Mahmoud M. Elmessalawy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mahmoud M. Elmessalawy has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
- Mohamed Elhattab
Concordia University, Montreal, QC, Canada
8 shared publications - Mostafa M. Abdelhakam
Department of Electronics and Communications Engineering, Faculty of Engineering, Helwan University, Cairo, Egypt
8 shared publications - Ahmed Abd El-Haleem
Department of Electronics and Communications, Faculty of Engineering, Helwan University, Cairo, Egypt
7 shared publications - Haitham H. Esmat
Department of Electrical and Computer Engineering, University of Massachusetts Amherst, Amherst, MA, USA
6 shared publications I. I. Ibrahim
Electronics, Communications and Computers Dept., Helwan University, Egypt, 11792
5 shared publications- Moustafa Mohammed Eissa
Electrical Power and Machines Engineering Department, Helwan University, Cairo, Egypt
4 shared publications
Latest Publications
- 24 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Networking Letters
- Vol. 8, 14-18
DRL-Driven Edge-Aware Utility Optimization for Multi-Slice 6G Networks
Virtual Reality (VR) services delivered over 6G networks demand ultra-low latency and high bandwidth to ensure seamless user experiences. This paper presents an intelligent resource allocation and edge caching framework for 6G O-RAN networks, leveraging Deep Q-Network (DQN) learning for optimizing edge caching and dynamic resource provisioning across multiple network slices within an O-RAN-compliant architecture.By incorporating DRL agents into the network control plane, the proposed system enables proactive and adaptive content distribution as well as real-time computational resource allocation that meets the quality-of-service demands of eMBB, URLLC, and especially the emerging MBRLLC slices essential for VR. Simulation results demonstrate that the DQN-based framework consistently outperforms traditional methods in reducing latency and improving throughput, leading to more reliable and responsive support for immersive VR applications in 6G environments.
- 24 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Networking Letters
- Vol. 8, 14-18
DRL-Driven Edge-Aware Utility Optimization for Multi-Slice 6G Networks
Virtual Reality (VR) services delivered over 6G networks demand ultra-low latency and high bandwidth to ensure seamless user experiences. This paper presents an intelligent resource allocation and edge caching framework for 6G O-RAN networks, leveraging Deep Q-Network (DQN) learning for optimizing edge caching and dynamic resource provisioning across multiple network slices within an O-RAN-compliant architecture.By incorporating DRL agents into the network control plane, the proposed system enables proactive and adaptive content distribution as well as real-time computational resource allocation that meets the quality-of-service demands of eMBB, URLLC, and especially the emerging MBRLLC slices essential for VR. Simulation results demonstrate that the DQN-based framework consistently outperforms traditional methods in reducing latency and improving throughput, leading to more reliable and responsive support for immersive VR applications in 6G environments.
- Open Access
- 10 November 2022
- Published by MDPI AG in Applied Sciences
- Vol. 12 (22) , 11416
An Intelligent Approach for Fair Assessment of Online Laboratory Examinations in Laboratory Learning Systems Based on Student’s Mouse Interaction Behavior
The COVID-19 pandemic has made the world focus on providing effective and fair online learning systems. As a consequence, this paper proposed a new intelligent, fair assessment of online examinations for virtual and remotely controlled laboratory experiments running through Laboratory Learning Systems. The main idea is to provide students with an environment similar to being physically present in a Laboratory while conducting practical experiments and exams and detecting cheating with high accuracy at a minimal cost. Therefore, an intelligent assessment module is designed to detect cheating students by analyzing their mouse dynamics using Artificial Intelligence. The mouse interaction behavior method was chosen because it does not require any additional resources, such as a camera and eye tribe tracker, to detect cheating. Various AI algorithms, such as KNN, SVC, Random Forest, Logistic Regression, XGBoost, and LightGBM have been used to classify student mouse behavior to detect cheating, and many metrics are used to evaluate their performance. Moreover, experiments have been conducted on students answering online laboratory experimentations while cheating and when answering the exams honestly. Experimental results indicate that the LightGBM AI algorithm achieves the best cheat detection results up to an accuracy of 90%, precision of 88%, and degree of separation of 95%.
