Are you an EPFL student looking for a semester project?

Work with us on data science and visualisation projects, and deploy your project as an app on top of GraphSearch.

Graph
Search

fr|en

Concept

Signal

Related publications (100)

Design and development of a haptic feedback demonstrator using time-reversal method and FPGAS

Andrés Chica Linares

The EPFL-LAI is investigating on haptic technologies to give vibrotactile feedback in digital musical interfaces. The aim is to develop an interactive surface that can render a multi-touch vibrotactile stimulus using piezoelectric actuators. The signals that will be sent to the piezoelectric system are generated using deep-learning strategies. Then, they will be simultaneously sent using a FPGA. After this stage there will be a power supply circuit to amplify the digital signals and finally send them to the piezoelectric system. When using these haptic technologies, the sensations of playing a real instrument will be recreated and the player will enjoy the digital instrument through the touch as well. This will make the experience of playing a digital instrument richer.

2021

Nondyadic and nonlinear multiresolution image approximations

Maria Arrate Munoz Barrutia

This thesis focuses on the development of novel multiresolution image approximations. Specifically, we present two kinds of generalization of multiresolution techniques: image reduction for arbitrary scales, and nonlinear approximations using other metrics than the standard Euclidean one. Traditional multiresolution decompositions are restricted to dyadic scales. As first contribution of this thesis, we develop a method that goes beyond this restriction and that is well suited to arbitrary scale-change computations. The key component is a new and numerically exact algorithm for computing inner products between a continuously defined signal and B-splines of any order and of arbitrary sizes. The technique can also be applied for non-uniform to uniform grid conversion, which is another approximation problem where our method excels. Main applications are resampling and signal reconstruction. Although simple to implement, least-squares approximations lead to artifacts that could be reduced if nonlinear methods would be used instead. The second contribution of the thesis is the development of nonlinear spline pyramids that are optimal for lp-norms. First, we introduce a Banach-space formulation of the problem and show that the solution is well defined. Second, we compute the lp-approximation thanks to an iterative optimization algorithm based on digital filtering. We conclude that l1-approximations reduce the artifacts that are inherent to least-squares methods; in particular, edge blurring and ringing. In addition, we observe that the error of l1-approximations is sparser. Finally, we derive an exact formula for the asymptotic Lp-error; this result justifies using the least-squares approximation as initial solution for the iterative optimization algorithm when the degree of the spline is even; otherwise, one has to include an appropriate correction term. The theoretical background of the thesis includes the modelisation of images in a continuous/discrete formalism and takes advantage of the approximation theory of linear shift-invariant operators. We have chosen B-splines as basis functions because of their nice properties. We also propose a new graphical formalism that links B-splines, finite differences, differential operators, and arbitrary scale changes.

EPFL2002

Parallel architecture for high-speed analog-to-digital conversion

Guillaume Ding

Nowadays digital signal processing systems used for radar applications, communication systems or RF measurement equipments, require very high sample-rates. Sometimes these sample-rates are beyond the possibilities offered by conventional ADCs. To overcome these limits parallel architectures have been developed. The most commonly used it the one called "time-interleaved" conversion. This technique allows to achieve very-high sample-rates with circuits working at a lower frequency. The accuracy of "time-interleaved" systems is sensitive to sample-time errors. Some calibration techniques have been developed to reduce this sensitivity. They involve very sophisticated digital signal processing and, in most of the cases, they are not directly implemented on silicon but applied on measurement results in software. The goal of this thesis is to study the feasibility of a new parallel architecture for analog-to-digital conversion. This architecture must present a higher robustness to sample-time errors. The first part of this work is dedicated to time-interleaved converter. An analysis of their sensitivity to several imperfections, such as mismatches and systematic and random sample-time error is presented. This analysis is followed by a description of time-interleaved converter evolution, since the first implemented prototype to the current state of the art of the domain. The second part of this thesis focuses on the development of the new conversion technique called "frequency-interleaved". Two different approaches are studied: the first one is based on a Fourier series decomposition of the signal to convert and the second one is based on a Walsh series decomposition. During this study, theoretical and practical aspects are faced the one with the other, to combine signal processing and microelectronics together. It appears that the Fourier series approach offers modest performances and presents serious problems of implementation. Based on this study, the design of functional blocks of a Walsh series based system is proposed.

