MAMohamed Awadallah

Mohamed Awadallah

*This is an algorithmically generated author record based on publicly available information.

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


Mohamed Awadallah has a total of 10 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Explainable Ai for Knowledge Graph-Based Drug Repurposing: Methods, Challenges, and Interpretability Frameworks

    by Marwa Mustafa, Firas Al–Doghman, Farookh Hussain, Mohamed Awadallah, Mohammed Ali

    The increasing reliance on Artificial Intelligence (AI) in drug repurposing has introduced interpretability concerns, particularly when using complex models over biomedical Knowledge Graphs (KGs). This paper critically investigates the role of Explainable Artificial Intelligence (XAI) in enhancing the transparency and interpretability of AI-driven predictions in KG-based drug repurposing. This paper presents a structured taxonomy of XAI approaches - including graphbased feature attribution, path-based reasoning, counterfactuals, and rule-based logic - tailored for knowledgerich biomedical graphs. Furthermore, an analysis is conducted on current limitations in XAI evaluation, the tension between explainability and predictive accuracy, and the need for audience-specific interpretability frameworks in biomedical contexts. By positioning XAI as a bridge between opaque machine learning models and clinically relevant reasoning, this paper contributes to the development of trustworthy AI systems in biomedical knowledge environments.


  • The Role of Explainable AI in Knowledge Graph-Based Drug Repurposing: Bridging Trust and Discovery

    by Marwa Mustafa, Firas Al-Doghman, Farookh Hussain, Mohamed Awadallah, Mohammed Ali

    Drug repurposing presents a cost-effective and accelerated alternative to traditional drug development. With Knowledge Graphs (KGs) structuring biomedical relationships and AI models performing predictive link analysis, Explainable AI (XAI) emerges as a critical enabler for interpretability, transparency, and trust. This paper explores the central role of XAI in enhancing the reliability, ethical transparency, and regulatory compliance of AI-powered, KG-based drug repurposing. One of XAI's most important contributions is its ability to explain why a drug is predicted to treat a particular disease - by revealing key features, paths, or biological relationships that support the prediction. This allows researchers and clinicians to verify the biological plausibility of AI-driven suggestions, which is crucial for preclinical validation and clinical decision-making. The paper also discusses key XAI techniques - including feature attribution, counterfactual analysis, and path reasoning - and the challenges and ethical implications tied to biomedical use cases.

  • Using a low-cost bluetooth torque sensor for vehicle jerk and transient torque measurement

    by Peter Tawadros, Mohamed Awadallah, Paul Walker, Nong Zhang

    This paper presents the use and development of a specific wireless torque measurement system that is used to obtain the transient torque performance of vehicle transmissions. The torque sensor is strain-based, using surface-mounted strain gauges on a prop shaft. The gauges are connected to a compact printed circuit board, which is clamped to the shaft next to the strain gauges using a three-dimensional printed housing. The printed circuit board contains an amplifier, low-pass filter, analog-to-digital converter, microcontroller and bluetooth transceiver. The printed housing is impact resistant carbon-reinforced nylon and securely retains the printed circuit board and the battery powering the device. The transmitted torque data are received by a transceiver, which is interfaced to a PC through an RS-232 connection. NI LabVIEW is used to process, display and save data. The wireless torque sensor was installed to the Unit Under Test at the output shaft of the five-speed manual transmission. The Unit Under Test was installed on a dynamometer for verification purposes and the transient torque was recorded under various operational conditions. The transient output torque of the manual transmission is measured and compared with results obtained from simulations performed under similar operating conditions. The two sets of transient responses show a good correlation with each other and hence demonstrate that the torque sensor meets the major design specifications. The data obtained will be used to enhance the fidelity of the software model.


  • Comparative fuel economy, cost and emissions analysis of a novel mild hybrid and conventional vehicles

    by Mohamed Awadallah, Peter Tawadros, Paul Walker, Nong Zhang

    Mild hybrid vehicles have been explored as a potential pathway to reduce vehicle emissions cost-effectively. The use of manual transmissions to develop novel hybrid vehicles provides an alternate route to producing low cost electrified powertrains. In this paper, a comparative analysis examining a conventional vehicle and a mild hybrid electric vehicle is presented. The analysis considers fuel economy, capital and ongoing costs and environmental emissions, and includes developmental analysis and simulation using mathematical models. Vehicle emissions (nitrogen oxides, carbon monoxide and hydrocarbons) and fuel economy are computed, analysed and compared using a number of alternative driving cycles and their weighted combination. Different driver styles are also evaluated. Studying the relationship between the fuel economy and driveability, where driveability is addressed using fuel-economical gear shift strategies. Our simulation suggests the hybrid concept presented can deliver fuel economy gains of between 5 and 10%, as compared to the conventional powertrain.


