This study introduces a fully automated collision detection and notification system specifically engineered to safeguard parked vehicles against accidental impacts in densely populated areas such as commercial parking lots and urban streets. The system employs an integrated network of front and rear cameras, proximity sensors, and vibration sensors to provide continuous environmental monitoring around a stationary vehicle. When a foreign object or vehicle encroaches within a predefined proximity, the system initiates real-time surveillance by activating on-board cameras. Simultaneously, visual alert mechanisms, such as high-intensity flashing lights, are triggered to attract the attention of nearby drivers and prevent potential collisions. In the event of physical contact, the system immediately begins continuous video recording, capturing high-resolution footage of the incident. This evidence is securely transmitted to the vehicle owner's mobile device via a dedicated application, delivering instant notification and remote access to the recorded material. The design emphasizes affordability and accessibility, ensuring that advanced vehicle protection is available to a broad user base. By combining proactive collision deterrence with post-incident documentation and real-time communication, the proposed system offers a comprehensive and practical solution to mitigate the risk and consequences of parked vehicle collisions. Experimental validation confirms the system's reliability, responsiveness, and effectiveness in real-world parking scenarios, demonstrating its value as a robust enhancement to vehicular safety infrastructure.

Diabetic retinopathy (DR), a leading cause of vision loss among individuals with diabetes, necessitates accurate and timely diagnosis for effective management. This paper evaluates two classification models: the gray-level co-occurrence matrix (GLCM) and the convolutional neural network (CNN) ResNet-50 architecture, for automated DR diagnosis. The study employs retinal images from Kaggle and Zenodo datasets, assesses model performance, and optimizes the ResNet-50 parameters to enhance classification accuracy. The results demonstrate the superior performance of ResNet-50 compared to GLCM. The achieved accuracies for distinguishing normal and diabetic retinal images are $\mathbf{9 7. 8 8 9 \%}$ and $\mathbf{9 2. 0 5 3} \%$, based on Kaggle and Zenodo datasets, respectively. This indicates a robust performance of ResNet- 50 in multi-class classification tasks and highlights its potential for improving DR diagnosis systems. These findings underscore the significance of advanced computational techniques in early DR detection, offering enhanced diagnostic efficiency and potentially alleviating healthcare burdens.

In recent years, the proliferation of data and sensor measurements in various scientific fields, particularly within the realm of the Internet of Things, has opened new avenues for knowledge extraction through advanced data analysis techniques. However, the presence of outliers and anomalies poses significant challenges, leading to inaccuracies that can compromise analytical outcomes. Outliers are defined as data points that deviate markedly from other observations, often resulting from measurement errors or inconsistencies within the dataset. Their detection and removal during the data cleaning process are crucial for enhancing data quality and ensuring robust analysis. This study systematically investigates the impact of outliers and their detection on the accuracy and performance of various machine learning algorithms and statistical models in regression and classification tasks. A series of MATLAB simulations is conducted on standard datasets to evaluate the effects of outliers and validate the performance of different methodologies. The findings highlight the critical importance of effective outlier detection, demonstrating a marked improvement in the accuracy and reliability of analytical results.

In this research, a comparison between three optimization techniques in the context of hybrid-direct extraction of GaN on Si HEMTs elements that characterize the substrate/buffer loading effect has been developed. The optimization techniques used are Genetic Algorithm (GA), Grey Wolf Optimization (GWO), and Harris-Hawks Optimization (HHO). The open de-embedded structure's Z-parameter have been used to simulate this effect on GaN on Si HEMTs devices. Comparing the measured Z-parameters with the simulated ones has been considered the validation of the techniques. The convergence speed and extraction time have been utilized for the evaluation of the techniques' efficiencies. The results showed the advantage of GWO in terms of fitting error and convergence speed.

This paper demonstrates the applicability of in-house developed Genetic-Algorithm (GA) based Neural Network (NN) model for designing linear- and switching-mode power amplifiers. The modeling procedure was applied on 1-mm GaN HEMT and the developed model was implemented on Advanced-Design-System (ADS) Computer-Aided Design (CAD) software. It is used to design class-AB, class E and class F amplifiers at the frequency of 6 GHz and 10 GHz. The designed Class AB power amplifier at 6 GHz exhibited a maximum gain of 12.4 dB and maximum power added efficiency of 56%. The same amplifier achieved a maximum power added efficiency of 40 % at 10 GHz frequency of operation. The amplifier power added efficiency was then improved using Doherty technique to obtain 78% at 6 GHz and 72% at 10 GHz. The model was also demonstrated by implementing switching-mode Class E and Class F power amplifiers. The results of this paper validate the applicability of our proposed model for the design of linear and nonlinear application circuits.

Worldwide COVID-19 pandemic is currently affecting all countries and led to loss of human life. A lot of scientific research are conducted in different areas to improve the future response. The purpose of the project is to use Machine learning (ML) techniques in predicting COVID-19 deaths which will enhance the hospitals response. This paper contributes by developing models that can predict COVID-19 deaths based on three factors: total number of elderly patients (greater than 65 years), diabetic patients, and smoking patients. Gaussian Process Regression (GPR), Support Vector Regression (SVR), Artificial Neural Network-Multi Layer Perceptron (ANN-MLP), and Artificial Neural Network- nonlinear autoregressive network with exogenous inputs (ANN-NARX) approaches are used to build the predictive models. All models are trained and tested using trusted data reported by the World Health Organization (WHO) in various countries. The developed models revealed very good results with excellent prediction rate and performance, especially GPR, which has the best performance. Also, it showed that region-based predictive models are more suitable than a single general model. The GPR predictive model showed the best performance compared to other models.

Driver's drowsiness is one of the leading causes of road accidents in the UAE and around the globe. Many lives are daily lost because of drowsy driving making an automatic driver drowsiness detection system an urgent necessity for our modern society. Over the past few years, many such systems have been investigated in academic and industry research but none are yet widely used in our day to day life due to high cost, or limited effectiveness. In this paper, we present the first steps of a realtime, non-intrusive, smart drowsiness detection system that works in different real-world scenarios and lighting conditions. Our system utilizes computer vision techniques to detect the driver's face in an infrared video, then a deep neural network predicts whether the driver is drowsy or not based only on their face. Our initial experiments report a promising 94.39% prediction accuracy which outperforms many previously published work especially in night time conditions or scenarios where the driver is wearing sunglasses.

This paper presents an efficient parameter extraction method applied to GaN high electron mobility transistors. The procedure only relies on ${S}$ -parameter measurements at cold bias conditions to extract the extrinsic parameters of a 19-element small-signal model. Hybrid technique of particle-swarm-optimization and direct fitting has been developed and implemented. The extraction procedure has been optimized to consider measurements uncertainty and improve the reliability of the extraction. The procedure has been validated by multibias extraction for different device sizes. A very good agreement between simulations and measurements has been obtained.

In this paper, a reliable small-signal model parameter extraction method for GaN high electron mobility transistor (HEMT) on Si substrate has been developed and validated with respect to different gate width devices. The main advantage of this approach is its accuracy and dependency on only pinched-off and unbiased S-parameter measurements. The developed procedure shows reliable and physically relevant results for the investigated devices and scaled with the gate width. A very good agreement is obtained between small-and large-signal simulations and measurements of the considered GaN HEMTs.