bit Error Rate -BER- as low as 10-7 in back-to-back) have been achieved over more than 100km. The electro-optic intensity chaos setup developed by the our group not only showed an innovative photonic architecture for broadband chaos generation of strong fundamental interest, but it also showed extremely high performances in terms of synchronisation quality and bandwidth capability.
National projects
- 2015-2017: BIPHOPROC project
Brain Inspired PHOtonic PROCessing is funded by the ANR (French funding agency for research), in the framework of the OH-RISQUE call. About 200k€ funding shared between 3 partners, FEMTO-ST, FCLAB (research federation around Fuel Cell systems), and LMB. The objective is to demonstrate the effective capability of Reservoir Computing to solve complex task of the real world in the framework of advanced technological applications, such as diagnostic and prognostic of Fuel Cell systems, and such as real time control of the focusing conditions in fs laser machining systems. See the project webpage for more information.
- 2012-2020: Labex ACTION
Funded by ANR (French funding agency for research). So-called Labex (Laboratory of Excellence) project coordinated by FEMTO-ST and associating two other French labs, the ICB in Dijon, and the LNIO in Troyes. This is a broad area and long term research initiative with an ANR funding of 1M€/year during 8 years, centered around the "Smart Systems" topic. Within this ACTION project constructed along 5 disciplinary WorkPackages and 5 demonstrators, I am responsible for the DEMO 3, "Photonic Neuromorphic Computer". It can be viewed as a straightforward continuation of the PHOCUS FP7 project, with a long term target towards a photonic integrated chip providing Neuromorphic computing capabilities.
Main achievements of this on-going project include fundamental results about the actual analogy between spatio-temporal dynamics and delay dynamics supporting the relevance to use delay dynamics as a way to emulate virtual neural networks. The discovery of Chimera states in delay dynamics was an important result obtained within the ACTION project in 2013 (Phys.Rev.Lett 2013, funding for Bogdan Penkovsky's thesis).
A collaboration with the Mathematics Laboratory of Besançon was also triggered through the Labex ACTION, with researchers working on Stochastic predictions for Econometry. Very interesting results on such predictions done through Delay dynamics-based Nonlinear Transient Computing (or Reservoir Computing) principles, have led to a very first paper on the topic in the journal Neural Networks (accepted March 2014, to be published soon).
Beyond the scientific achievements, the Labex ACTION is aimed in a broader spirit to animate the community on the topics to be explored. This has already concerned an international Workshop in Besançon, on 14th to 15th of October 2013, and the contribution to the organization of the 2014 edition of the Nonlinear Dynamics summer school in Peyresq.
- 2006-2009: O2E (Oscillateurs Opto-Electroniques)
Funded by ANR (French funding agency for research). We have coordinated this project dealing with a particular class of microwave oscillator involving both photonic and electronic devices, the OptoElectronic Oscillator (OEO). The project is aimed at acquiring the missing experience, and also at bringing original contributions, into a field of research which is essentially trusted by one group worldwide, the Jet Propulsion Laboratory at the NASA in the USA, together with a start-up company explicitely created on the concept in 2000, OEwaves.
Our contribution at FEMTO-ST was concerned by the nonlinear dynamics approach of the OEO architecture, leading to the discovery of unexpected instabilty mechanisms potentially existing in OEOs. We indeed observed experimentally, and demonstrated both analytically and numerically, an unusual envelope destabilization mechanism leading to a slow amplitude modulation of the microwave oscillation (Neimarck-Sacker bifurcation), with an envelop period related to twice the time delay. From the more applied perspective, we achieved a nearly state-of-the-art for the phase noise spectrum of a standard OEO, developing also the adequate means for the measurement of such short-term ultra-high spectral stability (cross-correlation phase noise measurement bench with fiber delay lines). We also investigated, in collaboration with the CNES, the potential of millimeter diameter (hence with microwave Free Spectral Range) disk resonators for OEO, and we developed fabrication techniques for such crystalline disks in MgF2. Last but not least, we also proposed in the frame of this very fruitful project, an original optical pulse generator based on both OEO architecture, and pulse compression techniques.
The other partners of the project were the ENSSAT in Lannion (exploring fundamental aspects of nonlinear optical cavities for OEOs, and developping accurate measurement and characterization techniques for high-Q optical resonators), and the LAAS in Toulouse, investigating fiber-based ring resonators. - 2005-2009: ACSCOM (Apport du Chaos dans la Sécurité des systèmes Communicants Optiques et Mobiles)
Funded by ANR (French funding agency for research). This project was coordinated by the IREENA lab. in Nantes, and involed XLIM brive, the LESIA in Toulouse, and 2 industrial partners, Orange Labs and Arbos ingénierie. Our main contribution concerned the demonstration of a pulsed laser based chaos communication system, allowing for a switched packet data transmission protocole, and for an "on-demand" discrete time optical chaos encryption system. Together with our partners, we also developed a hybrid optical and electronic FM chaos generator transposition our original concept of delay dynamics to the RF microwave communication channel. - 2002-2005: Transchaos (Le Chaos pour la Sécurité des Transmissions)
Funded by ACI, the former ANR (French funding agency for research). - 1999-2000: Electronic chaos generation for secure radio-transmission
funded by the ANVAR (French funding agency for intellectual property and innovation in research). - 1994-1997: Wavelength chaos for optical encryption
Funded by CNET / France Telecom, program "Fonctions Optiques Nouvelles pour les Télécommunications". This project brought the financial support for the experiments developed during my PhD Thesis.
Regional projects
- 2009-2011: Light manipulation for advanced optical signal processing
Funded by the Region Franche-Comté.
The project was jointly conducted with the nanophotonics group (coordinator Maria Pilar Bernal, photonics structures in integrated optics devices), the nonlinear optics group (Gil Fanjoux, slow light and nonlinear optics effects), and the OPTO group (nonlinear dynamics and chaos communications). Our main task was targeted at the demonstration of the possibility of a customized hardware key (in collaboration with the Kylia) company), consisting in a 3-wave imbalanced passive interferometer. This device allowed the demonstration of chaos synchronization and chaos-based secure optical transmission, with a customized nonlocal nonlinear transformation, which element is at the deterministic origine of the chaotic motions masking the information to be securely transmitted.