Article
Engineering, Electrical & Electronic
Jose O. Vargas, Ricardo Adriano
Summary: A subspace-based conjugate-gradient method (S-CGM) is proposed in this article to improve the performance of the linearized CGM. By retrieving the deterministic part of the variational-induced current, the S-CGM estimates the total electric field more accurately, resulting in faster convergence speed and higher accuracy.
IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION
(2022)
Article
Computer Science, Artificial Intelligence
Bangti Jin, Zeljko Kereta
Summary: In this work, we study the use of stochastic gradient descent (SGD) for solving linear inverse problems in Banach spaces. SGD and its variants have proven to be highly successful optimization methods in various fields, such as machine learning, imaging, and signal processing. By using a single datum or a small subset of data at each iteration, SGD allows for highly scalable approaches to large-scale inverse problems. However, the theoretical analysis of SGD-based methods for inverse problems has been primarily limited to Euclidean and Hilbert spaces. This work presents a novel convergence analysis of SGD for linear inverse problems in general Banach spaces, demonstrating almost sure convergence to the minimum norm solution and establishing regularization properties for suitable stopping criteria. Numerical results are also provided to illustrate the effectiveness of the approach.
SIAM JOURNAL ON IMAGING SCIENCES
(2023)
Article
Thermodynamics
Andrzej Frackowiak, Agnieszka Wroblewska, Michal Cialkowski
Summary: This paper presents a concept of solving the inverse heat conduction problem using Trefftz functions and provides two examples to validate the effectiveness of the method.
INTERNATIONAL JOURNAL OF THERMAL SCIENCES
(2022)
Article
Mathematics, Applied
Keyvan Amini, Parvaneh Faramarzi
Summary: Spectral conjugate gradient methods are efficient for solving unconstrained optimization problems. This paper introduces a modified method and proves its global convergence for general nonlinear functions.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2023)
Article
Mathematics, Applied
Xiangli Li, Wenjuan Zhao, Xiaoliang Dong
Summary: The proposed new conjugate gradient algorithm based on the self-scaling memoryless BFGS update with Wolfe line search shows descent and global convergence under mild conditions. Numerical experiments demonstrate the effectiveness of this method.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2021)
Article
Mathematics, Applied
Ruixue Gu, Bo Han, Shanshan Tong, Yong Chen
Summary: This paper proposes and analyzes a novel method for solving inverse problems in Banach spaces, by combining homotopy perturbation iteration and Kaczmarz method with uniformly convex penalty terms. The method shows convergence in the exact data case and is able to reconstruct special features of solutions such as sparsity and piecewise constancy.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2021)
Article
Automation & Control Systems
Rui Wang, Yuesheng Xu
Summary: The paper systematically studies solutions of interpolation problems in Banach spaces, aiming to obtain explicit representer theorems for their solutions and establish fixed-point equation formulation. Unlike in a Hilbert space, solutions in Banach spaces may not be reduced to truly finite dimensional problems, but in the special case of the Banach space l(1)(N), this obstacle can be removed.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Mathematics, Applied
Jajati Keshari Sahoo, Pradeep Boggarapu, Ratikanta Behera, M. Zuhair Nashed
Summary: Mosic and Djordjevic introduced the notation of the gDMP inverse for a linear operator on Hilbert space in their 2018 paper by considering generalized Drazin inverse with the Moore-Penrose inverse. This article introduces two new classes of inverses: GD1 (generalized Drazin and inner) inverse and 1GD (inner and generalized Drazin) inverse for Banach space operators. The existence and uniqueness of the GD1 (also 1GD) inverse are discussed along with some properties through core-quasinilpotent decomposition and closed range decomposition operator. We further establish a few explicit representations of the GD1 inverse and their interconnections with generalized Drazin inverse. In addition, we discuss a few properties of GD1 (also 1GD) inverse through binary relation.
COMPUTATIONAL & APPLIED MATHEMATICS
(2023)
Article
Mathematics, Applied
Ruixue Gu, Hongsun Fu, Bo Han
Summary: This paper generalizes inexact Newton regularization methods to solve nonlinear inverse problems and can handle various types of noise. The method has fast convergence through the inner scheme and accelerated version.