- 10 November 2022
- Published by MDPI AG in Applied Sciences
- Vol. 12 (22) , 11416
Open AccessAn Intelligent Approach for Fair Assessment of Online Laboratory Examinations in Laboratory Learning Systems Based on Student’s Mouse Interaction Behavior
The COVID-19 pandemic has made the world focus on providing effective and fair online learning systems. As a consequence, this paper proposed a new intelligent, fair assessment of online examinations for virtual and remotely controlled laboratory experiments running through Laboratory Learning Systems. The main idea is to provide students with an environment similar to being physically present in a Laboratory while conducting practical experiments and exams and detecting cheating with high accuracy at a minimal cost. Therefore, an intelligent assessment module is designed to detect cheating students by analyzing their mouse dynamics using Artificial Intelligence. The mouse interaction behavior method was chosen because it does not require any additional resources, such as a camera and eye tribe tracker, to detect cheating. Various AI algorithms, such as KNN, SVC, Random Forest, Logistic Regression, XGBoost, and LightGBM have been used to classify student mouse behavior to detect cheating, and many metrics are used to evaluate their performance. Moreover, experiments have been conducted on students answering online laboratory experimentations while cheating and when answering the exams honestly. Experimental results indicate that the LightGBM AI algorithm achieves the best cheat detection results up to an accuracy of 90%, precision of 88%, and degree of separation of 95%.
- 22 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 292-297
Geofencing-based Congestion Control in Workplaces Environment using Sequential Pattern Mining
To control congestion in the workplace environment especially in crises like the COVID-19 pandemic, this requires careful control of highly crowded workplace locations. Therefore, innovative technologies, such as geofencing and sequential pattern mining can be used to estimate people movement pattern and combat the spread of COVID-19. In this paper, the workplace area is divided into a set of geofences by using geofencing technology. Then, the movement profiles of each user are estimated to control the possible congestion in the workplace's enviroment. To accomplish this, the user's historical geofence transitions are used to anticipate the next time the user will leave the current geofence. The Sequential Pattern Discovery using Equivalence classes (CM-SPADE), Succinct BWT-based Sequence prediction model (SuBSeq) and Compact Prediction Tree + (CPT+) algorithms are adopted to predict the user's next geofence. In the CM-SPADE algorithm, a vertical database is obtained from the available database and the frequent sequence is found based on relative support, confidence, and lift measures. Meanwhile, in the training phase of the SuBSeq algorithm, Ferragina and Manzini (FM)-index and Burrows-Wheeler Transform string are generated. Then, in the ready-to-predict phase, the next geofence is anticipated. The CPT+ algorithm is based on generating Prediction Tree (PT), Lookup Table (LT), and Inverted Index (IIdx) for the training data. Then, Frequent Subsequence Compression (FSC) and Simple Branches Compression (SBC) are used to reduce the size of the PT. In addition, the Prediction with improved Noise Reduction (PNR) method is utilized to reduce the execution time. The results show remarkable superiority for SuBSeq algorithm over CM-SPADE and CPT+ with the accuracy greater than 90% withh an average of 8 input geofences to predict the next geofence.
- 22 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 292-297
Geofencing-based Congestion Control in Workplaces Environment using Sequential Pattern Mining
To control congestion in the workplace environment especially in crises like the COVID-19 pandemic, this requires careful control of highly crowded workplace locations. Therefore, innovative technologies, such as geofencing and sequential pattern mining can be used to estimate people movement pattern and combat the spread of COVID-19. In this paper, the workplace area is divided into a set of geofences by using geofencing technology. Then, the movement profiles of each user are estimated to control the possible congestion in the workplace's enviroment. To accomplish this, the user's historical geofence transitions are used to anticipate the next time the user will leave the current geofence. The Sequential Pattern Discovery using Equivalence classes (CM-SPADE), Succinct BWT-based Sequence prediction model (SuBSeq) and Compact Prediction Tree + (CPT+) algorithms are adopted to predict the user's next geofence. In the CM-SPADE algorithm, a vertical database is obtained from the available database and the frequent sequence is found based on relative support, confidence, and lift measures. Meanwhile, in the training phase of the SuBSeq algorithm, Ferragina and Manzini (FM)-index and Burrows-Wheeler Transform string are generated. Then, in the ready-to-predict phase, the next geofence is anticipated. The CPT+ algorithm is based on generating Prediction Tree (PT), Lookup Table (LT), and Inverted Index (IIdx) for the training data. Then, Frequent Subsequence Compression (FSC) and Simple Branches Compression (SBC) are used to reduce the size of the PT. In addition, the Prediction with improved Noise Reduction (PNR) method is utilized to reduce the execution time. The results show remarkable superiority for SuBSeq algorithm over CM-SPADE and CPT+ with the accuracy greater than 90% withh an average of 8 input geofences to predict the next geofence.
