We study the impact of reliability mechanisms introduced at the link layer on the performance of transport protocols in the context of 4G satellite links. Specifically, we design a software module that performs realistic analysis of the network performance, by utilizing real physical layer traces of a 4G satellite service. Based on these traces, our software module produces equivalent link layer traces, as a function of the chosen link layer reliability mechanism. We further utilize the link layer traces within the ns-2 network simulator to evaluate the impact of link layer schemes on the performance of selected TCP variants. We consider erasure coding, ARQ and Hybrid-ARQ link layer mechanisms, and TCP Cubic, Compound, Hybla, New Reno and Westwood. We show that, for all target TCP variants, when the throughput of the transport protocol is close to the channel capacity, using the ARQ mechanism is most beneficial for TCP performance improvement. In conditions where the physical channel error rate is high, Hybrid-ARQ results in the best performance for all TCP variants considered, with up to 22% improvements compared to other schemes.

This paper introduces a new dictionary learning strategy based on atoms obtained by translating the composition of K convolutions with S-sparse kernels of known support. The dictionary update step associated with this strategy is a non-convex optimization problem. We propose a practical formulation of this problem and introduce a Gauss–Seidel type algorithm referred to as alternative least square algorithm for its resolution. The search space of the proposed algorithm is of dimension KS, which is typically smaller than the size of the target atom and much smaller than the size of the image. Moreover, the complexity of this algorithm is linear with respect to the image size, allowing larger atoms to be learned (as opposed to small patches). The conducted experiments show that we are able to accurately approximate atoms such as wavelets, curvelets, sinc functions or cosines for large values of K. The proposed experiments also indicate that the algorithm generally converges to a global minimum for large values of K and S.

This paper studies a new Bayesian algorithm for fusing hyperspectral and multispectral images. The observed images are related to the high spatial resolution hyperspectral image to be recovered through physical degradations, e.g., spatial and spectral blurring and/or subsampling defined by the sensor characteristics. In this work, we assume that the spectral response of the multispectral sensor is unknown as it may not be available in practical applications. The resulting fusion problem is formulated within a Bayesian estimation framework, which is very convenient to model the uncertainty regarding the multispectral sensor characteristics and the scene to be estimated. The high spatial resolution hyperspectral image is then inferred from its posterior distribution. More precisely, to compute the Bayesian estimators associated with this posterior, a Markov chain Monte Carlo algorithm is proposed to generate samples asymptotically distributed according to the distribution of interest. Simulation results demonstrate the efficiency of the proposed fusion method when compared with several state-of-the-art fusion techniques.

This paper addresses the problem of ultrasound image restoration within a Bayesian framework. The distribution of the ultrasound image is assumed to be a generalized Gaussian distribution (GGD). The main contribution of this work is to propose a hierarchical Bayesian model for estimating the GGD parameters. The Bayesian estimators associated with this model are difficult to be expressed in closed form. Thus we investigate a Markov chain Monte Carlo method which is used to generate samples asymptotically distributed according to the posterior of interest. These generated samples are finally used to compute the Bayesian estimators of the GGD parameters. The performance of the proposed Bayesian model is tested with synthetic data and compared with the performance obtained with the expectation maximization algorithm.

Les produits d’intermodulation passifs ne suivent pas la loi classique d’augmentation de puissance de 3 dB par dB de puissance d’entrée. Un modèle basé sur des fonctions non-analytique permet de simuler correctement ce comportement pour deux ou plusieurs porteuses. Le modèle explique l’amélioration du rapport C/I lorsque le nombre de porteuses augmente et permet de calculer cette amélioration à partir des mesures à deux porteuses et de la pente. Ceci permet de relâcher les spécifications à deux porteuses de 4 dB si la pente est de 2,5 dB/dB et de 8 dB si la pente est de 2 dB/dB pour une même performance en multi-porteuse. Cette relaxation peut permettre d’utiliser une technologie d’antenne de masse ou de coût plus faible alors qu’elle n’aurait pas été acceptée en l’absence de modèle.

We propose to use memristors as memory nonlinear circuits to build behavioral models useful in the simulation of passive inter-modulation in RF and microwave devices such as filters, antennas and in general connections

The orbit of a satellite around the earth is constantly disturbed by various factors, such as variations in the gravitational field and the solar wind pressure. The drift of the satellite position can compromise the mission, and even lead to a crash or a fall in the atmosphere. The station-keeping operations therefore consist in performing an accurate measurement of the satellite trajectory and then in using its thrusters to correct the drift. The conventional solution is to measure the position with the help of a ground based radar. This solution is expensive and does not allow to have the satellite position permanently : the trajectory corrections are therefore infrequent. A positioning and autonomous navigation system using constellations of navigation satellites, called Global Navigation Satellite System (GNSS), allows a significant reduction in design and operational maintenance costs. Several studies have been conducted in this direction and the first navigation systems based on GPS receivers, are emerging. A receiver capable of processing multiple navigation systems, such as GPS and Galileo, would provide a better service availability. Indeed, Galileo is designed to be compatible with GPS, both in terms of signals and navigation data. Continuous knowledge of the position would then allow a closed loop control of the station keeping. Initially, we defined what the specifications of a multi-mission space receiver are. Indeed, the constraints on such a receiver are different from those for a receiver located on the surface of the Earth. The analysis of these constraints, and the performance required of a positioning system, is necessary to determine the specifications of the future receiver. There are few studies on the subject. Some of them are classified ; others have, in our view, an analytical bias that distorts the determination of specifications. So we modeled the system : GNSS and receivers satellite orbits, radio frequency link. Some parameters of this link are not given in the specification or manufacturers documents. Moreover, the available theoretical data are not always relevant for realistic modeling. So we had to assess those parameters using the available data. The model was then used to simulate various scenarios representing future missions. After defining analysis criteria, specifications were determined from the simulation results. Calculating a position of a satellite navigation system involves three main phases. For each phase, there are several possible algorithms, with different performance characteristics, the circuit size or the computation load. The development of new applications based on navigation also drives the development of new adapted algorithms. We present the principle for determining a position, as well as GPS and Galileo navigation signals. From the signal structure, we explain the phases of the demodulation and localization. Through the use of GPS and Galileo constellations, standard algorithms achieve the performance required for space applications. However, these algorithms need to be adapted, thus some parts were specifically designed. In order to validate the choice of algorithms and parameters, we have simulated the various operating phases of the receiver using real GPS signals. Finally, impact and prospects are discussed in the conclusion.

