Fethi Choubani
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Fethi Choubani has a total of 38 co-authors. Here are the top co-authors by number of shared publications:
- Mohamed Hayouni
InnovCom Laboratory, Higher School of Communications (SUP’COM), University of Carthage, Tunis 1054, Tunisia
5 shared publications - Emna Jebabli
Université de Carthage
4 shared publications - Sofiane Ben Mbarek
CEMSE, King Abdullah University of Science and Technology (KAUST) , Thuwal
4 shared publications - Aymen Ben Saada
, University of Carthage, Tunisia
2 shared publications - Abdeldjalil Aissa-El-Bey
IMT Atlantique, Lab-STICC, UMR CNRS 6285, Brest, France
1 shared publication - Amani Cherif
Institut d’Electronique et des Technologies du numeRique (IETR), Université de Rennes, CNRS, IETR-UMR 6164, F-35000 Rennes, France
1 shared publication
Latest Publications
- Open Access
HydroNeuro: A Data-Efficient IoT Sensing and Edge-AI Framework for Real-Time Hydraulic Anomaly Detection
Reliable monitoring of hydraulic networks is essential for efficient and sustainable water management in agriculture. To address the growing need for intelligent, low-latency anomaly detection in such systems, we propose HydroNeuro, a domain-aware embedded framework that integrates hydraulic domain knowledge with data-driven neural inference for the real-time detection of leaks and obstructions. Rather than embedding physical equations directly into the learning objective, we leverage established hydraulic principles, including Bernoulli’s equation and the Darcy–Weisbach formulation, to structure the experimental design, interpret pressure–flow relationships, and ensure physical consistency of the learned representations. These principles confirm that pressure deviations induced by leaks or obstructions are causally explainable and measurable. We employ a fractional factorial design (FFD) to optimize valve activation combinations and sensor configurations during dataset acquisition, thereby reducing redundant experiments, water circulation, and energy consumption while limiting mechanical stress on system components. We deploy a lightweight neural network on an ESP32 microcontroller using TensorFlow Lite for Microcontrollers to enable energy-efficient, low-latency edge inference under severe hardware constraints. Our experimental validation on a laboratory-scale hydraulic testbed demonstrates anomaly detection accuracy exceeding 96%, with strong robustness under sensor noise and hydraulic perturbations. Compared to a multiple linear regression baseline, the proposed neural model reduces the prediction error from an RMSE of 0.58 to 0.12. By coupling physically consistent experimental modeling with embedded neural inference, HydroNeuro provides a scalable and practically deployable solution for autonomous hydraulic monitoring in precision irrigation and smart water distribution systems.
Open AccessHydroNeuro: A Data-Efficient IoT Sensing and Edge-AI Framework for Real-Time Hydraulic Anomaly Detection
Reliable monitoring of hydraulic networks is essential for efficient and sustainable water management in agriculture. To address the growing need for intelligent, low-latency anomaly detection in such systems, we propose HydroNeuro, a domain-aware embedded framework that integrates hydraulic domain knowledge with data-driven neural inference for the real-time detection of leaks and obstructions. Rather than embedding physical equations directly into the learning objective, we leverage established hydraulic principles, including Bernoulli’s equation and the Darcy–Weisbach formulation, to structure the experimental design, interpret pressure–flow relationships, and ensure physical consistency of the learned representations. These principles confirm that pressure deviations induced by leaks or obstructions are causally explainable and measurable. We employ a fractional factorial design (FFD) to optimize valve activation combinations and sensor configurations during dataset acquisition, thereby reducing redundant experiments, water circulation, and energy consumption while limiting mechanical stress on system components. We deploy a lightweight neural network on an ESP32 microcontroller using TensorFlow Lite for Microcontrollers to enable energy-efficient, low-latency edge inference under severe hardware constraints. Our experimental validation on a laboratory-scale hydraulic testbed demonstrates anomaly detection accuracy exceeding 96%, with strong robustness under sensor noise and hydraulic perturbations. Compared to a multiple linear regression baseline, the proposed neural model reduces the prediction error from an RMSE of 0.58 to 0.12. By coupling physically consistent experimental modeling with embedded neural inference, HydroNeuro provides a scalable and practically deployable solution for autonomous hydraulic monitoring in precision irrigation and smart water distribution systems.