- 15 October 2022
- Published by Springer Nature
- p. 210-221
Efficient Collaborative and Cooperative Laboratory Experimentation System for Online Engineering, Science and Technology Education
- 15 October 2022
- Published by Springer Nature
- p. 210-221
Efficient Collaborative and Cooperative Laboratory Experimentation System for Online Engineering, Science and Technology Education
- Open Access
A Machine Learning Approach for Movement Monitoring in Clustered Workplaces to Control COVID-19 Based on Geofencing and Fusion of Wi-Fi and Magnetic Field Metrics
The ubiquitous existence of COVID-19 has required the management of congested areas such as workplaces. As a result, the use of a variety of inspiring tools to deal with the spread of COVID-19 has been required, including internet of things, artificial intelligence (AI), machine learning (ML), and geofencing technologies. In this work, an efficient approach based on the use of ML and geofencing technology is proposed to monitor and control the density of persons in workplaces during working hours. In particular, the workplace environment is divided into a number of geofences in which each person is associated with a set of geofences that make up their own cluster using a dynamic user-centric clustering scheme. Different metrics are used to generate a unique geofence digital signature (GDS) such as Wi-Fi basic service set identifier, Wi-Fi received signal strength indication, and magnetic field data, which can be collected using the person’s smartphone. Then, these metrics are utilized by different ML techniques to generate the GDS for each indoor geofence and each building geofence as well as to detect whether the person is in their cluster. In addition, a Layered-Architecture Geofence Division method is considered to reduce the processing overhead at the person’s smartphone. Our experimental results demonstrate that the proposed approach can perform well in a real workplace environment. The results show that the system accuracy is about 98.25% in indoor geofences and 76% in building geofences.
Open AccessA Machine Learning Approach for Movement Monitoring in Clustered Workplaces to Control COVID-19 Based on Geofencing and Fusion of Wi-Fi and Magnetic Field Metrics
The ubiquitous existence of COVID-19 has required the management of congested areas such as workplaces. As a result, the use of a variety of inspiring tools to deal with the spread of COVID-19 has been required, including internet of things, artificial intelligence (AI), machine learning (ML), and geofencing technologies. In this work, an efficient approach based on the use of ML and geofencing technology is proposed to monitor and control the density of persons in workplaces during working hours. In particular, the workplace environment is divided into a number of geofences in which each person is associated with a set of geofences that make up their own cluster using a dynamic user-centric clustering scheme. Different metrics are used to generate a unique geofence digital signature (GDS) such as Wi-Fi basic service set identifier, Wi-Fi received signal strength indication, and magnetic field data, which can be collected using the person’s smartphone. Then, these metrics are utilized by different ML techniques to generate the GDS for each indoor geofence and each building geofence as well as to detect whether the person is in their cluster. In addition, a Layered-Architecture Geofence Division method is considered to reduce the processing overhead at the person’s smartphone. Our experimental results demonstrate that the proposed approach can perform well in a real workplace environment. The results show that the system accuracy is about 98.25% in indoor geofences and 76% in building geofences.
- Open Access
- 1 January 2020
- Published by Institution of Engineering and Technology (IET) in IET Communications
- Vol. 14 (1) , 11-20
Energy‐efficient BBU pool virtualisation for C‐RAN with quality of service guarantees
Cloud radio access network (C‐RAN) has been introduced as a promising network paradigm for improving the spectral and energy efficiency of next‐generation mobile systems. In C‐RAN, the computation resources of the centralised baseband units (BBUs) can be virtualised and dynamically shared among cells for energy‐efficient BBU pool utilisation. In this study, a BBU virtualisation scheme is proposed to minimise the total power consumption in the BBU pool subject to constraints on users’ quality of service in terms of real‐time requirements, individual fronthaul capacity and BBU capacity. As the BBU processing time and transmission delay for each user data can be compromised to meet the user's real‐time requirements while minimising the BBU power consumption, a priori user association phase is proposed and formulated as an optimisation problem to maximise the users’ transmission rate, and hence minimising their transmission delay. Then, the BBU processing allocation phase is formulated as a bin‐packing problem to minimise the overall power consumption in the BBU pool. Since this problem is combinatorial, a heuristic algorithm is proposed based on best‐fit‐decreasing algorithm to solve it. Extensive simulations show that the proposed scheme outperforms the comparable ones in terms of power consumption with reduction up to 33%.
