A statistical model for detecting changes in remote sensing images has recently been proposed in (Prendes et al., 2014, 2015). This model is sufficiently general to be used for homogeneous images acquired by the same kind of sensors (e.g., two optical images from Pléiades satellites, possibly with different acquisition conditions), and for heterogeneous images acquired by different sensors (e.g., an optical image acquired from a Pléiades satellite and a synthetic aperture radar (SAR) image acquired from a TerraSAR-X satellite). This model assumes that each pixel is distributed according to a mixture of distributions depending on the noise properties and on the sensor intensity responses to the actual scene. The parameters of the resulting statistical model can be estimated by using the classical expectation-maximization algorithm. The estimated parameters are finally used to learn the relationships between the images of interest, via a manifold learning strategy. These relationships are relevant for many image processing applications, particularly those requiring a similarity measure (e.g., image change detection and image registration). The main objective of this paper is to evaluate the performance of a change detection method based on this manifold learning strategy initially introduced in (Prendes et al., 2014, 2015). This performance is evaluated by using results obtained with pairs of real optical images acquired from Pléiades satellites and pairs of optical and SAR images.

Global Navigation Satellite Systems (GNSS) are increasingly present in our everyday life. New users are emerging with further operational needs implying a constant evolution of the current GNSS systems. A significant part of the new applications are found in environments with difficult reception conditions such as urban areas, where there are many obstacles such as buildings or trees. Therefore, in these obstructed environments, the signal emitted by the satellite is severely degraded. The signal received by the user has suffered from attenuations, as well as refractions and diffractions, making difficult the data demodulation and the user position calculation. GNSS signals being initially designed in an open environment context, their demodulaton performance is thus generally studied in the associated AWGN propagation channel model. But nowadays, GNSS signals are also used in degraded environments. It is thus essential to provide and study their demodulation performance in the urban propagation channel models. Nevertheless, GNSS modernization with new signals design such as GPS L1C or Galileo E1 OS, takes into account these new obstructed environments constraints. Since they have been designed especially for urban propagation channels [1], they are expected to have better demodulation performance compared with current GNSS signals. It is thus particularly interesting to firstly provide their demodulation performance in urban environments (not available in the literature) and secondly to compare it with the new GNSS signal designed in this PhD research context. In this way, their performance could be used as a benchmark for the future signals design. However these modernized signals are not yet available for this moment (for example, GPS L1C is expected to be operational over the entire constellation in 2026). It is thus essential to provide and study their demodulation performance in urban environments through simulations. It is in this context that this PhD thesis is related, the final goal being to improve GNSS signals demodulation performance in urban areas, proposing a new signal. In order to be able to provide and study GNSS signals demodulation performance in urban environments, a simulation tool has been developed : SiGMeP for "Simulator for GNSS Message Performance". It allows simulating the entire emission/reception GNSS signal chain in urban environment getting away from dependence of real signals availability, controlling the simulation parameters and testing new configurations. Existing and modernized signals demodulation performance has thus been computed with SiGMeP in urban environments. Since the classical way to compute and represent GNSS signals demodulation performance assumes an AWGN propagation channel model, and since the urban environments are really different from the AWGN channel, this classical method is not satisfactory in our urban context. Thus, in order to represent GNSS signals demodulation performance faithfully to reality, a new methodology more adapted to the user environment is proposed. It is based on the fundamental characteristics of a GNSS system, as well as on the urban environment impact on the received signal analysis. GNSS signals demodulation performance is thus provided in urban environments thanks to this new methodology, and compared with the classic methodology used in the AWGN case. Then, to improve GNSS signals demodulation performance in urban environments, many strategies are possible. However, the research axis on this thesis focuses on the "Channel Coding" aspect. It is thus this field which will be privileged to improve GNSS signals demodulation performance in urban environments. Each message, in addition to containing the useful information, carries redundant informaton, which is in fact the channel coding result, applying to the useful information. The message thus needs to be decoded at the reception. In order to decode the transmitted useful information, the receiver computes a detection function at the decoder imput. But the detection function used in classic receivers correspond to an AWGN propagation channel. This dissertation thus proposes an advanced detection function which is adapting to the propagation channel where the user is moving. This advanced detection function computation considerably improves demodulation performance, just in modifying the receiver part system. Finally, in order to design a new signal with better demodulation performance in urban environments than one of existing and future signals, a new LDPC channel code profile has been proposed, optimizing for a CSK modulation in an AWGN channel for iterative decoding. Indeed, the CSK modulation is a promising modulation in the spread signals world, which permits to free from limitations in terms of data rate implied by current GNSS signals modulations. Moreover, LDPC codes belong to the modern codes family, the first being able to approach the channel capacity.They thus represent promising achievable performance.