EPFL2009

An Iterative Algorithm for Joint Signal and Interference Alignment

Feng Xue

We consider the problem of signal design for transmission over an interference channel, along the lines of the recently proposed strategy of Interference Alignment. In the framework of interference alignment, each receiver maintains an interference subspace. The transmitters shape their transmissions such that all the interference seen at a particular receiver falls into its interference subspace, hence allowing the remaining receiver space to be used for reception of the useful signal. In contrast to previously proposed iterative interference alignment algorithms for the general interference channel, our proposed technique not only minimizes the interference power that is spilled outside the interference subspace, but also minimizes the signal power that is spilled over into the interference subspace. A weighted optimization technique is proposed to tradeoff between the relative importance associated with the leakage interference and the leakage signal powers. It is proven that the algorithm converges monotonically, and simulation results reveal that the proposed technique outperforms other known algorithms in terms of throughput. Further discussions on generalizing the algorithm are also presented.

IEEE Service Center, 445 Hoes Lane, Po Box 1331, Piscataway, Nj 08855-1331 Usa2010

The Global Navigation Satellite System (GNSS) refers to a constellation of satellites emitting signals from space, used to provide Position, Navigation and Timing (PNT) services to the receivers on Earth. Nowadays, the GNSS is exploited in a wide range of applications among which ground mapping, transportation, machine control, timing, surveying, defense, or aerial photogrammetry [1]. However, as we become more dependent on GNSS technology, we also become more vulnerable to its limitations [2]. The GNSS operates with satellites orbiting in Medium Earth Orbits (MEO, approximately 20200 [km] in altitude). On one hand, this great distance allows to have a large footprint of the satellite signal on Earth, but on the other hand, it requires the signal to cross a wide path in the atmosphere before reaching the receiver. This results in a decrease of signal strength, which is not an issue in open-sky condition, but becomes a limitation in deep attenuation environments, with little or no service in urban canyon, in dense cities or indoors [3]. In this contribution, a possible strategy to cope with this issue is proposed and evaluated through a simulation process. It implies the use of Low Earth Orbits (LEO) satellites as a support for the GNSS PNT service. Since nowadays no broadband big LEO constellation broadcasting navigation messages exist, in this project various LEO constellation setups have been simulated and tested. Because LEOs are much closer to Earth (200 to 1500 [km]), they move faster in the sky, passing overhead in minutes instead of hours like MEOs. This gives rise to the possibility to exploit the Doppler effect to accurately locate the user. Therefore, three different PNT algorithms, employing the Kalman filter, are implemented and the evaluation of the localization quality is made at four distinct locations on the globe. The results show that, for most of the cases, compared to the standard GPS pseudorange only PNT, the addition of LEO pseudoranges already improve the localization quality. The performance of the PNT algorithm increases even more with the exploitation of both pseudoranges and pseudorangerates (Doppler) measurements. Nevertheless, the choice of constellation parameters such as satellites altitude, total number of satellites, orbit inclination or the frequency of the emitted signal, has a great impact on the magnitude of the improvement in the localization quality. The factor that has the most significant influence seems to be the signal frequency and for that reason it must be carefully chosen: the higher the better. Moreover, it has been demonstrated that at high elevation cut-off angles (denied environments), the addition of LEO constellation is a valid option and allows us to obtain an accurate position even when the GNSS only can not provide a solution. Finally, the LEO only system seems to be a promising alternative to cope with GNSS jamming or spoofing. However, the optimal LEO parameters providing an accurate solution for all the location on the globe has not yet been identified. In conclusion, the result of this effort is the awareness that, once the optimal parameters for the LEO system have been found, it is a valid option to augment the GNSS, or even as a standalone backup, especially when the Doppler-based positioning is applied. Nevertheless, we are still far from that end and much more research needs to be done.

