Ahmed Fawzy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Ahmed Fawzy has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Mohamed Fathy Abo Sree
Department of Electronics and Communications Engineering, Arab Academy for Science, Technology and Maritime Transport, Cairo 11799, Egypt
2 shared publications Ahmed Magdy
Suez Canal University
1 shared publication- Ahmed Magdy
Suez Canal University
1 shared publication - Amir El-Azazy
Electrical Engineering Department, Faculty of Engineering at Shoubra, Benha University, Cairo 11629, Egypt
1 shared publication - Amr Hesham
Department of Electronics and Communication, College of Engineering, AAST, Egypt
1 shared publication Ashraf Kamal Eessaa
Nanotechnology Central Lab Electronics Research Institute El Tahrir St. Giza 12622 Egypt
1 shared publication
Latest Publications
- Open Access
- 6 November 2026
- Published by MDPI AG in Future Internet
- Vol. 18 (11) , 611
Structured Compression and Recovery Fine-Tuning of Egyptian Arabic Hybrid RNN-T/CTC ASR for Offline Raspberry Pi Deployment
Offline Egyptian Arabic automatic speech recognition (ASR) requires balancing recognition accuracy, model size, and edge device performance. This study investigates structured feed-forward network (FFN) compression and recovery fine-tuning of a 114.6M-parameter FastConformer Hybrid RNN-T/CTC model, followed by offline Raspberry Pi 4 deployment. Reducing FFN width from 2048 to 1792 produced a 105.7M-parameter model, a 7.78% reduction. Structured compression initially degraded recognition accuracy, while subsequent recovery fine-tuning restored performance without restoring the removed FFN channels. On a frozen 10,361-utterance test set, the compressed-and-recovered model achieved 25.01% word error rate (WER) and 11.96% character error rate (CER), compared with 25.88% WER and 12.58% CER for the reference model. A dense control given the same recovery training reached 24.81% WER and 11.87% CER; its paired WER difference from the compressed-and-recovered model was −0.20 percentage points (95% confidence interval −0.44 to +0.04; one training run per model). The earlier advantage over the reference therefore does not indicate a compression-specific benefit. On Raspberry Pi 4, the reference and compressed-and-recovered models achieved aggregate real-time factors of 0.470 and 0.466 and mean latencies of 1.185 and 1.176 s, respectively. The parameter reduction yielded no measured Raspberry Pi speed or memory advantage; traffic-command recognition was not evaluated.
- 6 November 2026
- Published by MDPI AG in Future Internet
- Vol. 18 (11) , 611
Open AccessStructured Compression and Recovery Fine-Tuning of Egyptian Arabic Hybrid RNN-T/CTC ASR for Offline Raspberry Pi Deployment
Offline Egyptian Arabic automatic speech recognition (ASR) requires balancing recognition accuracy, model size, and edge device performance. This study investigates structured feed-forward network (FFN) compression and recovery fine-tuning of a 114.6M-parameter FastConformer Hybrid RNN-T/CTC model, followed by offline Raspberry Pi 4 deployment. Reducing FFN width from 2048 to 1792 produced a 105.7M-parameter model, a 7.78% reduction. Structured compression initially degraded recognition accuracy, while subsequent recovery fine-tuning restored performance without restoring the removed FFN channels. On a frozen 10,361-utterance test set, the compressed-and-recovered model achieved 25.01% word error rate (WER) and 11.96% character error rate (CER), compared with 25.88% WER and 12.58% CER for the reference model. A dense control given the same recovery training reached 24.81% WER and 11.87% CER; its paired WER difference from the compressed-and-recovered model was −0.20 percentage points (95% confidence interval −0.44 to +0.04; one training run per model). The earlier advantage over the reference therefore does not indicate a compression-specific benefit. On Raspberry Pi 4, the reference and compressed-and-recovered models achieved aggregate real-time factors of 0.470 and 0.466 and mean latencies of 1.185 and 1.176 s, respectively. The parameter reduction yielded no measured Raspberry Pi speed or memory advantage; traffic-command recognition was not evaluated.