- 24 June 2025
- Published by Springer Nature in Journal of Infrared, Millimeter, and Terahertz Waves
- Vol. 46 (7) , 1-17
A High Gain Switched Beamforming Antenna Array with Hemispherical Lens for Ka-Band 5G Applications
- 24 June 2025
- Published by Springer Nature in Journal of Infrared, Millimeter, and Terahertz Waves
- Vol. 46 (7) , 1-17
A High Gain Switched Beamforming Antenna Array with Hemispherical Lens for Ka-Band 5G Applications
- Open Access
- 1 January 2024
- Published by Wiley in International Journal of RF and Microwave Computer-Aided Engineering
- Vol. 2024 (1)
Mutual Coupling Reduction in a Switched Beamforming Multilayer Antenna Array System
In this paper, we present a millimeter‐wave beamforming antenna array for the fifth generation (5G) of mobile networks and beyond. One of our priority objectives was to minimize the antenna array’s total physical size, by paying attention to avoid radiation performance degradation, such as bandwidth and efficiency. Another priority objective was to avoid mutual coupling (MC) between radiation elements and enhance the gain of the antenna to meet the requirements and challenges of the 5G. In these prospects, we have designed a 4 × 3 Chebyshev modified rectangular antenna array, engraved on the dielectric substrate of the top layer. The antenna array is powered across a 4 × 4 Butler matrix, engraved on the dielectric substrate of the bottom layer. The prototype overall size is 35.76 × 45.56 × 1.589 mm3. Hence, four switched main lobes pointed according to the angles ‐14°, +26°, ‐23° and +12°. The corresponded maximum gains at 28 GHz are 11.1, 10.5, 11, and 11.5 dB, respectively. Moreover, a reduction of the MC is ensured by using a CSRR metamaterial carved into the ground plane (GP) of the upper dielectric substrate layer. Experimental and simulation results concord well and show a multiband behavior at 26 GHz and in Ka‐band. The impedance bandwidths are 1.5%, 2.7%, and 2.43% at 26, 28, and 40 GHz, respectively. A low MC, between the antenna array’s elements, is also achieved, and the levels of the transmission coefficients vary between ‐11 dB and ‐68 dB.
- 1 January 2024
- Published by Wiley in International Journal of RF and Microwave Computer-Aided Engineering
- Vol. 2024 (1)
Open AccessMutual Coupling Reduction in a Switched Beamforming Multilayer Antenna Array System
In this paper, we present a millimeter‐wave beamforming antenna array for the fifth generation (5G) of mobile networks and beyond. One of our priority objectives was to minimize the antenna array’s total physical size, by paying attention to avoid radiation performance degradation, such as bandwidth and efficiency. Another priority objective was to avoid mutual coupling (MC) between radiation elements and enhance the gain of the antenna to meet the requirements and challenges of the 5G. In these prospects, we have designed a 4 × 3 Chebyshev modified rectangular antenna array, engraved on the dielectric substrate of the top layer. The antenna array is powered across a 4 × 4 Butler matrix, engraved on the dielectric substrate of the bottom layer. The prototype overall size is 35.76 × 45.56 × 1.589 mm3. Hence, four switched main lobes pointed according to the angles ‐14°, +26°, ‐23° and +12°. The corresponded maximum gains at 28 GHz are 11.1, 10.5, 11, and 11.5 dB, respectively. Moreover, a reduction of the MC is ensured by using a CSRR metamaterial carved into the ground plane (GP) of the upper dielectric substrate layer. Experimental and simulation results concord well and show a multiband behavior at 26 GHz and in Ka‐band. The impedance bandwidths are 1.5%, 2.7%, and 2.43% at 26, 28, and 40 GHz, respectively. A low MC, between the antenna array’s elements, is also achieved, and the levels of the transmission coefficients vary between ‐11 dB and ‐68 dB.
