From the 80’s to today, all AIRBUS civil aircraft are equipped with electrical flight control systems (EFCS). This technology now constitutes an industrial standard for commercial applications. This allows a more sophisticated aircraft control (advanced flight laws, more available autopilot...) and the setting up of specific protection functions of the flight enveloppe. In the framework of a global aircraft optimisation, for future and upcoming programs, current research efforts are dedicated to a more easy-to-handle aircraft, more efficient and so on more environmentally-friendly, resulting in augmented EFCS availability. The industrial state of practice, for all aircraft manufacturers, is to develop high levels of hardware redundancy. Therefore several sensors (for instance three angle of attack probes, three pitot probes) provide flight parameter measurements which are necessary for the computation of the flight laws, as an example. For each of these measurements, a choice or computation is performed to provide a unique and valid value among the redundant sensors. In parallel, a monitoring is done to discard a measure in case of a failure. Both processes are called « consolidation ». The aim of the Ph.D. is to provide new detection strategies to detect a failure on each sensor (monosensor monitoring) and then to design new data fusion methods to act as the actual « consolidation » process. The main idea proposes to create « software » sensors which actually are flight parameter estimators (measured by external sensors) created thanks to other dissimilar flight parameters (in our case inertial parameters, measured by inner sensors, from a different technology). The partial least squares regression (PLS) is used to perform this estimation. Detection strategies and fusion methods are following from its properties.

From the 80’s to today, all AIRBUS civil aircraft are equipped with electrical flight control systems (EFCS). This technology now constitutes an industrial standard for commercial applications. This allows a more sophisticated aircraft control (advanced flight laws, more available autopilot...) and the setting up of specific protection functions of the flight enveloppe. In the framework of a global aircraft optimisation, for future and upcoming programs, current research efforts are dedicated to a more easy-to-handle aircraft, more efficient and so on more environmentally-friendly, resulting in augmented EFCS availability. The industrial state of practice, for all aircraft manufacturers, is to develop high levels of hardware redundancy. Therefore several sensors (for instance three angle of attack probes, three pitot probes) provide flight parameter measurements which are necessary for the computation of the flight laws, as an example. For each of these measurements, a choice or computation is performed to provide a unique and valid value among the redundant sensors. In parallel, a monitoring is done to discard a measure in case of a failure. Both processes are called « consolidation ». The aim of the Ph.D. is to provide new detection strategies to detect a failure on each sensor (monosensor monitoring) and then to design new data fusion methods to act as the actual « consolidation » process. The main idea proposes to create « software » sensors which actually are flight parameter estimators (measured by external sensors) created thanks to other dissimilar flight parameters (in our case inertial parameters, measured by inner sensors, from a different technology). The partial least squares regression (PLS) is used to perform this estimation. Detection strategies and fusion methods are following from its properties.

Many measurement results on passive intermodulation products exhibit slopes of third order intermodulation product level as a function of input level different from the classical 3 dB/dB slope. Even-integer and real values between 1 and 3 are commonly reported for telephony base station towers antennas and filters. No classical model has been able to approximate these measurements up to now. We propose a non-analytic model that explains this behaviour and may serve as theoretical basis to find a physical model.

Most nonlinear behavioural models of amplifiers are based on functions that are analytic at the origin and thus can be replaced by their Taylor series development around this point, e.g. polynomials of the input signal. Chebyshev Transforms can be used to compute the harmonic response of the model to a sine input signal. These responses are polynomials of the input signal amplitude. A second application of the Chebyshev transform to the first harmonic response or RF characteristic will lend the carriers and intermodulation (IM) products for a 2-carrier input signal, again polynomials. An important class of non-analytic nonlinear behaviour encountered in practice, such as hard limiters and detectors are either empirically treated or only approximated by an analytic function such as the hyperbolic tangent. This work proposes to generalize the polynomial nonlinearity theory by adding non-analytic at the origin functions that, like polynomials, are invariant elements of the Chebyshev Transform. Devices modelled with these non-analytic at the origin functions exhibit intermodulation behaviour significantly different from that of classical polynomial models, giving theoretical foundation to a number of important unexplained practical measurement observations.