Article
Mathematics, Applied
Yan-Long Fang, Daniel Lesnic, Moataz Alosaimi
Summary: Understanding the properties of biological tissues is crucial for monitoring and preventing abnormalities that can affect organ functioning. This paper investigates the uniqueness and stable reconstruction of the space-dependent perfusion coefficient in the thermal-wave hyperbolic model of bio-heat transfer using Carleman estimates. It also considers Robin boundary conditions and provides stronger stability estimates for coefficient identification problems.
Article
Mathematics, Applied
Vu Huu Nhu
Summary: This paper discusses the Levenberg-Marquardt method for solving ill-posed inverse problems in Banach spaces with general regularization terms. The findings show that the method performs well in dealing with specific problems.
Article
Telecommunications
Zaid Albataineh
Summary: Massive MIMO wireless systems play a crucial role in 5G communication systems, offering advantages in range, spectral efficiency, and coverage but also encountering challenges in increased computational complexity. The use of multiple search direction conjugate gradient method is proposed to reduce data detection complexity and has been shown to outperform existing methods in terms of computational complexity for large-scale MIMO systems.
WIRELESS PERSONAL COMMUNICATIONS
(2021)
Article
Engineering, Mechanical
Shashi Kant Mishra, Mohammad Esmael Samei, Suvra Kanti Chakraborty, Bhagwat Ram
Summary: This paper proposes a modified q-DY conjugate gradient algorithm based on q-gradient for solving unconstrained optimization problems, demonstrating strong global convergence under standard Wolfe conditions. Numerical results show the efficiency of the proposed algorithm.
NONLINEAR DYNAMICS
(2021)
Article
Mathematics, Applied
Junaid Ahmad, Kifayat Ullah, Reny George
Summary: This paper introduces numerical algorithms for solving nonlinear problems in Hilbert and Banach spaces, including functional equations, split feasibility problems, and variational inequality problems. The algorithms are based on the Thakur-Thakur-Postolache iterative algorithm and the class of mean nonexpansive mappings. Convergence results are provided in the Banach space setting, and a numerical example demonstrates the faster convergence of the TTP algorithm compared to other iterative algorithms. Additionally, new TTP type projection iterative algorithms are proposed for solving SFPs and VIPs in the Hilbert space setting, offering effective numerical methods for finding approximate solutions to nonlinear problems.
Article
Engineering, Electrical & Electronic
Yuan Fang, Kazem Bakian-Dogaheh, Mahta Moghaddam
Summary: A new multifrequency inverse algorithm is proposed for 3-D quantitative microwave imaging, which is capable of reconstructing the 3-D dielectric relaxation model using measured data at multiple transmit frequencies. The algorithm performance is evaluated through experiments and demonstrates accuracy in image reconstruction with no a priori information.
IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION
(2023)
Article
Mathematics, Applied
Brigida Bonino, Claudio Estatico, Marta Lazzaretti
Summary: This article studies one-step iterative algorithms for solving ill-posed inverse problems in the framework of variable exponent Lebesgue spaces L-p(.). It explores gradient descent iteration schemes in Banach spaces and establishes a deep connection between regularization iterative schemes and convex optimization. Furthermore, it applies the Landweber and Conjugate Gradient methods to deblurring imaging problems in L-p(.) space and proposes an effective strategy for selecting the point-wise variable exponent function p(.). Numerical tests demonstrate the advantages of considering variable exponent Lebesgue spaces over the standard L-2 Hilbert and constant exponent Lebesgue spaces in terms of reconstruction quality and convergence speed.
NUMERICAL ALGORITHMS
(2023)
Article
Chemistry, Analytical
Valentina Schenone, Claudio Estatico, Gian Luigi Gragnani, Matteo Pastorino, Andrea Randazzo, Alessandro Fedeli
Summary: A microwave characterization technique for inspecting subsurface scenarios is proposed and numerically assessed in this paper. The method combines finite element electromagnetic modeling with an inversion procedure in Lebesgue spaces with variable exponents, allowing for accurate description of the measurement system and subsurface scenario, as well as improved results in the inverse scattering problem. Numerical simulations of two-layered environments with planar and non-planar air-soil interfaces demonstrate the method's effectiveness in detecting buried objects under different operative conditions.