2020

This report presents a semi-supervised method to jointly extract audio-visual sources from a scene. It consist of applying a supervised method to segment the video signal followed by an automatic process to properly separate the audio track. This approach starts with the user interaction to select the audio-visual target we want to cut. This labels the problem into two parts. One is the foreground, the object we aim to extract. The background is the remaining content which sets the other part. The segmentation method will be performed over the video signal to split its visual content into foreground and background. It is based on describing the video information as a volume. It is composed of pixels and relationships among them which set the features of the problem. This 3D structure can be interpreted as a graph. Therefore the segmentation is achieved by applying a Graph Cuts algorithm which provides the suitable results. In this work the audio separation process is an automatic task. Due to this we need to extract prior information from the audio-visual data available. Therefore a link between audio and video channels is deﬁned by an audio-visual motion map. This is a video sequence which provides the amount of synchronous motion with the audio track. An audio-based video diﬀusion method have been developed to obtain this kind of signal. Applying the video segmentation labelling to the audio-visual motion map we obtain the distribution of the amount motion into foreground and background. According to the range of these values it is possible to asses when the corresponding audio is on/oﬀ. At this point, audio samples of sources can be extract when the audio activity is due to only one. This is the prior information to learn audio models by means of spectral GMM estimation to assess the audio signals. Finally the separation method is applied. It consist of looking for the more suitable couple of states given the mixture spectrum of the total audio track. Therefore we achieve the audio separation goal. In this work we will study in detail all the stages towards the ﬁnal aim of our method, analyzing the performances of our approach.

2010

Graph Laplacians for Rotation Equivariant Convolutional Neural Networks

Martino Milani

A fundamental problem in signal processing is to design computationally efficient algorithms to filter signals. In many applications, the signals to filter lie on a sphere. Meaningful examples of data of this kind are weather data on the Earth, or images of the sky. It is then important to design filtering algorithms that are computationally efficient and capable of exploiting the rotational symmetry of the problem. In these applications, given a continuous signal

f: \mathbb S^2 \rightarrow \mathbb R

on a 2-sphere

\mathbb S^2 \subset \mathbb R^3

, we can only know the vector of its sampled values

\mathbf f \in \mathbb R^N:\ (\mathbf f)_i = f(\mathbf x_i)

in a finite set of points

\mathcal P \subset \mathbb S^2,\quad \mathcal P = \{\mathbf x_i\}_{i=0}^{n-1}

where our sensors are located. Perraudin et al. in \cite{DeepSphere} construct a sparse graph

G

on the vertex set

\mathcal P

and then use a polynomial of the corresponding graph Laplacian matrix

\mathbf L \in \mathbb R^{n\times n}

to perform a computationally efficient -

\mathcal O (n)

- filtering of the sampled signal

\mathbf f

. In order to study how well this algorithm respects the symmetry of the problem - i.e it is equivariant to the rotation group SO(3) - it is important to guarantee that the spectrum of

\mathbf L

and spectrum of the Laplace-Beltrami operator

\Delta_\mathbb S^2

are somewhat ``close''. We study the spectral properties of such graph Laplacian matrix in the special case of \cite{DeepSphere} where the sampling

\mathcal P

is the so called HEALPix sampling (acronym for \textbf Hierarchical \textbf Equal \textbf Area iso\textbf Latitude \textbf {Pix}elization) and we show a way to build a graph

G'

such that the corresponding graph Laplacian matrix

\mathbf L'

shows better spectral properties than the one presented in \cite{DeepSphere}. We investigate other different methods of building the matrix

\mathbf L

better suited to non uniform sampling measures. In particular, we studied the Finite Element Method approximation of the Laplace-Beltrami operator on the sphere, and how FEM filtering relates to graph filtering, showing the importance of non symmetric discrete Laplacians when it comes to non uniform sampling measures. We finish by showing how the graph Laplacian

\mathbf L'

proposed in this work improved the performances of DeepSphere in a well known classification task using different sampling schemes of the sphere, and by comparing the different Discrete Laplacians introduced in this work.

2019

Nowadays, a wide variety of terminals are proposed to nomadic users. Generally these terminals provide wireless communication operating at frequencies between one and a few GHz. For technical reasons, including multiple access to the communication channel and battery autonomy, these terminals transmit only during very short periods i.e. transmission bursts. For a direct observation of certain characteristics of the transmitted signals radiated by such terminals, only a few measurement setups exist. This article proposes such a novel real-time 3D electromagnetic field measurement instrument with direct visualization. The prototype used for validation is based on an array of probes regularly attached on a non-conductive rigid loop which is put into fast rotation around the terminal under test. To cite this article: J. Rioult et al., C. R. Physique 10 (2009). (C) 2008 Academie des sciences. Published by Elsevier Masson SAS. All rights reserved.