The orbit of a satellite around the earth is constantly disturbed by various factors, such as variations in the gravitational field and the solar wind pressure. The drift of the satellite position can compromise the mission, and even lead to a crash or a fall in the atmosphere. The station-keeping operations therefore consist in performing an accurate measurement of the satellite trajectory and then in using its thrusters to correct the drift. The conventional solution is to measure the position with the help of a ground based radar. This solution is expensive and does not allow to have the satellite position permanently : the trajectory corrections are therefore infrequent. A positioning and autonomous navigation system using constellations of navigation satellites, called Global Navigation Satellite System (GNSS), allows a significant reduction in design and operational maintenance costs. Several studies have been conducted in this direction and the first navigation systems based on GPS receivers, are emerging. A receiver capable of processing multiple navigation systems, such as GPS and Galileo, would provide a better service availability. Indeed, Galileo is designed to be compatible with GPS, both in terms of signals and navigation data. Continuous knowledge of the position would then allow a closed loop control of the station keeping. Initially, we defined what the specifications of a multi-mission space receiver are. Indeed, the constraints on such a receiver are different from those for a receiver located on the surface of the Earth. The analysis of these constraints, and the performance required of a positioning system, is necessary to determine the specifications of the future receiver. There are few studies on the subject. Some of them are classified ; others have, in our view, an analytical bias that distorts the determination of specifications. So we modeled the system : GNSS and receivers satellite orbits, radio frequency link. Some parameters of this link are not given in the specification or manufacturers documents. Moreover, the available theoretical data are not always relevant for realistic modeling. So we had to assess those parameters using the available data. The model was then used to simulate various scenarios representing future missions. After defining analysis criteria, specifications were determined from the simulation results. Calculating a position of a satellite navigation system involves three main phases. For each phase, there are several possible algorithms, with different performance characteristics, the circuit size or the computation load. The development of new applications based on navigation also drives the development of new adapted algorithms. We present the principle for determining a position, as well as GPS and Galileo navigation signals. From the signal structure, we explain the phases of the demodulation and localization. Through the use of GPS and Galileo constellations, standard algorithms achieve the performance required for space applications. However, these algorithms need to be adapted, thus some parts were specifically designed. In order to validate the choice of algorithms and parameters, we have simulated the various operating phases of the receiver using real GPS signals. Finally, impact and prospects are discussed in the conclusion.

Undoubtedly, the idea of global network is the founding concept of the next generation of communication systems. In the future, a user should therefore be able to maintain an excellent quality of communication regardless of where he is or moves without noticing the underlying technology changes. As a niche market of communications, be part of this dynamic is crucial for satellite communications. However, in the recent years, many studies have shown that the integration of the satellite segment in a global context was complicated by the different characteristics of considered technologies. In fact, some of the most important protocols such as TCP that warranty the quality of a communication are strongly suffering from the RTT duration and so are inadequate for the satellite link. Since the 2000s, satellite community therefore deployed specific solutions in the form of TCP-PEP. They offer very good performance, but marginalize the satellite link from the others by breaking the essential concept of end-to-end communication, and then make its integration among other technologies difficult. Also, if a solution that enables a better integrability and efficiency is not found, the satellite may be excluded from this ambitious project, despite its numerous strengths. We first conducted an extensive set of studies regarding the behavior of TCP latest flavors initiated by the main Operating Systems. They suggest that relevant end-to-end solutions, designed to fit terrestrial networks, can eventually offer similar performance in a satellite environment than the TCP-PEP solution. However, these optimisations only improve long-lived connections. The poor performance of short-lived connections, that are a majority in the Internet, continues therefore to justify the use of TCP-PEP. Consequently, we focused on improving end-to-end transport protocols for short-lived connections and proposed a mechanism called Initial Spreading that allows significant performance improvements regardless of the context. Its simple concept aims to overcome the RTT dependence that strongly damages the short-lived flows by emitting a large amount of data segments just after the connection establishment. We pay close attention to the consequences of releasing a large group of segments (burst) in a congested network. So while solutions such as RFC 6928, proposed by Google, see their performance sharply deteriorated in such an environment, our mechanism ensures very good performance using an accurateand regulated spreading of the first sent segments. Many simulations in NS2 first allowed us to validate the usefulness and scope of our mechanism. A mathematical model of short-lived TCP connections then allowed us to corroborate these results and to understand in an accurate way the consequences of the transmission of bursts of segments on the average performance of a communication. Finally, we implemented the Initial Spreading in the Linux kernel in order to test its behavior and efficiency in terrestrial and satellite networks and show the merits of our proposal in a real environment. All these evaluations allowed us to refine our mechanism to significantly improve the performance of short-lived TCP connections regardless of the context in question and the state of the network. We finally submitted our proposal to the IETF in the form of an “Internet Draft”.