Global Navigation Satellite Systems (GNSS) are increasingly present in our everyday life. New users are emerging with further operational needs implying a constant evolution of the current GNSS systems. A significant part of the new applications are found in environments with difficult reception conditions such as urban areas, where there are many obstacles such as buildings or trees. Therefore, in these obstructed environments, the signal emitted by the satellite is severely degraded. The signal received by the user has suffered from attenuations, as well as refractions and diffractions, making difficult the data demodulation and the user position calculation. GNSS signals being initially designed in an open environment context, their demodulaton performance is thus generally studied in the associated AWGN propagation channel model. But nowadays, GNSS signals are also used in degraded environments. It is thus essential to provide and study their demodulation performance in the urban propagation channel models. Nevertheless, GNSS modernization with new signals design such as GPS L1C or Galileo E1 OS, takes into account these new obstructed environments constraints. Since they have been designed especially for urban propagation channels [1], they are expected to have better demodulation performance compared with current GNSS signals. It is thus particularly interesting to firstly provide their demodulation performance in urban environments (not available in the literature) and secondly to compare it with the new GNSS signal designed in this PhD research context. In this way, their performance could be used as a benchmark for the future signals design. However these modernized signals are not yet available for this moment (for example, GPS L1C is expected to be operational over the entire constellation in 2026). It is thus essential to provide and study their demodulation performance in urban environments through simulations. It is in this context that this PhD thesis is related, the final goal being to improve GNSS signals demodulation performance in urban areas, proposing a new signal. In order to be able to provide and study GNSS signals demodulation performance in urban environments, a simulation tool has been developed : SiGMeP for "Simulator for GNSS Message Performance". It allows simulating the entire emission/reception GNSS signal chain in urban environment getting away from dependence of real signals availability, controlling the simulation parameters and testing new configurations. Existing and modernized signals demodulation performance has thus been computed with SiGMeP in urban environments. Since the classical way to compute and represent GNSS signals demodulation performance assumes an AWGN propagation channel model, and since the urban environments are really different from the AWGN channel, this classical method is not satisfactory in our urban context. Thus, in order to represent GNSS signals demodulation performance faithfully to reality, a new methodology more adapted to the user environment is proposed. It is based on the fundamental characteristics of a GNSS system, as well as on the urban environment impact on the received signal analysis. GNSS signals demodulation performance is thus provided in urban environments thanks to this new methodology, and compared with the classic methodology used in the AWGN case. Then, to improve GNSS signals demodulation performance in urban environments, many strategies are possible. However, the research axis on this thesis focuses on the "Channel Coding" aspect. It is thus this field which will be privileged to improve GNSS signals demodulation performance in urban environments. Each message, in addition to containing the useful information, carries redundant informaton, which is in fact the channel coding result, applying to the useful information. The message thus needs to be decoded at the reception. In order to decode the transmitted useful information, the receiver computes a detection function at the decoder imput. But the detection function used in classic receivers correspond to an AWGN propagation channel. This dissertation thus proposes an advanced detection function which is adapting to the propagation channel where the user is moving. This advanced detection function computation considerably improves demodulation performance, just in modifying the receiver part system. Finally, in order to design a new signal with better demodulation performance in urban environments than one of existing and future signals, a new LDPC channel code profile has been proposed, optimizing for a CSK modulation in an AWGN channel for iterative decoding. Indeed, the CSK modulation is a promising modulation in the spread signals world, which permits to free from limitations in terms of data rate implied by current GNSS signals modulations. Moreover, LDPC codes belong to the modern codes family, the first being able to approach the channel capacity.They thus represent promising achievable performance.