- 14 December 2023
- Published by SAGE Publications in International Journal of Applied Electromagnetics and Mechanics
- Vol. 73 (4) , 321-338
Modeling of the interaction between human body and electromagnetic waves near resonance using machine learning
Human-Like digital models have been around for quite some time. They significantly contributed to the increase of the accuracy of the whole-body-average specific absorption rate estimations. However, the anatomical and morphological diversity between human beings has not yet been embraced by the actual anthropomorphic models for several reasons such as financial costs, excessive exposure of volunteers to electromagnetic waves, and the required number of technical experts needed to build one voxelized model. Recently, machine learning has been used to reduce the complexity of certain tasks. Yet, at least, having an anthropomorphic model per nation is still far away to achieve. To reduce the building cost of new human-like models, we build on the success of anthropomorphic models and machine learning to derive mathematical equations that make it possible to predict the Whole-body-average SAR from low frequencies up to twice the resonance frequency without any cost and excessive electromagnetic exposure of new volunteers. The completely new machine learning based equations are applicable for any age, ethnic group, and for both genders. They depend only on the human body’s morphological (height and weight) and anatomical parameters (tissue weights). In this work, we first address the whole-body-average SAR peak and we present a set of two estimators. In second, we show that the resonance frequency is not only a function of the height of the human body, to end up with a third estimation for the resonance frequency. These completely new estimators are finally combined into a novel equation that links the whole-body-average SAR to the frequency. It shows the accurate prediction for low frequencies (10 MHz) up to twice the resonance frequency. The derived estimators for the maximum WBASAR and the resonance frequencies showed better results for low frequency exposure.
- 14 December 2023
- Published by SAGE Publications in International Journal of Applied Electromagnetics and Mechanics
- Vol. 73 (4) , 321-338
Modeling of the interaction between human body and electromagnetic waves near resonance using machine learning
Human-Like digital models have been around for quite some time. They significantly contributed to the increase of the accuracy of the whole-body-average specific absorption rate estimations. However, the anatomical and morphological diversity between human beings has not yet been embraced by the actual anthropomorphic models for several reasons such as financial costs, excessive exposure of volunteers to electromagnetic waves, and the required number of technical experts needed to build one voxelized model. Recently, machine learning has been used to reduce the complexity of certain tasks. Yet, at least, having an anthropomorphic model per nation is still far away to achieve. To reduce the building cost of new human-like models, we build on the success of anthropomorphic models and machine learning to derive mathematical equations that make it possible to predict the Whole-body-average SAR from low frequencies up to twice the resonance frequency without any cost and excessive electromagnetic exposure of new volunteers. The completely new machine learning based equations are applicable for any age, ethnic group, and for both genders. They depend only on the human body’s morphological (height and weight) and anatomical parameters (tissue weights). In this work, we first address the whole-body-average SAR peak and we present a set of two estimators. In second, we show that the resonance frequency is not only a function of the height of the human body, to end up with a third estimation for the resonance frequency. These completely new estimators are finally combined into a novel equation that links the whole-body-average SAR to the frequency. It shows the accurate prediction for low frequencies (10 MHz) up to twice the resonance frequency. The derived estimators for the maximum WBASAR and the resonance frequencies showed better results for low frequency exposure.
- 8 November 2023
- Published by Springer Nature in Applied Physics A
- Vol. 129 (12) , 1-10
A millimeter-wave beamforming antenna array with mutual coupling reduction using metamaterial surface
- 8 November 2023
- Published by Springer Nature in Applied Physics A
- Vol. 129 (12) , 1-10
A millimeter-wave beamforming antenna array with mutual coupling reduction using metamaterial surface
- Open Access
- 1 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. c1-c5
ComNet 2023 Cover Page
- 1 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. c1-c5
Open AccessComNet 2023 Cover Page
- Open Access
- 5 July 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (13) , 7878
Radiation Beam Width and Beam Direction Electronic Control of Transparent and Compact Vivaldi Antennas
In this paper, we present a study on a broadband transparent tapered slot antenna. In general, the objective of achieving optical transparency is to enable antennas to seamlessly integrate into windows, offering an aesthetically pleasing and inconspicuous appearance. The aim of our research is to develop antennas that possess the ability to adjust horizontal plane beams across a wide frequency range, from 24 to 28 GHz, for 5G applications. This structure combines three antennas into a single unit, providing an advantage in terms of saving space. Furthermore, this structure offers the possibility of choosing between using a single antenna to obtain a directional beam in the −90°, 0°, or +90° directions (depending on the activated antenna) corresponding to three states, or the combination between two states to obtain another three additional states. The combination of the three states also allows for the acquisition of another state. At this point, the total number of states is 23 − 1. Only three PIN diodes are employed to switch between all states. Additionally, by adjusting the bias values of the PIN diodes, which function as variable resistors, the antenna beamwidth can be adjusted in order to achieve a coverage of 300°, offering more radiation pattern reconfigurability. The proposed method offers several advantages, including simplicity and feasibility in controlling the beamwidth and the beam direction electronically. This structure can be easily integrated into the development of fifth-generation communication systems.