- Open Access
Accelerating Deep Learning-Based Morphological Biometric Recognition with Field-Programmable Gate Arrays
Convolutional neural networks (CNNs) are increasingly recognized as an important and potent artificial intelligence approach, widely employed in many computer vision applications, such as facial recognition. Their importance resides in their capacity to acquire hierarchical features, which is essential for recognizing complex patterns. Nevertheless, the intricate architectural design of CNNs leads to significant computing requirements. To tackle these issues, it is essential to construct a system based on field-programmable gate arrays (FPGAs) to speed up CNNs. FPGAs provide fast development capabilities, energy efficiency, decreased latency, and advanced reconfigurability. A facial recognition solution by leveraging deep learning and subsequently deploying it on an FPGA platform is suggested. The system detects whether a person has the necessary authorization to enter/access a place. The FPGA is responsible for processing this system with utmost security and without any internet connectivity. Various facial recognition networks are accomplished, including AlexNet, ResNet, and VGG-16 networks. The findings of the proposed method prove that the GoogLeNet network is the best fit due to its lower computational resource requirements, speed, and accuracy. The system was deployed on three hardware kits to appraise the performance of different programming approaches in terms of accuracy, latency, cost, and power consumption. The software programming on the Raspberry Pi-3B kit had a recognition accuracy of around 70–75% and relied on a stable internet connection for processing. This dependency on internet connectivity increases bandwidth consumption and fails to meet the required security criteria, contrary to ZYBO-Z7 board hardware programming. Nevertheless, the hardware/software co-design on the PYNQ-Z2 board achieved an accuracy rate of 85% to 87%. It operates independently of an internet connection, making it a standalone system and saving costs.
Open AccessAccelerating Deep Learning-Based Morphological Biometric Recognition with Field-Programmable Gate Arrays
Convolutional neural networks (CNNs) are increasingly recognized as an important and potent artificial intelligence approach, widely employed in many computer vision applications, such as facial recognition. Their importance resides in their capacity to acquire hierarchical features, which is essential for recognizing complex patterns. Nevertheless, the intricate architectural design of CNNs leads to significant computing requirements. To tackle these issues, it is essential to construct a system based on field-programmable gate arrays (FPGAs) to speed up CNNs. FPGAs provide fast development capabilities, energy efficiency, decreased latency, and advanced reconfigurability. A facial recognition solution by leveraging deep learning and subsequently deploying it on an FPGA platform is suggested. The system detects whether a person has the necessary authorization to enter/access a place. The FPGA is responsible for processing this system with utmost security and without any internet connectivity. Various facial recognition networks are accomplished, including AlexNet, ResNet, and VGG-16 networks. The findings of the proposed method prove that the GoogLeNet network is the best fit due to its lower computational resource requirements, speed, and accuracy. The system was deployed on three hardware kits to appraise the performance of different programming approaches in terms of accuracy, latency, cost, and power consumption. The software programming on the Raspberry Pi-3B kit had a recognition accuracy of around 70–75% and relied on a stable internet connection for processing. This dependency on internet connectivity increases bandwidth consumption and fails to meet the required security criteria, contrary to ZYBO-Z7 board hardware programming. Nevertheless, the hardware/software co-design on the PYNQ-Z2 board achieved an accuracy rate of 85% to 87%. It operates independently of an internet connection, making it a standalone system and saving costs.