Microsoft Kinect which has been primarily aimed at the computer gaming industry has been used in bio-kinematic research related implementations. A multi-Kinect system can be useful in exploiting spatial diversity to increase measurement accuracy. One of the main problems in deploying multi-Kinect systems is to estimate the pose, including the position and orientation of each Kinect. In this paper, a singular value decomposition (SVD) least-squares algorithm is extended to a more generic time-series based approach to solve this pose estimation problem utilising 3D positions of one or more joints in skeletons obtained from a multi-Kinect system. Additionally, computer simulations are performed to demonstrate the use and to evaluate the efficiency of the proposed algorithm. The former is further validated with a commercial Vicon system.

The installation of a Condition Monitoring System (CMS) on a mechanical machine (e.g., on a wind turbine) aims to reduce the operating costs by applying a predictive maintenance strategy. The CMS is composed of sensors acquiring signals from which system health indicators are computed and monitored. Part of those indicators are predefined depending on the monitored system kinematic and are computed by averaging large or narrow spectral bands. The averaging and the need for predefined thresholds for default detection may induce lots of false alarms while reducing the ability to detect the default early. To get precise health indicators reflecting each local meaningful spectral content, the AStrion software proposes a new data-driven monitoring strategy without any a priori on the measured signals. First, an automatic spectral analysis is applied to detect, characterize and classify the different spectral structures of the successive measured signals. These spectral structures can be either single spectral peaks, either peaks grouped in harmonic series or in modulation sidebands [1]. Second, these spectral structures are characterized by several features, including for example the number of peaks, the characteristic frequencies and the energy. This gives a snapshot of the system health at the signal acquisition time. To perform an automatic diagnosis of the system, the spectral evolution should be tracked along the time snapshots. In this paper, we propose a time tracking method based on McAulay & Quatieri algorithm [2] which has been designed originally for speech signals acquired on a continuous temporal basis. We have adapted [2] in order to account not only for single spectral peak evolution but also for the evolution of more complex structures such as harmonic series or modulation sidebands, even in the case of signals acquired on a non-regular temporal basis, as it is often the case. Moreover, an added sleep state makes the proposed method robust against nondetected spectral structures at a given time. Finally, the temporal evolution of the spectral structure features can be monitored and used as precise health indicators. The following figure is a result of the proposed method applied on real signals coming from a test bench designed in KAStrion project for simulating a wind turbine operation and for which the inner race of the main bearing has been damaged. Above, the time frequency map displays a zoom of the spectral peaks detected (around 20.000 per snapshot, represented by circles) and shows in blue the tracking from 44 to 189 operating hours of a spectral peak at 3.45 Hz. This particular peak evolves at 129 hours to become an harmonic series with more and more peaks and energy. Its energy evolution (plotted below) shows an increase which mirrors out a failure. In a following step [3], this spectral structure has been associated with the ball pass frequency of the inner ring of the main bearing. A dismantling of this bearing has confirmed the failure. This result shows the potential of the proposed data-driven method to create automatically relevant health indicators.

The mobile context of Mobile Wireless Sensor Networks (MWSN) limits the existence of a direct route from source to destination. A Disruption Tolerant Networking (DTN) architecture fits the requirements for such a context where messages need to be stored, carried and forwarded. For this kind of DTN applications, the goal is to achieve a high delivery ratio at low transmission cost with the lowest latency. Some DTN routing protocols use this ACK information to decrease the number of useless transmissions. Nevertheless in memory-constrained environments, the proportion of memory allocated to ACKs is a problem to study. This paper focuses on the use of acknowledgements (ACKs). We model the network with a Markov chain, to study the effect of ACKs on buffer time occupancy. Finally, an extensive set of simulations is run to analyse the influence of the memory proportion allocated to ACKs, on the performance.