- 5 July 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (13) , 7878
Open AccessRadiation Beam Width and Beam Direction Electronic Control of Transparent and Compact Vivaldi Antennas
In this paper, we present a study on a broadband transparent tapered slot antenna. In general, the objective of achieving optical transparency is to enable antennas to seamlessly integrate into windows, offering an aesthetically pleasing and inconspicuous appearance. The aim of our research is to develop antennas that possess the ability to adjust horizontal plane beams across a wide frequency range, from 24 to 28 GHz, for 5G applications. This structure combines three antennas into a single unit, providing an advantage in terms of saving space. Furthermore, this structure offers the possibility of choosing between using a single antenna to obtain a directional beam in the −90°, 0°, or +90° directions (depending on the activated antenna) corresponding to three states, or the combination between two states to obtain another three additional states. The combination of the three states also allows for the acquisition of another state. At this point, the total number of states is 23 − 1. Only three PIN diodes are employed to switch between all states. Additionally, by adjusting the bias values of the PIN diodes, which function as variable resistors, the antenna beamwidth can be adjusted in order to achieve a coverage of 300°, offering more radiation pattern reconfigurability. The proposed method offers several advantages, including simplicity and feasibility in controlling the beamwidth and the beam direction electronically. This structure can be easily integrated into the development of fifth-generation communication systems.
- Open Access
- 7 March 2023
- Published by MDPI AG in Electronics
- Vol. 12 (6) , 1282
Mathematical Channel Modeling of Electromagnetic Waves in Biological Tissues for Wireless Body Communication
In the wireless body area network (WBAN), radio propagations from devices that communicate with the human body are very complex and distinctive compared to other environments. As we know, the human body is a lossy channel that significantly attenuates the propagation of electromagnetic waves (EMW). Therefore, channel models are critical in evaluating the communication link. One of the most predominant models is the path loss channel model, which is used to cover a wide range of communication channels and frequency bands in WBAN. This paper investigates the EMW in a human model irradiated by an incident electromagnetic plane wave. A planar multilayer structure is used for modeling human tissue. Moreover, the steady-state electromagnetic distribution is calculated by solving the differential and integral equations (DIE) by using the method of moments (MoM). The obtained results demonstrate the great use of performing a theoretical analysis for path loss (PL) and power loss density (PLD) estimation. The magnitude of the electric field inside muscle tissue at various depths, and with the most important frequencies in medical applications, is evaluated. This investigation provides evidence that the penetration of EMW in biological tissue strongly depends on the frequency and thickness of the tissue involved. Thus, for different examined conditions, an excellent agreement between recent results that were obtained by an analytical method, finite element (FEM), and the proposed MoM method is verified to be valid in this investigation, and it is found that the distribution of the field, PL, and PLD for different communication scenarios is very promising to determine the quality of communication for WBAN technology.
- 7 March 2023
- Published by MDPI AG in Electronics
- Vol. 12 (6) , 1282
Open AccessMathematical Channel Modeling of Electromagnetic Waves in Biological Tissues for Wireless Body Communication
In the wireless body area network (WBAN), radio propagations from devices that communicate with the human body are very complex and distinctive compared to other environments. As we know, the human body is a lossy channel that significantly attenuates the propagation of electromagnetic waves (EMW). Therefore, channel models are critical in evaluating the communication link. One of the most predominant models is the path loss channel model, which is used to cover a wide range of communication channels and frequency bands in WBAN. This paper investigates the EMW in a human model irradiated by an incident electromagnetic plane wave. A planar multilayer structure is used for modeling human tissue. Moreover, the steady-state electromagnetic distribution is calculated by solving the differential and integral equations (DIE) by using the method of moments (MoM). The obtained results demonstrate the great use of performing a theoretical analysis for path loss (PL) and power loss density (PLD) estimation. The magnitude of the electric field inside muscle tissue at various depths, and with the most important frequencies in medical applications, is evaluated. This investigation provides evidence that the penetration of EMW in biological tissue strongly depends on the frequency and thickness of the tissue involved. Thus, for different examined conditions, an excellent agreement between recent results that were obtained by an analytical method, finite element (FEM), and the proposed MoM method is verified to be valid in this investigation, and it is found that the distribution of the field, PL, and PLD for different communication scenarios is very promising to determine the quality of communication for WBAN technology.