Buenos días,

Esperando que sea de vuestro interés, os remito información sobre una propuesta de PhD por parte de las universidades ENSTA-Paristech y Telecom-Paristech.

Si estáis interesados en hacer un Doctorado en París, podéis poneros en contacto directamente con el profesor a través del siguiente email:

Benoît Geller: benoit.geller@ensta-paristech.fr

Podéis encontrar la propuesta de PhD a continuación:

----------------------------------------------------------------------------------------------------------------

BAYESIAN METHODS FOR SUBMARINE SENSOR NETWORK COMMUNICATIONS

Sensor networks are one of the major challenges for the next decade. Realizing the interface between man and all the parameters of the physical world, those sensor networks must allow the creation of a so-called ambient intelligence, where decisions can be taken on masses of information mapped in space and time. This strategic objective applies in particular to the marine space, which covers the majority of our planet, because civilian and military applications abound : sensitive areas monitoring, tsunami detection, looking for damaged objects lost into the sea, but also monitoring of the ecosystem and help to drilling deeper and deeper and finding precious resources. However, the realization of sensor networks is particularly challenging in the underwater environment. Indeed, the systems are energy constrained, limiting the complexity of the algorithms used, however, this medium is recognized as the most difficult for wave propagation:

- Acoustic bandwidths are narrow and the center frequency changes according to the distance covered. This severely limits the flow.

- The inter-symbol effect can be particularly long due to the relatively lows peed of the acoustic waves.

- Synchronization is very difficult due to both Doppler effect and significantly time-varying propagation delay.

The first objective of this PhD is to propose some innovative solutions to improve the synchronization of sensors for underwater acoustics; this performance improvement is permitted when taking into account the dynamic of the parameters to be estimated, which is allowed by the Bayesian framework. Unfortunately, this increase in performance is most often achieved at the cost of increased complexity of the algorithms developed. This is unacceptable in the context of undersea sensors networks as energy resources are constrained. Recent works at ParisTech focused on the theme of estimating dynamic phase [1]-[5]. The model adopted for this phase is that adopted in many standards namely Brownian walk with Doppler shift. In [1]-[3], we have computed Cramer-Rao bounds (BCRB, HCRB) adapted to this problem (see curves HCRB on the figure displayed below). Then in[4]-[5], we have derived systematically a digital S-PLL phase loop S-phase which is a Bayesian estimator of phase capable of asymptotically reach the previous bounds (see curves "Sim" with high SNR of the figure taken below from [5]). This figure high lights the superiority of the loop phase called S-PLL (denoted "off-line" as opposed to conventional PLL called "on-line") ; the superiority of the S-PLL comes from the fact that it is able to take into account the dynamics over the whole block phase trajectory, unlike the conventional PLL on-line PLL.

We now want to test this kind of algorithm in a realistic case submarine channel and extend previous results to the estimation of other parameters of telecom systems, namely that of the full synchronization in the time-frequency plane. The base for this is to develop algorithms similar to the algorithm described in the international patent [6]. This algorithm estimating phase amplitude channel was tested in a simple (single-carrier without inter-symbol interference) cases with encouraging results. Models found in the literature (eg[7]- [9]) are too simplistic: usually the parameters to be estimated are assumed to be constant or totally independent from one moment to the next. The Bayesian approach in this case is artificial and provides norealgain because it ignores the dynamics of parameters to be estimated. Instead, the proposed approach will be similar to the phase estimate by a single S-PLL phase loop and exemplified by the above figure. At first, the PhD student will derive different Cramér-Rao bounds related to the study. Then he will get systematically synchronization estimators in the time-frequency plane. Very importantly, some specially designed modulation will be analyzed and chosen according to some low-complexity demodulation arguments. Correspondingly a Bayesian equalizer and interference canceler will also have to be developed so as to work at lower signal to noise ratios and larger distances. The algorithms will be developed for different scenarios (Non Data Aided, Code Aided and Data Aided) and the study will systematically be extended to the MIMO configuration.

Since Sea web [10]-[11] in the early 2000s which involved 17 nodes spread over 16km2 to detect submarines with a precision of less than ten meters, submarine sensor networks (see.g.[12]-[14]) are growing fast [15]-[16]. Previous Bayesian temporal synchronization algorithms should finally be applied to a TDOA localization in a multi-frame distributed sensors. A better location is expected due to the higher precision expected from the previous time synchronization. All signal processing algorithms will be tested on real data recorded in a test tank.