- Open Access
- 17 May 2024
- Published by Akademia Baru Publishing in Journal of Advanced Research in Applied Sciences and Engineering Technology
- Vol. 45 (1) , 224-238
EC-MAC: Energy-Aware Cooperative MAC Protocol in Wireless Sensor Network
Nowadays, cooperative communication algorithms have been utilized in Wireless Sensor Network (WSN) to enhance the overall network performance. As is well known, a WSN needs to consider a number of important factors, such as the energy effectiveness and the longevity of the sensor nodes. The energy-aware cooperative medium access control (EC-MAC) protocol is a novel protocol proposed in this paper for usage in WSNs. EC-MAC protocol allows the source nodes to use the intermediate nodes as relays that can be used to transmit the source's data to the access point (AP). This paper demonstrates how the suggested relay selection method can be used by the EC-MAC protocol to choose the best relay node. After channel state information (CSI) has been calculated and acquired, the best relay should have the highest residual energy and the quickest transmission time. Then, by establishing suitable cooperative links, data transmission from a source node to an AP can be carried out. The effectiveness of the EC-MAC protocol in terms of system energy efficiency is examined in this study using the MATLAB simulation tool and compares the outcomes with other cooperative protocols like Modified Cooperative Access MAC Protocol (MCA-MAC) and Throughput and Energy aware Cooperative MAC Protocol (TEC-MAC) and the performance of WSNs employing the suggested EC-MAC protocol is examined in this research for both ideal and dynamic channel conditions. EC-MAC protocol achieved energy efficiency improvements of 20%, and 40% respectively, more than MCA-MAC and TEC-MAC protocols. The results indicated that EC-MAC protocol offers a higher level of energy efficiency for the WSN than other cooperative protocols currently in use.
- 17 May 2024
- Published by Akademia Baru Publishing in Journal of Advanced Research in Applied Sciences and Engineering Technology
- Vol. 45 (1) , 224-238
Open AccessEC-MAC: Energy-Aware Cooperative MAC Protocol in Wireless Sensor Network
Nowadays, cooperative communication algorithms have been utilized in Wireless Sensor Network (WSN) to enhance the overall network performance. As is well known, a WSN needs to consider a number of important factors, such as the energy effectiveness and the longevity of the sensor nodes. The energy-aware cooperative medium access control (EC-MAC) protocol is a novel protocol proposed in this paper for usage in WSNs. EC-MAC protocol allows the source nodes to use the intermediate nodes as relays that can be used to transmit the source's data to the access point (AP). This paper demonstrates how the suggested relay selection method can be used by the EC-MAC protocol to choose the best relay node. After channel state information (CSI) has been calculated and acquired, the best relay should have the highest residual energy and the quickest transmission time. Then, by establishing suitable cooperative links, data transmission from a source node to an AP can be carried out. The effectiveness of the EC-MAC protocol in terms of system energy efficiency is examined in this study using the MATLAB simulation tool and compares the outcomes with other cooperative protocols like Modified Cooperative Access MAC Protocol (MCA-MAC) and Throughput and Energy aware Cooperative MAC Protocol (TEC-MAC) and the performance of WSNs employing the suggested EC-MAC protocol is examined in this research for both ideal and dynamic channel conditions. EC-MAC protocol achieved energy efficiency improvements of 20%, and 40% respectively, more than MCA-MAC and TEC-MAC protocols. The results indicated that EC-MAC protocol offers a higher level of energy efficiency for the WSN than other cooperative protocols currently in use.
- Open Access
- 10 April 2024
- Published by Akademia Baru Publishing in Journal of Advanced Research in Applied Sciences and Engineering Technology
- Vol. 43 (1) , 211-226
Energy Harvesting Fractal Antenna for Charging Low Power Devices
Radio frequency energy harvesting (RFEH) is a prospective technique that uses electromagnetic waves which waste in the air to create energy. This cutting-edge technology offers the ability to wirelessly charge battery-free gadgets, making it a potential alternative energy source for use in the future. Fractal antenna systems are a type of antenna that is characterized by their self-similarity and ability to operate over a wide range of frequencies. The proposed design combined two popular fractal shapes in one design. This one design combines the advantages of the other two designs. This research presents a novel multi-band fractal slot antenna design and optimization method for RFEH systems. The fractal antenna designs that were simulated using the CAD design suite, as well as the final design and performance that captured multiple bands are described. The rectifying and matching networks are next evaluated, and the ADS (Advanced Construct System) software is used to construct and simulate each circuit. Additionally, the voltage regulator, energy storage, and application are addressed along with the required voltage to completely function a wireless sensor node, which may be utilized in many applications including wearable technology and intelligent traffic systems (ITS) that control roads to improve environmental quality. Finally, experimental results demonstrated good agreement with simulations that had a high radiation efficiency and gain.