Wideband radars are promising systems that may provide numerous advantages, like simultaneous detection of slow and fast moving targets, high range-velocity resolution classification, and electronic counter-countermeasures. Unfortunately, classical processing algorithms are challenged by the range-migration phenomenon that occurs then for fast moving targets. We propose a new approach where the range migration is used rather as an asset to retrieve information about target velocities and, subsequently, to obtain a velocity dealiased mode. More specifically three new complex spectral estimators are devised in case of a single low-PRF (pulse repetition frequency) wideband waveform. The new estimation schemes enable one to decrease the level of sidelobes that arise at ambiguous velocities and, thus, to enhance the discrimination capability of the radar. Synthetic data and experimental data are used to assess the performance of the proposed estimators.

Analysis and comparison of linear and hybrid/compact dual-polarization (dual-pol) synthetic aperture radar (SAR) imagery have gained a wholly new importance in the last few years, in particular, with the advent of new spaceborne SARs such as the Japanese ALOS PALSAR, the Canadian RADARSAT-2, and the German TerraSAR-X. Compact polarimetry, hybrid dual-pol, and quad-pol modes are newly promoted in the literature for future SAR missions. In this paper, we investigate and compare different hybrid/compact and linear dual-pol modes in terms of the estimation of the degree of polarization (DoP). The DoP has long been recognized as one of the most important parameters characterizing a partially polarized electromagnetic wave. It can be effectively used to characterize the information content of SAR data. We study and compare the information content of the intensity data provided by different hybrid/compact and linear dual-pol SAR modes. For this purpose, we derive the joint distribution of multilook SAR intensity images. We use this distribution to derive the maximum likelihood and moment-based estimators of the DoP in hybrid/compact and linear dual-pol modes.We evaluate and compare the performance of these estimators for different modes on both synthetic and real data, which are acquired by RADARSAT-2 spaceborne and NASA/JPL airborne SAR systems, over various terrain types such as urban, vegetation, and ocean.

Linear spectral unmixing is a challenging problem in hyperspectral imaging that consists of decomposing an observed pixel into a linear combination of pure spectra (or endmembers) with their corresponding proportions (or abundances). Endmember extraction algorithms can be employed for recovering the spectral signatures while abundances are estimated using an inversion step. Recent works have shown that exploiting spatial dependencies between image pixels can improve spectral unmixing. Markov random fields (MRF) are classically used to model these spatial correlations and partition the image into multiple classes with homogeneous abundances. This paper proposes to define the MRF sites using similarity regions. These regions are built using a self-complementary area filter that stems from the morphological theory. This kind of filter divides the original image into flat zones where the underlying pixels have the same spectral values. Once the MRF has been clearly established, a hierarchical Bayesian algorithm is proposed to estimate the abundances, the class labels, the noise variance, and the corresponding hyperparameters. A hybrid Gibbs sampler is constructed to generate samples according to the corresponding posterior distribution of the unknown parameters and hyperparameters. Simulations conducted on synthetic and real AVIRIS data demonstrate the good performance of the algorithm.

DVB-SH (Digital Video Broadcasting- Satellite Handled) is a hybrid satellite terrestrial broadcasting standard dedicated to provide video or audio services for handheld terminals. On the satellite part, this standard can make use of interleaving mechanisms to mitigate the effects of the Land Mobile Satellite (LMS) channel. As result, these mechanisms enables the in-time distribution of a codeword over a duration ranging from 100 ms to about 30 s, depending of their parameters. This mechanism significantly improves the error recovery performance of the code however, in the literature, a theoretical evaluation at system level of this improvement is missing. Moreover, carrying out Monte-Carlo simulations implementing real decoding processes on significant traveled distances is time prohibitive. We propose hereafter a prediction method compatible with fast simulations to quantitatively evaluate the system performance in function the Packet Error Rate (PER), Erroneous Second Ratio (ESR) and zapping time. This method is based on the computation of the mutual information between emitted and received symbols for QPSK modulation and turbo coding. We demonstrate that our method reaches a prediction precision of the order of 0.1 dB, which is significantly better than two classical prediction methods. Moreover, our solution reduces the simulation time by a factor of 500 compared to Monte-Carlo. Beyond DVB-SH application, the presented approach can be applied in a large panel of satellite mobile systems and is completely new for the satellite community.