Arthur Hugeat : " Optimization of Radiofrequency Digital Signal Processing Chains and Application to a Cascade of Filters

Arthur HUGEAT PhD Works  : "Optimization of Radiofrequency Digital Signal Processing Chains and Application to a Cascade of Filters"

Abstract  :

The thesis tackles along two axis the challenge of optimizing digital implementations of radiofrequency filters and signal processing chains for stable oscillator noise characterization: on one hand, from the digital signal processing perspective, the resolution and resources occupied by efficient algorithms to analyze signals from ultra stable oscillators (phase, frequency) are characterized. We are not interested here in the implementation of these algorithms but analyze the different characteristics such as input, output, resource, latency... Then, on the other hand, having identified these characteristics, they are exploited in the context of an optimal scheduling search to obtain the best performances (in terms of resolution and bandwidth) of a given hardware platform designed to collect data from ultra stable oscilators. The question of the limitation induced by the numerical resolution of the processing steps is approached with regard to the intrinsic noise of the components (analog-todigital converters, clocks, voltage reference noise, jitter on sampling). This question has already been dealt with as part of another thesis defended at the Time-Frequency department. Thus, the thesis work considered here is clearly focused on the IT aspects of radiofrequency signal processing but by addressing classical problem solving and scheduling problems in the specific context of time-frequency metrology (correlations, Fourier transform at various scales aiming at more than 6-orders of magnitude between the highest and the slowest frequency range - 10 MHz to less than 1 Hz from the carrier - anti-aliasing filters, phase extraction from sampled signals, digital local oscillator). To put into context the complexity of the task, let us recall that a noise level of -160 dBrad²/Hz results relevant to 16 decimal places in the calculation result.

Jury Commposition :

M. Frédéric Pétrot, Professeur à l'Institut Polytechnique de Grenoble TIMA, Reviewer
M. Tanguy Risset, Professeur à l'Insa-Lyon Département Télécommunications, Services et Usages, Reporter
M. Emmanuel Casseau, Professeur à l'Université de Rennes, INRIA, Reporter
M. Jean-Michel Friedt, Maître de Conférence HDR à Université de Franche-Comté, PhD Director
M. Julien Bernard, Maître de Conférence à l'Université de Franche-Comté, PhD  Co-Director
M. Pierre-Yves Bourgeois, Chargé de Recherche CNRS à FEMTO-ST, PhD  Co-Director

Localization : Amphithéâtre J-J Gagnepain, FEMTO-ST- TEMIS - 15B, Avenue des Montboucons - 25000 Besançon