Electrical Engineering Open Day

Students: Rashid Deemas, Abdelrahman Abdelfatah, Khalifa Al Shamsi

Supervised by: Dr. Hasan Al-Nashash

The main objective of this report is to design a portable device that can continuously monitor cognitive vigilance. The device should also contain a feature that produces an alarm when there is a decrement in vigilance. The proposed system is based on calculating the EEG signal power in different EEG bands, then comparing them to a baseline ratio to conclude if there is a decrease in vigilance. The device comprises electrodes, a differential amplifier, a high pass filter, a low pass filter, a notch filter, a non-inverting amplifier, an analog to digital converter, a microcontroller and a Wi-Fi module.

Students: Maryam Bin Sulaiman Al Malik, Osama Al Alawi, Shaikha Mira Al Mualla

Supervised by: Dr. Mostafa Shaaban, Dr. Mahmoud Ibrahim

The project aims to redesign an available golf cart by maximizing solar energy and introducing smart capabilities that will monitor measurements such as voltage, current and temperature. Those readings will be displayed on a digital screen to the user.

Students: Muath Mohammed, Mahmoud Hudaib

Supervised by: Dr. Mahmoud Ibrahim, Dr. Amer Zakaria

In this project, an Arduino-based water pollution monitoring system is designed and implemented. The system will be monitoring several parameters to help determine the quality of the water. The parameters were decided based on research and recommendations by the World Health Organization. The device takes in readings from the water through five built-in sensors. The sensors transmit the data to the Arduino, which conducts the necessary conversions to translate the sensor readings from mV into an actual value which can be used. This data is then sent to a web server using a 4G module. The end user can finally access the data using a dedicated website on the internet. The sensors, microcontroller and 4G module are all powered using our dedicated solar panel with its built-in battery and charge controller.

Students: Tala Al Abweh, Zaineh Al Nasser, Ramzi Al Sharawi, Yasser Juran

Supervised by: Dr. Nasser Qaddoumi, Dr. Shayok Mukhopadhyay, Dr. Amer Zakaria

A novel, contactless and battery-operated microwave based sensor is used to continuously monitor the mechanical vibrations of any structure. The sensor can be operated in any environment without the need for direct line of sight. Data from the sensor is remotely sent to the operator’s PC where the signal processing is done and the data is displayed in the form of a graph showing the vibration frequencies. The sensor is mounted on a robotic base/arm allowing for the system to be mobile where the operator can orient and move the system using a remote controller.

Students: Abdulla Qassim, Mohammad AlFalasi

Supervised by: Dr. Ming Foey Tang

The students will present a low-cost portable function generator that will enable other students to practice electronics and build circuits anywhere in the world without the need to access a lab.

Students: Nour Jamal, Mohmad Saeed, Khawla Hashimi

Supervised by: Dr. Usman Tariq

We demonstrate a deep learning algorithm to hide an image within another image. The hidden image can then be recovered back. This is known as image steganography and can be applied in secretive communication of information. We use convolutional autoencoders to both hide and then later recover a secret image from a cover image. We train the autoencoders in a way that when the image is hidden inside the cover image, it looks very similar to the original cover image. However, when the image is recovered, it looks very similar to the hidden secret image.

Student: Mohammed Farook Maricar
Supervised by: Dr. Amer Zakaria, Dr. Nasser Qaddoumi

This research implements a deep neural network to solve electromagnetic inverse scattering problems. The conventional approaches using inversion algorithms encounter challenges such as high contrast and computational costs. Various deep-learning techniques have been proposed to tackle these issues over the years. In this work, a deep convolutional neural encoder-decoder architecture is implemented and tested with complex inputs obtained by backpropagating (BP) the measured scattered field. The simulation results show that the implemented deep learning method can effectively enhance the image reconstruction quality in much less time.

Students: Hassan Alhilo, Muhammad Yousuf Virani, Joseph Matthew

Supervised by: Dr. Habib Ur Rehman

We demonstrate a smart photovoltaic system that uses a neural network to determine the optimal angle that generates maximum power at a certain time.

Students: Carine Saeed and Ahmed Soliman

Supervised by: Dr. Ahmed Osman and Dr. Mohamed Hassan

We aim to find a way to connect a super capacitor to a lithium ion battery to increase the battery life of an electric car, and reduce the need of having frequent charging stops.