- 10 April 2024
- Published by Akademia Baru Publishing in Journal of Advanced Research in Applied Sciences and Engineering Technology
- Vol. 43 (1) , 211-226
Open AccessEnergy Harvesting Fractal Antenna for Charging Low Power Devices
Radio frequency energy harvesting (RFEH) is a prospective technique that uses electromagnetic waves which waste in the air to create energy. This cutting-edge technology offers the ability to wirelessly charge battery-free gadgets, making it a potential alternative energy source for use in the future. Fractal antenna systems are a type of antenna that is characterized by their self-similarity and ability to operate over a wide range of frequencies. The proposed design combined two popular fractal shapes in one design. This one design combines the advantages of the other two designs. This research presents a novel multi-band fractal slot antenna design and optimization method for RFEH systems. The fractal antenna designs that were simulated using the CAD design suite, as well as the final design and performance that captured multiple bands are described. The rectifying and matching networks are next evaluated, and the ADS (Advanced Construct System) software is used to construct and simulate each circuit. Additionally, the voltage regulator, energy storage, and application are addressed along with the required voltage to completely function a wireless sensor node, which may be utilized in many applications including wearable technology and intelligent traffic systems (ITS) that control roads to improve environmental quality. Finally, experimental results demonstrated good agreement with simulations that had a high radiation efficiency and gain.
- Open Access
- 1 April 2022
- Published by Elsevier in Alexandria Engineering Journal
- Vol. 61 (4) , 3175-3186
A piezoelectric MEMS microphone optimizer platform
- 1 April 2022
- Published by Elsevier in Alexandria Engineering Journal
- Vol. 61 (4) , 3175-3186
Open AccessA piezoelectric MEMS microphone optimizer platform
- 1 January 2022
- Published by Egyptian Knowledge Bank in JES. Journal of Engineering Sciences
- Vol. 50 (1) , 1-8
Membraneless Piezoelectric MEMS speakers based on AlN Thin Film
This paper reports piezoelectric aluminum nitride (AlN) based microelectromechanical systems (MEMS) speakers. We introduce a novel geometry of microspeakers based on AlN to meet the requirements of modern applications such as phones, tablets, laptops, and in ear applications. We introduce the principle, design, and characterization results. The speakers were fabricated on cavity silicon on insulator (SOI) substrate and characterized by using an electroacoustic tester. This paper considers the acoustic performance of the speakers. The results show that the speakers gave us a high sound pressure level (SPL) of more than 78 dB for circle geometry when applying 2 volts on the electrodes. These results are equal for PZT MEMS speakers moreover AlN opens the door to integrate the speakers and the ASIC on the same chip. The size of the different geometries’ speakers isn’t exceeded 3mm × 3mm. These geometries offer a breakthrough in acoustic performance, a frequency response, and low power consumption.
- 1 January 2022
- Published by Egyptian Knowledge Bank in JES. Journal of Engineering Sciences
- Vol. 50 (1) , 1-8
Membraneless Piezoelectric MEMS speakers based on AlN Thin Film
This paper reports piezoelectric aluminum nitride (AlN) based microelectromechanical systems (MEMS) speakers. We introduce a novel geometry of microspeakers based on AlN to meet the requirements of modern applications such as phones, tablets, laptops, and in ear applications. We introduce the principle, design, and characterization results. The speakers were fabricated on cavity silicon on insulator (SOI) substrate and characterized by using an electroacoustic tester. This paper considers the acoustic performance of the speakers. The results show that the speakers gave us a high sound pressure level (SPL) of more than 78 dB for circle geometry when applying 2 volts on the electrodes. These results are equal for PZT MEMS speakers moreover AlN opens the door to integrate the speakers and the ASIC on the same chip. The size of the different geometries’ speakers isn’t exceeded 3mm × 3mm. These geometries offer a breakthrough in acoustic performance, a frequency response, and low power consumption.
