It is known that there have been numerous improvements in the field of Intelligent Transportation Systems using VANETs which encouraged many to develop intelligent traffic control approaches to replace the current technology using traffic lights. One approach was proposed by Elhadef in [3], which was to improve the inVANETs-based intersection control algorithm that was developed by Wu et al. to adapt to real life traffic scenarios or accidents. It is implemented using vehicle-tovehicle or vehicle-to-infrastructure communications as the vehicles exchange messages to get the opportunity to cross the intersection. In this paper, we present our work in simulating inVANETsbased traffic control for future smart cities using OMNet++, SUMO, and VEINS simulation tools.
We demonstrate an integrated model to retrieve and map satellite image from the image warehouse. Our idea is to detect the geographical location of remote sensing images using deep convolutional neural network (CNN). We train the VGGNet-16 model employing Fully-Connected2 (FC2) features based on a reference dataset based on the closest match found. Once, we obtain the closest match, we use Speeded-Up Robust Features (SURF) to detect the tie points between the new image and reference image. The tie points can be used to register the test image with the reference image automatically. Performance evaluation of our proposed model is performed on satellite image acquired in 2015 using WorldView-2 satellite over Abu Dhabi, United Arab Emirates. We also perform experimentation with Google Earth image of different resolution to demonstrate the robustness of our approach.
Vehicular Ad Hoc Networks (VANETs) are emerging as an enabler for distributed transport application such as traffic management and multimedia sharing. In urban VANET, basic routing protocols are affected by the urban environment elements such as intersections and traffic lights which lead to frequently changing network topology and uneven vehicle distribution. Multiple routing protocols overcome this problem by introducing a street connectivity metric for relay selection. This paper presents a survey of routing protocols using street connectivity prediction in urban VANETs. Moreover, a discussion of the limitations in the surveyed protocols is presented.
In the last decade network science emerged as a new discipline that highly impacted scientific research. Data availability and advanced computational resources attracted researchers to dive in network science looking for new insights and discoveries in almost all disciplines. We are surrounded by networks and the size of these networks is growing exponentially. Bitcoin network is one example of the complex networks that are growing exponentially. In this paper we worked with Bitcoin user graph and applied network sampling to generate sample graphs. We analyzed the basic network properties for original and sampled user graphs to evaluate the sampling method we used in this paper.
In this paper, we propose a vision-based hand gesture recognition system for interaction with a Smart-TV under varying illumination conditions. Vision-based hand gesture recognition systems are employed as human-computer interfaces to increase the comfort of the user, and provide a more intuitive interaction. In our algorithm, a convolutional long short-term memory (LSTM) network is used to classify features extracted from video sequences of hand gestures captured under varying illumination conditions. We experimented this approach with our hand gesture detector, and report a superior classification accuracy on our hand gesture dataset, the dataset consists of eight different hand gestures performed at night-light room ambient lighting conditions six gestures of which are used to be recognised in this paper.
In this paper, the procedures that will be used to implement pathloss measurements for different indoor and outdoor scenarios at a frequency of 28 GHz in the United Arab Emirates will be discussed. The measurements will act as a vital step to understand the behavior of mmWave channels in the desert-like environment. It is expected that the results will show differences in channel behavior in comparison to other countries due to the special weather conditions existent in the United Arab Emirates environments.
The Pursuit of happiness is of much interest but what exactly we can do to measure the state of happy emotion and how to distinguish it from a fake emotion. The main aim of this project is to classify a fake and a real smile using EEG signals. It focuses on using the EEG signals in order to classify a True versus a fake smile. The project involves stages of data acquisition followed by feature extraction and Classification. The data collection involves the recording of EEG signals in response to external stimuli that can elicit a fake smile and a genuine smile. This recorded data is then preprocessed to remove artifacts. The preprocessed signals are utilized for feature extraction using source localization techniques. Source localization technique has been used for feature extraction due to fact that it provides a good estimate of the brain cortex activity and also due to the role of different brain regions to facial expression. Finally, the features extracted are fed onto a classifier.
There is an urgent increasing need for healthcare to be efficient. Healthcare is highly impacted by medical technologies since it is one of the main drivers of healthcare performance and cost. The number of medical equipment and their complexity force hospitals to adopt different maintenance strategies to enhance the performance of their devices in addition to attempting to reduce their maintenance cost and effort. In this work, we are proposing a predictive maintenance strategy that relies on real online data through using the Internet of Things (IoT) technology to predict failure before it occurs. This maintenance strategy along with IoT will form a successful combination to improve the reliability of medical devices and make good use of maintenance resources. We developed a simulation setup to test the methodology using two online tools developed by IBM to show how failure can be predicted and equipment's availability can be improved.
The fight against cancer has pushed scientist and researchers into exploring new effective and innovative forms of cancer therapy. Nanomedicine to deliver site specific chemotherapy is studied to enhance drug delivery by reducing side effects experienced by patients and drug toxicity to noncancerous cells, improving drug accumulation and specificity in the tumor site. With cancer affecting several organs and tissues, each type is treated differently by choosing the most effective nanocarriers, such as liposomes, dendrimers or micelles among others. The use of active targeting by attaching the suitable ligand to the nanocarriers will enhance the drug delivery process. Stealth properties, enhanced permeability and retention effect (EPR), biocompatibility and ease of syntheses are all factors that must be considered when choosing the most effective form of the drug delivery system
Asthma is a chronic inflammatory disease that is treatable but incurable affecting more than 14% of the UAE population. Asthma is still a clinical dilemma as there is no proper clinical definition of asthma, unknown definitive underlying mechanisms, no objective prognostic tool nor bedside noninvasive diagnostic test to predict complication or exacerbation. Here we used a novel in house bioinformatic method on publicly available database to identify novel gene signatures and pathways that can explain the heterogeneous nature of asthma. Our approach showed that Transcriptomic profiling of asthma cases can infer their different phenotypes and can shed light on the cellular pathways and molecular mechanism underlying asthma