REFERENCES

[1]"Analytic and Asymptotic Analysis of Bayesian Cramér-Rao Bound for Dynamical Phase Offset Estimation," S. Bay, C. Herzet, J.P. Barbot, J. M. Brossier, B. Geller, IEEE Transactions on Signal Processing, vol. 56, pp. 61-70, Jan. 2008.

[2] “On the Hybrid Cramer-Rao bound and its application to dynamical phase estimation,” S. Bay, B. Geller, A. Renaux, J.P. Barbot, J.M. Brossier, IEEE Signal Processing Letters, vol. 15, pp. 453-456, 2008.

[3] “Bayesian and Hybrid Cramér-Rao Bounds for the Carrier Recovery under Dynamic Phase Uncertain Channels,” J. Yang, B. Geller, S. Bay, IEEE Trans. on Signal Processing, vol. 59, no 2, Feb. 2011.

[4] “Near Optimum Low Complexity Smoothing Loops for Dynamical Phase Estimation - Application to BPSK Modulated Signals,” J. Yang, B. Geller, IEEE Trans. on Signal Processing, pp. 3704-3711, Sept. 2009.

[5]”Near-MAP Smoothing Loops for Code Aided QAM Dynamical Carrier Phase Estimation,” J. Yang, B. Geller, O. Rioul submitted to IEEE Trans. on Signal Processing, Dec. 2013.

[6] “Method for estimating the phase and the gain of observation data transmitted over a QAM-modulated transmission channel,” B. Geller, C. Vanstraceele, J.M. Brossier, J.P. Barbot, International patent CNRS, WO/2006/032768, PCT Fr2005/002301, 30th March 2006.

[7]"Joint estimation of channel and oscillator phase noise in MIMO systems,"H. Mehrpouyan, A. Nasir, S. Blostein, T. Erikson, G. Karagiannidis, T. Svensson, IEEE Trans. on Signal Processing, Sept. 2012.

[8] “Joint estimation of channel response, frequency offset, and phase noise in multi-carrier systems, ” D. Lin, R. Pacheco, T. Lim and D. Hatzinakos, IEEE trans. on Signal Processing, vol. 54, no 9, pp. 3443-3554, Sept. 2006.

[9]"Phase Noise in MIMO Systems: Bayesian Cramer-Rao Bounds and Soft-Input Estimation," A. Nasir, H. Mehrpouyan, R. Schober, and Y. Hua, IEEE Trans. on Signal Processing, vol. 61, no1, Jan. 2013.

[10] “Shallow water acoustic networks”, J. Proakis, E. Sozer, J. Rice, M. Stojanovic, IEEE communications magazine, vol. 39, Nov. 2001.

[11] “Undersea navigation via a distributed acoustic communication network,” M. Hahn, J. Rice, Proc. of TICA, Istanbul, July 2005.

[12]“A shallow water acoustic network for mine countermeasures operations with autonomous underwater vehicles”, L. Freitag, M. Grund, Ch; von Alt, RlStokey, T. Austin, Proc. of Undersea Defense Technology(UDT), Amsterdam, June 2005.

[13]“Heterogeneous underwater networks for ASW, R. Been, D.T. Hugues, A. Vermeij, Proc. of Undersea Defense Technology (UDT), Glasgow, June 2008.

[14]“Growth of underwater communication technology in the US Navy,” R. Headrick and L. Freitag, IEEE Commun. Mag., vol. 47, no 1, pp. 80-82, Jan. 2009.

[15]“Underwater networking acoustic techniques,” R. Otnes, A. Asterjadhi, P. Casari, M. Goetz, T. Husøy, I. Nissen, K. Rimstad, P. Van Walree, M. Zorzi, Springer 2012.

[16]“Underwater sensor networks : applications advances and challenges”, J. Heideman, M. Stojanovic, M. Zorzi, Transactions of the Royal Society Press, 2012.

Un saludo!

## Comentarios recientes

hace 18 semanas 6 días

hace 21 semanas 3 días

hace 21 semanas 6 días

hace 21 semanas 6 días

hace 21 semanas 6 días

hace 24 semanas 1 día

hace 26 semanas 2 días

hace 28 semanas 5 días

hace 38 semanas 1 día

hace 40 semanas 4 días