Article
Engineering, Electrical & Electronic
Igor Bisio, Chiara Garibotto, Fabio Lavagetto, Andrea Sciarrone
Summary: In recent years, significant importance and research efforts have been devoted to ensuring safety and security in the vehicular framework. Maritime surveillance is crucial in this regard as situation awareness is fundamental for guaranteeing safety conditions at sea. This study proposes a novel maritime surveillance system based on Video Content Analysis, leveraging a remote Machine Learning algorithm for automatic vessel detection. Experimental tests are conducted to analyze the impact of packet loss, compression rate, and transport protocol on ship-to-ground communication via satellite, as well as their combined effects on video quality, transmission time, and vessel detection accuracy.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2023)
Article
Computer Science, Information Systems
Igor Bisio, Chiara Garibotto, Halar Haleem, Fabio Lavagetto, Andrea Sciarrone
Summary: Drones have been widely used for road traffic monitoring, and this study focuses on the challenging task of monitoring the Region of Interest (RoI) in drone imagery. Two tasks are addressed: predicting the RoI using deep learning approaches and performing vehicle detection within the RoI. A custom aerial dataset and a Massachusetts roads dataset are utilized for evaluation, and the best performing models are combined to improve results. Experimental tests demonstrate the promising potential of this framework for drone-based road traffic monitoring.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Engineering, Manufacturing
Francesco Sillani, Samuel Poretti, Tommaso Pagani, Fatlind Hajdaj, Manfred Schmid, Andrea Randazzo, Matteo Pastorino, Konrad Wegener
Summary: Temperature monitoring during the cooldown in powder bed fusion of polymers is crucial for quality assurance and is an innovative approach in the field. This technology utilizes microwave tomography to assess the part cooldown history in an industry-grade EOS P110 machine, enabling relative and absolute temperature readings even for parts surrounded by powder. Experimental measurements revealed different cooldown rates depending on the polymer state: 0.56 degrees C min-1 in the liquid phase, 1.2 degrees C min-1 during supercooling, and 0.83 degrees C min-1 in the solid phase. This groundbreaking technology can contribute to the development of smarter PBF machines that can effectively control the entire PBF process, including the often overlooked cooldown phase which significantly impacts the final mechanical performance of parts.
ADDITIVE MANUFACTURING
(2023)
Article
Engineering, Electrical & Electronic
Riccardo Aramini, Massimo Brignone, Daniele Mestriner, Matteo Pastorino, Renato Procopio, Farhad Rachidi, Andrea Randazzo, Marcos Rubinstein
Summary: In this paper, we use numerical simulations to demonstrate that electromagnetic field data from the radiation of a return-stroke lightning discharge can be used to reconstruct the lightning channel's attenuation function when measured over a short-duration time-window. We discretize the current-field equations derived in the first part of this work and formulate an inversion algorithm that computes the attenuation function. Numerical tests show that satisfactory reconstructions can be obtained even for short-duration time-windows.
ELECTRIC POWER SYSTEMS RESEARCH
(2023)
Article
Engineering, Electrical & Electronic
Riccardo Aramini, Massimo Brignone, Daniele Mestriner, Matteo Pastorino, Renato Procopio, Farhad Rachidi, Andrea Randazzo, Marcos Rubinstein
Summary: In this study, a method for reconstructing the attenuation function of a return-stroke current was presented using simultaneous measurements of the channel-base current and radiated electromagnetic fields. However, the assumption that sensors can record the entire time-evolution of the lightning return-stroke is too restrictive for most real data. Therefore, this study aims to remove this restriction and develop a theoretical investigation.