  • An innovative control strategy for a hybrid energy storage system (HESS)

    by Li Sun, Nong Zhang, Mohamed Awadallah, Paul Walker

    Electric Vehicles (EVs) adopting both batteries and supercapacitors have attracted a significant amount of attention in research communities due to its unique power sharing capabilities. A Hybrid Energy Storage System (HESS) can effectively reduce power stress that would otherwise be applied to batteries alone, and whose weight and size is still a common concern when competing against conventional ICE-powered cars. In this paper, a high-level control strategy is developed to adaptively split the load between two sources for an electric vehicle adopting HESS under real-life load fluctuations. A converter - Supercapacitor Pack (SP) coupled HESS upon which such an algorithm is deployed on, is proposed to divert excess power into the SP via a smart Power Converter (PC) which is located in between in order to regulate both behaviors. Such a power split strategy (PSS) is designed in such a way to track real-time load profiles and determines one important variable - the cut-off frequency. A simplified HESS model is first developed. The power split algorithm is coded in Matlab and then applied to this HESS model. Finally, the overall system is tested comprehensively over 4 EPA driving cycles. Simulation results prove its effectiveness in coping with even the harshest driving scenarios in real life.


  • Eliminating the torque hole: Using a mild hybrid EV architecture to deliver better driveability

    by Mohamed Awadallah, Peter Tawadros, Paul Walker, Nong Zhang

    Hybrid vehicle engineering has traditionally and dominantly focused on fuel economy benefits and emissions reductions. Although the transient power delivery benefits of hybrid powertrains are well-understood, these are not a primary focus of the majority of research and development efforts, with some exceptions. Our approach to this problem is to deliver a low-cost, low-tech mild-hybrid powertrain, with unique power delivery features designed to appeal to price-sensitive, but aspirational consumers. The powertrain is a simple post-transmission parallel hybrid configuration. It utilizes a low-powered four-cylinder engine coupled to a four-speed manual transmission through a robotically-actuated clutch. A low-voltage BLDC motor is directly connected to the transmission output shaft, before the final drive. Our research focuses on bringing the benefits of HEV architecture to the world's developing cities, where, it can be confidently argued, local emissions reductions are needed the most. Crucial to the success of this research is the understanding that compared to an equivalent ICE-powered vehicle, an HEV competes at a price disadvantage, no matter how cost-effective the solution is. This disadvantage is amplified in regions of low-middle income, where price sensitivity is greatest. It must, therefore, present better value than an equivalent conventional vehicle if it is to be commercially successful in these particularly price-sensitive markets. We discuss the extent to which control can be used to deliver transient power delivery gains in such a setup, and offer an example powertrain for simulation. To validate the concept, simulation of this research is performed in MATLAB and Simulink. The prototype is based on a generic engine and a BLDC motor. The results mainly focus on the electric drive and comparison of the transient response of drivetrains.


  • Open Access

    An Electric Scooter with Super-Capacitor Drive and Regenerative Braking

    by Li Sun, Mohamed Awadallah, Lianhua Chi, Nong Zhang

    This paper presents a smart electric scooter system consisting of a microprocessor based vehicle controller (integrating an embedded regenerative braking controller), a 300W Permanent Magnet (PM) DC motor, two low-power DC-DC converters to form a higher power DC-DC converter pack, a motor controller, a supercapacitor bank and a capacitor cell balancing sub-system.During acceleration or forward motoring mode, the vehicle controller sets the DC motor into motoring mode to further utilizing motor controller regulate wheel speed and acceleration torque, whereas during deceleration or forward braking mode, sets the DC motor into braking mode and further utilizing regenerative braking controller regulate wheel speed and braking torque, as well as functions as a constant current (whose reference value is adjustable via a potentiometer) generator to charge the supercapacitor bank in a controllable fashion, hence not only successfully replacing frictional braking to certain degree, but also increasing the total energy efficiency dramatically owing to the low internal resistance and larger capacitance of the supercapacitor compared with other conventional regenerative braking systems via batteries.General structure of the smart system, control principle of the controllers, realization of measurement platform, experimental test setup as well as validation with results are all presented within this paper.