- 1 January 2020
- Published by Institution of Engineering and Technology (IET) in IET Communications
- Vol. 14 (1) , 11-20
Open AccessEnergy‐efficient BBU pool virtualisation for C‐RAN with quality of service guarantees
Cloud radio access network (C‐RAN) has been introduced as a promising network paradigm for improving the spectral and energy efficiency of next‐generation mobile systems. In C‐RAN, the computation resources of the centralised baseband units (BBUs) can be virtualised and dynamically shared among cells for energy‐efficient BBU pool utilisation. In this study, a BBU virtualisation scheme is proposed to minimise the total power consumption in the BBU pool subject to constraints on users’ quality of service in terms of real‐time requirements, individual fronthaul capacity and BBU capacity. As the BBU processing time and transmission delay for each user data can be compromised to meet the user's real‐time requirements while minimising the BBU power consumption, a priori user association phase is proposed and formulated as an optimisation problem to maximise the users’ transmission rate, and hence minimising their transmission delay. Then, the BBU processing allocation phase is formulated as a bin‐packing problem to minimise the overall power consumption in the BBU pool. Since this problem is combinatorial, a heuristic algorithm is proposed based on best‐fit‐decreasing algorithm to solve it. Extensive simulations show that the proposed scheme outperforms the comparable ones in terms of power consumption with reduction up to 33%.
- Open Access
- 23 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 158623-158633
User Association With Mode Selection in LWA-Based Multi-RAT HetNet
- 23 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 158623-158633
Open AccessUser Association With Mode Selection in LWA-Based Multi-RAT HetNet
- 1 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 1, 170-173
Experimental OAI-based Testbed for Evaluating the Impact of Different Functional Splits on C-RAN Performance
Cloud Radio Access Network (C-RAN) is an innovative concept introduced to enhance the performance of the Fifth Generation (5G) mobile networks. The main idea of CRAN is to split the functionalities of the ordinary Evolved Node B (eNB) into two main components. The first one is the Remote Radio Head (RRH) that contains the Radio Frequency (RF) functionality in which multiple RRHs will be connected to the other shared component that is called the Baseband Processing Unit (BBU) which performs the other functionalities. The link between these two components will be referred to as the fronthaul interface. This interface has many different technologies in which we are interested in this paper in the Next Generation Fronthaul Interface (NGFI). According to NGFI, various functional splits have different associated tasks to both the BBU and the RRH which impacts on the C-RAN performance. In this paper, OpenAirInterface (OAI) platform is used to implement and conFigure C-RAN testbed. Then, the impact of different NGFI splits on C-RAN performance is evaluated and concluded using the implemented C-RAN testbed in terms of various perfomance metrics such as required fronthaul data rate, percentage of CPU usage, and memory usage.
- 1 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 1, 170-173
Experimental OAI-based Testbed for Evaluating the Impact of Different Functional Splits on C-RAN Performance
Cloud Radio Access Network (C-RAN) is an innovative concept introduced to enhance the performance of the Fifth Generation (5G) mobile networks. The main idea of CRAN is to split the functionalities of the ordinary Evolved Node B (eNB) into two main components. The first one is the Remote Radio Head (RRH) that contains the Radio Frequency (RF) functionality in which multiple RRHs will be connected to the other shared component that is called the Baseband Processing Unit (BBU) which performs the other functionalities. The link between these two components will be referred to as the fronthaul interface. This interface has many different technologies in which we are interested in this paper in the Next Generation Fronthaul Interface (NGFI). According to NGFI, various functional splits have different associated tasks to both the BBU and the RRH which impacts on the C-RAN performance. In this paper, OpenAirInterface (OAI) platform is used to implement and conFigure C-RAN testbed. Then, the impact of different NGFI splits on C-RAN performance is evaluated and concluded using the implemented C-RAN testbed in terms of various perfomance metrics such as required fronthaul data rate, percentage of CPU usage, and memory usage.