ELECTRIC POWER SYSTEMS RESEARCH
(2023)
Proceedings Paper
Engineering, Electrical & Electronic
Valentina Schenone, Alessandro Fedeli, Claudio Estatico, Igor Bisio, Fabio Lavagetto, Andrea Sciarrone, Matteo Pastorino, Andrea Randazzo
Summary: Stroke diagnostics can benefit from early detection and continuous monitoring using electromagnetic imaging systems. However, retrieving quantitative microwave images of the dielectric properties of the head tissues presents several challenging points. In this paper, a simultaneous processing of multifrequency data is proposed, and initial results from a numerical assessment with a detailed three-dimensional phantom of the human head are presented and discussed, focusing on a simulated hemorrhagic event.
2023 17TH EUROPEAN CONFERENCE ON ANTENNAS AND PROPAGATION, EUCAP
(2023)
Article
Engineering, Electrical & Electronic
Matteo Bruno Lodi, Nicola Curreli, Chiara Dachena, Alessandro Fedeli, Rosa Scapaticci, Andrea Randazzo, Matteo Pastorino, Alessandro Fanti
Summary: Magnetic biomaterials are investigated as theranostic platforms for biomedical applications, particularly for performing local hyperthermia treatment after bone cancer resection. This study explores the feasibility of using a microwave monitoring system to track the treatment process. Numerical simulations show that using magnetic scaffolds for interstitial hyperthermia results in significant variations in the transmission coefficient, which can be correlated with the average tumor temperature.
IEEE JOURNAL OF ELECTROMAGNETICS RF AND MICROWAVES IN MEDICINE AND BIOLOGY
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Valentina Schenone, Alessandro Fedeli, Claudio Estatico, Matteo Pastorino, Andrea Randazzo
Summary: Antenna arrays play a crucial role in various applications such as radar, mobile and satellite communication systems, and electromagnetic imaging. It is important to identify defective components in antenna arrays, as it allows for fixing the faulty elements instead of replacing the entire antenna.
ADVANCES IN SYSTEM-INTEGRATED INTELLIGENCE, SYSINT 2022
(2023)
Proceedings Paper
Engineering, Electrical & Electronic
Alessandro Fedeli, Valentina Schenone, Matteo Pastorino, Andrea Randazzo
Summary: This work introduces a neural network based on long short-term memory cells for dealing with the inverse problem of electromagnetic imaging of dielectric targets at microwave frequencies. The proposed network performs a preliminary processing of the scattered field and is validated in a simulated environment.
2022 16TH EUROPEAN CONFERENCE ON ANTENNAS AND PROPAGATION (EUCAP)
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Valentina Schenone, Alessandro Fedeli, Matteo Pastorino, Andrea Randazzo, Claudio Estatico
Summary: This paper presents a hybrid electromagnetic imaging method for subsurface prospection, consisting of a time-domain qualitative reconstruction algorithm and a multi-frequency quantitative inverse-scattering technique. The scattered-field data estimated by adaptive filtering are used in the first phase to generate an initial image, and then a multifrequency inexact-Newton inversion approach is applied in the second phase for qualitative-quantitative combination strategies. Numerical simulations were conducted to validate the proposed method and an analysis on the effects of the combination strategies was presented.
2022 IEEE INTERNATIONAL CONFERENCE ON IMAGING SYSTEMS AND TECHNIQUES (IST 2022)
(2022)
Article
Engineering, Electrical & Electronic
V. A. L. E. N. T. I. N. A. SCHENONE, A. L. E. S. S. A. N. D. R. O. FEDELI, C. L. A. U. D. I. O. ESTATICO, M. A. T. T. E. O. PASTORINO, A. N. D. R. E. A. RANDAZZO
Summary: This paper presents a novel approach for diagnosing antenna array failures by solving the inverse problem using a non-Hilbertian Lebesgue-space L-p technique. The solution enables the retrieval of faulty feed excitations distribution from far-field measurements. Numerical validations on planar arrays with different failure rates and distributions demonstrate the method's effectiveness in detecting damaged regions.
IEEE OPEN JOURNAL OF ANTENNAS AND PROPAGATION
(2022)
