Recherche
Article de journal
Foldings of Periodic Nonuniform Samplings
IEEE Transactions on Circuits and Systems II, vol. 69, issue 3, pp. 1862-1868, March 2022.
Periodic Nonuniform Samplings of order N (PNSN) are interleavings of periodic samplings. For a base period T, simple algorithms can be used to reconstruct functions of spectrum included in an union of N intervals δk of length 1/T. In this paper we study the behavior of these algorithms when applied to any function. We prove that they result in N (or less) foldings on , each of δk holding at most one folding.
Traitement du signal et des images / Autre
Article de conférence
A Multiscale Anomaly Detection Framework for AIS Trajectories via Heat Graph Laplacian Diffusion
In Proc. 32nd EUropean SIgnal Processing COnference (EUSIPCO 2024), Lyon, France, August 26-30, 2024.
The monitoring of abnormal ship behavior is an important task for maritime surveillance for which the automatic identification system (AIS) has been widely exploited. Several works have proposed graph-based anomaly detection (AD) methods on spatial AIS points to provide further information regarding the interactions between the observed data through graph structures. This paper studies a new AD framework on graphs constructed from AIS trajectories. This framework considers a diffusion kernel at multiple scales of the graph Laplacian matrix, referred to as multiscale AD for AIS trajectories (MADAIS). MAD-AIS builds an attributed graph from a set of AIS trajectories, where nodes encode spatio-temporal trajectories and edges connect them via a similarity measure. In a second stage, AD is performed by computing scaled versions of the graph Laplacian matrix that are used to assess the graph connectivity. Simulation results are first conducted on synthetic data with controlled ground truth showing that the proposed MAD-AIS can effectively detect the abnormal behavior of ships in terms of spatio-temporal irregularities. Simulations conducted on real AIS subtrajectories (i.e., segments of AIS trajectories) show that abnormal features/attributes can be localized along AIS paths.
Traitement du signal et des images / Observation de la Terre et Autre
Improving AI Monitoring of Early Life Satellites Using Transfer Learning
In Proc. 17th International Conference on Space Operations (SpaceOps), Dubaï, United Arab Emirates, March 6-10, 2023.
In the last decades, many space domain actors such as the Centre National d’Etudes Spatiales (CNES) have begun to use Artificial Intelligence to monitor spacecraft housekeeping telemetry. These novel techniques are able to identify atypical behaviours and potential satellite anomalies that cannot be detected by more standard monitoring approaches. However, AI methods have an important drawback: they need a significant amount of data to be able to “learn” the nominal behaviour of a spacecraft and then detect novelties in new telemetry, which is not suitable for a satellite in the beginning of life where in-flight telemetry is very scarce. One way to bypass the scarcity of data is Transfer Learning (TL). Depending on the use case, operators may have already-available telemetry either from on-the-ground Assembly, Integration, and Test (AIT) of the spacecraft, from full-digital or hybrid simulators, or from in-flight telemetry of one or multiple “twin-spacecraft” in case of a constellation with already-launched units. This already-available telemetry is often close, but not perfectly similar, to in-flight telemetry of the newly-launched spacecraft to be monitored. The idea of TL is therefore to use this large and existing database (the source database), coupled with the first in-flight telemetry from the new spacecraft (the target database), to be able to mathematically-design a relevant AI learning model. In 2022, CNES and TéSA laboratory have worked together and have identified two TL methods to detect anomalies in telemetry of early life satellites with few data, by working directly on the telemetry dataset (the learning domain) or on the model learned from the target database. The first TL method consists in mathematically modifying the decision boundary estimated by a One-Class Support Vector Machine (OC-SVM) algorithm applied to the source database to match the target database. The second method based on “Domain Transfer” consists in building an “extended” learning domain made up with the relevant data from both the source and target databases, which is used to build a learning model. These two algorithms have been evaluated with real Earth Observation satellite telemetry. The preliminary outcomes of this research show promising results. Further work will consist in implementing these methods operationally so that AI monitoring methods can be used from the very beginning-of-life of CNES satellites. The main conclusion of this work is that TL can be an interesting tool to monitor spacecraft housekeeping telemetry during the first 6 months after the launch of a satellite.
Traitement du signal et des images / Systèmes spatiaux de communication
Modèles de Markov cachés appliqués au masquage de perte de paquets en voix sur IP
In Proc. Groupement de Recherche en Traitement du Signal et des Images (GRETSI), September 8-11, 2009.
Packet loss due to misrouted or delayed packets in voice over IP leads to huge voice quality degradation. This paper presents a packet loss concealement algorithm which is inde pendant from the vocoder. This method relies on hidden Marko v model (HMM). A new voicing parameter is introduced to get ove r voiced/unvoiced sound separation and use a unique HMM. Since best parameter for prediction are not necessarily the best ones for synthesis, we introduce two separate vectors: the first one dedicated to the analysis of the signal and the second one featured for the synthesis of missing part. Performances of the proposed system are evaluated on parts of well-known speech corpora, leading to promising results.
Traitement du signal et des images / Systèmes spatiaux de communication
A New Feature Vector for HMM-Based Packet Loss Concealment
In Proc. European Signal and Image Processing Conference (EUSIPCO), Glasgow, Scotland, August 24-28, 2009.
Packet loss due to misrouted or delayed packets in voice over IP leads to huge voice quality degradation. Packet loss con- cealment algorithms try to enhance the quality of the speech. This paper presents a new packet loss concealment algorithm which relies on one hidden Markov model. For this purpose, we introduce a continuous observation vector well-suited for silence, voiced and unvoiced sounds. We show that having a global HMM is relevant for this application. The proposed system is evaluated using standard PESQ score in a real- world application.
Traitement du signal et des images / Autre
A Continuous Voicing Parameter in the Frequency Domain
In Proc. Int. Conf. on Speech and Computer (SPECOM), Saint Petersbourg, Russia, June 23-25, 2009.
In automatic speech analysis, voicing articulatory is often defined as a binary decision: voiced or unvoiced. Lin- guists agree that this articulatory should be continuous. In this paper, we present a new approach to compute a con- tinuous voicing indicator of a speech frame. This voic- ing percentage is then evaluated in both a segmentation process and a speech recognition task. Promising results show that this continuous voicing percentage may be used as a reliable voicing indicator.
Traitement du signal et des images / Autre
ADRESSE
7 boulevard de la Gare
31500 Toulouse
France