Elsevier2009

Parameter Estimation with GNSS-Reflectometry and GNSS Synthetic Aperture Techniques

Miguel Angel Ribot Sanfelix

Aside from intentional interference, multipath is the most significant error source for Global Navigation Satellite Systems (GNSS) receivers in many operational scenarios. In this thesis, we study the multipath estimation from two different perspectives: to retrieve useful information from it using GNSS-Reflectometry (GNSS-R) techniques; and to mitigate its effects or to estimate its direction-of-arrival (DOA) as well as the line-of-sight (LOS) signal¿s using synthetic aperture (SA) processing. The first part of the thesis focuses on precision bounds for GNSS-R techniques for ground-based receivers, in scenarios where a single antenna simultaneously receives the LOS signal and a specular reflection. First, we derive the Cramér-Rao bound (CRB) of the receiver¿s height and the reflection coefficient, with the latter depending on the surface¿s electrical properties. More specifically, we propose a CRB derivation applicable to GNSS-R techniques that make use of the phase information and long observation times, such as the interference pattern technique (IPT). The derivation is based on the parameter transformation of the Fisher information matrix. We study the dependence of the computed CRB on the scenario and the receiver bandwidth. The CRB results for the simulated scenarios are consistent with the precision reported for many GNSS-R techniques used in these scenarios. The proposed CRB is meant to benchmark and compare new and existing techniques. Besides the derived CRB, we propose an algorithm to obtain the maximum-likelihood (ML) estimator of the parameters of interest with the IPT: the segmented ML estimator (SML). The SML transforms a complex multivariate optimization problem into multiple simpler ones by dividing the parameter search space taking advantage of the cost function¿s particular structure. The SML is validated with simulated signal and asymptotically cross-validates the CRB results. The second part of the thesis is devoted to the study of the SA processing of GNSS signals. The goal is to estimate the DOA of the signals received, and mitigate errors in the navigation solution caused by interfering signals, such as multipath. We start by deriving the CRB for the SA context, as a function of the antenna trajectory. This CRB considers the effect of the antenna complex gain, and we show in simulations that it is possible to achieve meaningful DOA estimation only by changing the antenna¿s orientation. We continue by proposing a development framework built upon a signal tracking architecture integrating SA processing. Before any SA processing, it is necessary to estimate and compensate any carrier phase contribution not related to the antenna motion. To do so, we propose two new sequential techniques based on the extended Kalman filter (EKF): EKF1 and EKF2. Also, we develop an open-loop version of the proposed SA tracking architecture, more robust than its closed-loop counterpart. Finally, we validate the proposed architecture and SA-based techniques with synthetic GPS signals at first, and then with real signals, recorded using an antenna mounted on a mechanical rotating arm. The obtained results validate the implemented techniques and show how the proposed SA architecture can ultimately mitigate the position bias error observed in environments with severe multipath interference.

EPFL2018

Depth matching well logs acquired from multiple logging passes in a single well has been a longstanding challenge for the industry. The existing approaches employed in commercial platforms are typically based on classical cross-correlation and covariance measures of two signals, followed by manual adjustments. These solutions do not satisfy the rising demand to minimize user intervention to proceed towards automated data interpretation. We aimed at developing a robust and fully automatic algorithm and workflow for depth matching gamma-ray logs, which are commonly used as a proxy to match the depth of other well logs measured in multiple logging passes within the same well. This was realized by a supervised machine learning approach through a fully connected neural network. The training dataset was obtained by manually labeling a limited set of field data. As it is unrealistic to expect a perfect model from the initial training with limited manually labeled data, we developed a continuously self-evolving depth-matching framework. During the use of depth-matching service, this framework allows taking the user input and feedback to further train and improve the depth-matching engines. This is facilitated by an automatic quality-control module for that we developed a dedicated metric by combining a few different algorithms. We use this metric to assess the quality of the returned results from the depth-matching engine. The users review the results and do manual adjustments if some intervals are not ideally depth matched by the engine. Those manual adjustments can be used to further improve the machine-learning model. A well-designed framework enables automatic and continuous self-evolving of the depth-matching service. A key aspect of the developed framework is its generalization potential because it is independent of the signal type. It could be easily extended for other log types, especially when the correlation thereof is not obvious, provided that a sufficiently large volume of labeled data is available. This framework has been prototyped and tested on field data.

SOC PETROPHYSICISTS & WELL LOG ANALYSTS-SPWLA2019