Leveraging carrier phase observations within Global Navigation Satellite Systems receivers allows centimeter-level positioning accuracy. However, carrier phase observations are significantly affected by additive noise, which is assumed to follow a von Mises distribution, thereby degrading the performance of phase-based positioning estimators. To improve the modeling of carrier phase observations, we propose a novel approach that constrains the parameters of the von Mises distri-bution-specifically, the angular location modeling the phase and its dispersion parameter $\kappa$ modeling the noise-to evolve within the Lie group space $S O(2) \times \mathbb{R}^{+}$. To estimate these parameters, we employ a Lie group maximum likelihood estimator, solved through a Newton algorithm on Lie groups. This approach demonstrates advantages in terms of robustness and precision, especially when dealing with a small number of observations, compared to traditional Euclidean-based methods.

Plug-and-play algorithms have shown impressive results on imaging inverse problems, such as registration, super-resolution, denoising and inpainting. These methods rely on a neural network denoiser to learn an implicit prior of the image to be estimated. This paper investigates a new plug-and-play approach for 3D point cloud registration, which is crucial for a wide range of applications such as urban planning, archaeology and autonomous vehicles. The 3D point cloud registration problem is formulated as an inverse problem whose unknowns are the image to be estimated and the transformation between the two point clouds. A plug-and-play approach using an alternating optimization strategy is proposed for solving the registration problem. Experiments conducted on synthetic data and Li-DAR point clouds are presented showing the potential of the method.

This work studies a new Expectation-Maximization (EM) algorithm for solving the 2D-3D registration problem, which consists of estimating the position and orientation of a camera using a 3D map and a 2D image of the same scene. This algorithm associates each image feature coordinate to one vector of the 3D map using the pinhole camera model or to a class of outliers, making the registration robust to the presence of abnormal image features. It iteratively improves the camera pose by estimating the associations between the image features and the 3D map coordinates (using a robust mixture model) and minimizing the reprojection errors between the image and map points. Experimental results demonstrate that the proposed EM algorithm achieves competitive results in both absolute position and orientation compared to the Iterative Closest Point (ICP) approach.

This research paper proposes adaptations of three state-of-the-art anomaly detection algorithms, (One-Class Support Vector Machine, Isolation Forest and Local Outlier Factor), for detecting abnormal behavior in ship trajectories in an unsupervised way. These algorithms are adapted and tested using a wide range of similarity measures built specifically for time series, such as Dynamic Time Warping. The proposed methods are first applied on synthetic Automatic Identification System datasets with available ground truth. Then, they are generalized to handle pairs of Automatic Identification System and radar trajectories to detect unexpected activities, such as route deviations, delays and entering prohibited zones. The performances of the proposed methods are shown to be competitive when compared to the state-of-the-art for abnormal ship behavior detection.

Sparse regularisation has proven its worth and effectiveness in many fields, such as medical imaging. In this sense, nuclear magnetic resonance spectroscopy (MRS) is one of the modalities that could greatly benefit from sparse regularisation. This paper introduces a novel Bayesian approach for MRS restoration that accounts for possible errors in the observation linear operator. The algorithm is tailored to the complex nature of MRS data, incorporating both real and imaginary parts of the spectrum. An MCMC (Markov chain Monte Carlo) inference is conducted using a Gibbs sampler strategy. The method has been successfully validated on both synthetic and clinical data of high-grade brain tumor glioblastoma (GBM) patients. This study will enable further analysis of metabolites of interest not conventionally considered in clinics because of their undetectable concentration.

This letter explores time-delay and Doppler estimation in the presence of unknown heavy-tailed disturbance. Conventional methods for achieving optimal mean squared error performance rely on the maximum likelihood estimator (MLE), which is consistent and asymptotically efficient under the unrealistic assumption of a perfect a-priori knowledge of the noise distribution. However, in practical situations, the noise distribution is often unknown, and classical parametric estimation procedures are no longer able to guarantee the statistical efficiency. In this work, by relying on the semiparametric theory, we present an original rank-based and distribution-free $R$-estimator which have the remarkable property to be parametrically efficient, i.e. it attains the “classical” Cramér-Rao Bound, irrespective of the unknown noise distribution, provided that the latter belongs to the family of Complex Elliptically Simmetric (CES) distributions.

The estimation of high resolution spectrometer Instrument Spectral Response Functions (ISRFs) is crucial because an imperfect knowledge of these functions can induce errors in the measurements. The state-of-the-art for this problem currently relies on the use of parametric models, which frequently lack flexibility to accurately model real-world ISRFs. To address this limitation, this paper proposes and investigates the use of sparse representations for modeling and estimating ISRFs, where the ISRFs are decomposed in a fixed dictionary of atoms. To estimate the sparse coefficient vector, a novel sparsity inducing regularization of the problem based on quadratic envelopes is studied and compared to the classical LASSO estimator and to a greedy method based on the Orthogonal Matching Pursuit (OMP) algorithm. Results for simulated ISRFs from the MicroCarb mission indicate that the proposed spectral representations yield excellent ISRF estimates, and that the use of quadratic envelopes can yield significantly better precision than competing methods.

In this talk, we present a hybrid approach to predicting the future performance of a fuel cell. The idea is to combine the physical laws governing system dynamics with machine learning algorithms. The data used come from an aging campaign carried out under a dynamic load profile and variable operating conditions. The proposed approach is based on a simplified formulation of the well-known quasi-static model. Evolution laws are proposed for certain parameters related to activation and diffusion losses. The ohmic resistance, which determines the ohmic losses, is modeled by random forests. To estimate the evolution of other parameters over time, these methods are used in conjunction with an extended Kalman filter (EKF). To predict future performance, a long-term memory is used to learn the evolution of these parameters, estimated by the EKF.