Student: Usama Mozammil Iqbal

Supervised by: Dr. Lutfi Albasha, Dr Hasan Mir

We present the design of an antenna that can be employed for Frequency Diverse Arrays.

Students: Noura A Ali, Meera Al Suwaidi, Hamda Al Mohammed, Hind Al Falamarzi

Supervised by: Dr Lutfi Al Basha

Our project is about an electronic sensor that can detect humans. We plan to use it to detect kids in locked cars as there are many deaths that happen because parents forget their children in the cars.

Students: Ahmed Abdelaal, Faris AtaAllah, Shoaib Ahmed, Ahmad Rizwan

Supervised by: Dr. Habib Ur Rehman, Dr. Shayok Mukhopadhyay

Two battery energy management techniques (BEM) are investigated for an EV traction system which incorporates an indirect field oriented induction motor (IM) drive system. The main objective of the proposed BEM techniques is to regulate the IM's speed while minimizing the lithium ion (Li ion) battery bank state of charge (SOC) reduction and state of health (SOH) degradation

Student: Mennatalla Elbalki

Supervised by: Dr. Mostafa Shaaban, Dr. Ahmed Osman

Ensuring access to water and electricity is essential for human dignity and social development. Today, 2 billion people do not have access to safe drinking water and 730 million people lack access to electricity. It is necessary to optimize the ways in which these resources are consumed and produced. Furthermore, the electricity generation is associated with water production from the combined cycle cogeneration thermal plants. The UAE has to run the power plants at low efficiency during the winter when the electric demand is severely reduced but the water demand stays almost the same. Therefore, a major target currently is to shift to more sustainable water production and decouple electricity and water production.

Student: Omar Zaatar

Supervised by: Dr. Amer Zakaria, Dr. Nasser Qaddoumi

Microwave imaging for biomedical applications is a promising modality due to its numerous advantages. One of its medical applications is the estimation of bone density for monitoring osteoporosis. In this thesis, a novel wearable tomographic microwave imaging (MWI) system is designed and implemented to estimate the bones' electrical properties and thus their health, by imaging the lower limb bone of the human. The main two hardware components of the system are a microwave switch and an array of antennas. In the proposed system, a 2-by-32 microwave switch matrix is designed using a combination of Single-Pole Double-Throw (SPDT), Single-Pole Four-Throw (SP4T) and Single-Pole Eight-Throw (SP8T) switches. The switch matrix will connect a two-port vector network analyzer (VNA) to an array of thirty-two antennas. Further, the proposed design will enable a given antenna in the system to be used simultaneously as a transmitter or as a receiver. Moreover, a controller unit with a predefined logic will be devised to handle the operation of the switches. Further, for the array antennas, thirty-two antennas will be designed and simulated. The antenna type will be chosen due to directivity, size and operating frequency. After implementing the switch and array, the microwave imaging system will be controlled using a computer workstation where data will be processed to solve an inverse scattering problem. The scattering problem is cast into an optimization problem whose input are the calibrated collected data and output are colormaps representing the electrical properties of imaged objects.

Students: Mohammad Rais Ali, Maryam Alnuaimi, Abdalla Alsuwaidi
Supervised by: Dr. Mohamad Hasan, Dr. Ahmed Osman

Misalignment detection of EVs using auxiliary coils whose voltages determine the misalignment values.

Student: Muhannad Alkaddour
Supervised by: Dr. Abhinav Dhall, Dr. Usman Tariq, Dr. Hasan Al Nashash, Dr. Fares Al-Shargie

Research suggests that false news reaches more people than the truth. Malicious use of cheapfakes and  deepfakes includes propagating rumors, sowing discord and corroborating fake news. The detection of cheapfakes in news settings is useful in combating misinformation. Cheapfakes are images or text that have been manually altered to propagate a false narrative. Deepfakes, on the other hand, are media altered by means of deep learning, or AI.

We propose in this work to contextualize the sentiment of the captions during classification, by doing the following:

Sentiment analysis: Compute textual news caption sentiment features and fuse with additional features.
Data annotation: Provide a framework for automatic annotation of a labeled dataset.
Classification: Train an MLP on existing data and features to predict out-of-context labels.

Left (cheapfake): NATO bombing Serbian victim (left), Albanian refugee (right). Right (deepfake): Face of Italian prime minister Matteo Renzi (2019) on actor’s body, claiming he participated in a satirical show.