The evolution of satellite communication towards packet-based communication and the adoption of modern physical layer techniques, such as Adaptive Coding and Modulation, raise new issues, brought by this variability, for scheduling algorithms. This paper addresses the problem of GSE packets scheduling over DVB-S2 with QoS support by adopting a utility-based scheduling strategy. Utility functions are considered as a mean to encompass parameters from several layers into a single scheduling process. By making use of a first-order approximation, a low complexity scheduling algorithm is derived from the utility functions optimization problem, adaptable to a wide range of functions. The algorithm determines the characteristics of the next BBFrame to be send, and which GSE packets it should contain. We focus particularly on delay as a joint metric to represent QoS requirements, and assess the performance of our algorithm and the relevance of our assumptions using comparative simulations.

This paper proposes a multi-tone signal pattern designed for accurate and easy measurements of nonlinear devices linearity factors of merit. The stimulus signal we propose ensures that DUT’s third order intermodulation products won’t overlap. Thus, the relative phases of source tones do not affect the amplitudes of intermodulation products. The usual metrics for linearity factors, ACPR (Adjacent Channel Power Ratio) or NPR (Noise Power Ratio), can be acquired with a greater accuracy only with amplitude measurements. This work has been carried out with a sampler-based receiver, using FFT (Fast Fourier Transform) filtering for tone separation.

Background: The aim of the Endocardial T-Wave Alternans Study was to prospectively assess the presence of T-wave alternans (TWA) or beat-to-beat repolarization changes on implantable cardioverterdefibrillator (ICD)-stored electrograms (EGMs) immediately preceding the onset of spontaneous ventricular tachycardia (VT) or fibrillation (VF). Methods: Thirty-seven VT/VF episodes were compared to 116 baseline reference EGMs from the same 57 patients. A Bayesian model was used to estimate the T-wave waveform in each cardiac beat and a set of 10 parameters was selected to segment each detected T wave. Beat-by-beat differences in each T-wave parameter were computed using the absolute value of the difference between each beat and the following one. Fisher criterion was used for determining the most discriminant T-wave parameters, then top-M ranked parameters yielding a normalized cumulative Fisher score > 95% were selected, and analysis was applied on these selected parameters. Simulated TWA EGMs were used to validate the algorithm. Results: In the simulation study, TWA was detectable even in the case of the smallest simulated alternans of 25 μV. In 13 of the 37 episodes (35%) occurring in nine of 16 patients, significant larger beat-to-beat variations before arrhythmia onset were detected compared to their respective references (median one positive episode per patient). Parameters including the T-wave apex amplitude seem the more discriminant parameters. Conclusions: Detection of beat-by-beat repolarization variations in ICD-stored EGMs is feasible in a significant subset of cases and may be used for predicting the onset of ventricular arrhythmias.

Spectral estimation is an important classical problem that has received considerable attention in the signal processing literature. In this contribution, we propose a novel method for estimating the parameters of sums of complex exponentials embedded in additive noise from regularly or irregularly spaced samples. The method relies on Kronecker’s theorem for Hankel operators, which enables us to formulate the nonlinear least squares problem associated with the spectral estimation problem in terms of a rank constraint on an appropriate Hankel matrix. This matrix is generated by sequences approximating the underlying sum of complex exponentials. Unequally spaced sampling is accounted for through a proper choice of interpolation matrices. The resulting optimization problem is then cast in a form that is suitable for using the alternating direction method of multipliers (ADMM). The method can easily include either a nuclear norm or a finite rank constraint for limiting the number of complex exponentials. The usage of a finite rank constraint makes, in contrast to the nuclear norm constraint, the method heuristic in the sense that the problem is non-convex and convergence to a global minimum can not be guaranteed. However, we provide a large set of numerical experiments that indicate that usage of the finite rank constraint nevertheless makes the method converge to minima close to the global minimum for reasonably high signal to noise ratios, hence essentially yielding maximum-likelihood parameter estimates. Moreover, the method does not seem to be particularly sensitive to initialization and performs substantially better than standard subspace-based methods.