- 9 November 2021
- Published by Springer Nature
- p. 338-355
An Intelligent Model for Rapid Diagnosis of Patients with COVID-19 Based on ANFIS
- 9 November 2021
- Published by Springer Nature
- p. 338-355
An Intelligent Model for Rapid Diagnosis of Patients with COVID-19 Based on ANFIS
- Open Access
- 20 January 2021
- Published by Institution of Engineering and Technology (IET) in Micro & Nano Letters
- Vol. 16 (3) , 227-231
Design and analysis of piezoelectric MEMS micro‐speaker based on scandium‐doped AlN thin film
Piezoelectric MEMS microspeakers are promising candidates for miniaturized speakers, such as for in‐ear applications. Using lead‐free scandium doped aluminium nitride (ScAlN) thin film instead of lead zirconium titanate helps these speakers to be environmental friendly, CMOS compatible and potentially more linear. The piezoelectric coefficient of ScAlN is higher than undoped AlN, which leads to improved performance of MEMS actuators including speaker chips. The design of piezoelectric MEMS ScAlN speaker based on four triangular cantilevers is proposed. The design in this work has been verified by finite element method simulation. The displacement and sound pressure level are 39 μm and 56 dB, respectively. This MEMS loudspeaker has small size and simple structure. The model opens the door also to further optimization for higher energy efficiency and higher acoustic performance based on lead‐free ScAlN thin film.
- 20 January 2021
- Published by Institution of Engineering and Technology (IET) in Micro & Nano Letters
- Vol. 16 (3) , 227-231
Open AccessDesign and analysis of piezoelectric MEMS micro‐speaker based on scandium‐doped AlN thin film
Piezoelectric MEMS microspeakers are promising candidates for miniaturized speakers, such as for in‐ear applications. Using lead‐free scandium doped aluminium nitride (ScAlN) thin film instead of lead zirconium titanate helps these speakers to be environmental friendly, CMOS compatible and potentially more linear. The piezoelectric coefficient of ScAlN is higher than undoped AlN, which leads to improved performance of MEMS actuators including speaker chips. The design of piezoelectric MEMS ScAlN speaker based on four triangular cantilevers is proposed. The design in this work has been verified by finite element method simulation. The displacement and sound pressure level are 39 μm and 56 dB, respectively. This MEMS loudspeaker has small size and simple structure. The model opens the door also to further optimization for higher energy efficiency and higher acoustic performance based on lead‐free ScAlN thin film.
- Open Access
- 1 November 2018
- Published by Institution of Engineering and Technology (IET) in Micro & Nano Letters
- Vol. 13 (11) , 1516-1519
Energy gap variation due to Al content in SmFe 1− x Al x O 3 and its application in optics
The optical properties of Al‐doped SmFeO 3 (SmFe 1− x Al x O 3 , for x varying from 0 to 0.15), were monitored by diffuse reflectance spectroscopy (DRS). The diffuse reflectance spectra were used to study the surface properties of the samples. The DRS spectra exhibit three reflection bands at different regions. The product was characterised by X‐ray diffraction to confirm the formation of single‐phase crystals of the perovskite structure. The bandgap, a major factor in determining the optical performance of any material, was found to be tunable with Al content ( x ).
- 1 November 2018
- Published by Institution of Engineering and Technology (IET) in Micro & Nano Letters
- Vol. 13 (11) , 1516-1519
Open AccessEnergy gap variation due to Al content in SmFe 1− x Al x O 3 and its application in optics
The optical properties of Al‐doped SmFeO 3 (SmFe 1− x Al x O 3 , for x varying from 0 to 0.15), were monitored by diffuse reflectance spectroscopy (DRS). The diffuse reflectance spectra were used to study the surface properties of the samples. The DRS spectra exhibit three reflection bands at different regions. The product was characterised by X‐ray diffraction to confirm the formation of single‐phase crystals of the perovskite structure. The bandgap, a major factor in determining the optical performance of any material, was found to be tunable with Al content ( x ).