- 26 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 6 (6) , 9774-9782
Matching Game-Based Cell Association in Multi-RAT HetNet Considering Device Requirements
Multiple Radio Access Technologies heterogeneous network (multi-RAT HetNet) composed by cellular and WiFi cell Base Stations (BSs) is considered a promising paradigm for meeting the high traffic demand on mobile networks. In this paper, a new cell association algorithm for different Wireless Devices (WDs) having diverse requirements in multi-RAT HetNet is proposed. The algorithm is designed to consider the WDs’ uplink power consumptions and downlink achieved data rate while taking into account the proportional fairness among them. For this, the WDs association process is formulated as an optimization problem to maximize the WDs downlink achieved data rates considering proportional fairness, and to minimize their uplink transmitted power consumption. To solve this problem, a matching game-based association algorithm is proposed, where WDs are matched to their best cell BS according to well-designed utility functions that consider the rate and power requirements of each WD. Simulation results and performance analysis show that our proposed association strategy outperforms comparable strategies in terms of system throughput, WDs transmit power consumption, and fairness with significant gains.
- 26 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 6 (6) , 9774-9782
Matching Game-Based Cell Association in Multi-RAT HetNet Considering Device Requirements
Multiple Radio Access Technologies heterogeneous network (multi-RAT HetNet) composed by cellular and WiFi cell Base Stations (BSs) is considered a promising paradigm for meeting the high traffic demand on mobile networks. In this paper, a new cell association algorithm for different Wireless Devices (WDs) having diverse requirements in multi-RAT HetNet is proposed. The algorithm is designed to consider the WDs’ uplink power consumptions and downlink achieved data rate while taking into account the proportional fairness among them. For this, the WDs association process is formulated as an optimization problem to maximize the WDs downlink achieved data rates considering proportional fairness, and to minimize their uplink transmitted power consumption. To solve this problem, a matching game-based association algorithm is proposed, where WDs are matched to their best cell BS according to well-designed utility functions that consider the rate and power requirements of each WD. Simulation results and performance analysis show that our proposed association strategy outperforms comparable strategies in terms of system throughput, WDs transmit power consumption, and fairness with significant gains.
- 1 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 307-312
Joint Network and Mode Selection in 5G Multi RAT Heterogeneous Networks
The unexpected explosion of mobile data traffic has led to that the cellular networks are overloaded. Long term evolution (LTE) wireless local area networks (WLAN) aggregation (LWA) has recently emerged as a developed solution of mobile data offloading. In this paper, a multi-RAT heterogeneous network (HetNet) is considered, where three mode categories are defined, LTE mode, WiFi mode, and aggregation mode. A non-collocated LWA scenario is considered for the aggregation mode, in which a wireless backhaul link is connecting LTE with WLAN networks. We use the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) in order to arrange the available modes and their own stations, according to several metrics and based on the user requirements. Thus, helping for optimum mode and network selection that will improve the system performance while satisfying the user requirements. The analytical results show the importance of taking the effect of the backhaul during the mode selection process and how the backhaul link affects the overall performance.
- 1 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 307-312
Joint Network and Mode Selection in 5G Multi RAT Heterogeneous Networks
The unexpected explosion of mobile data traffic has led to that the cellular networks are overloaded. Long term evolution (LTE) wireless local area networks (WLAN) aggregation (LWA) has recently emerged as a developed solution of mobile data offloading. In this paper, a multi-RAT heterogeneous network (HetNet) is considered, where three mode categories are defined, LTE mode, WiFi mode, and aggregation mode. A non-collocated LWA scenario is considered for the aggregation mode, in which a wireless backhaul link is connecting LTE with WLAN networks. We use the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) in order to arrange the available modes and their own stations, according to several metrics and based on the user requirements. Thus, helping for optimum mode and network selection that will improve the system performance while satisfying the user requirements. The analytical results show the importance of taking the effect of the backhaul during the mode selection process and how the backhaul link affects the overall performance.