Dr. Yazan A Alqudah

This paper investigates the tracking of Unmanned Aerial Vehicles (UAV) using video feed. Gaussian Mixture Models (GMM) are used to detect the presence of the UAV from a video feed. If the feed is reliable the position of the UAV is detected and tracked. In case the feed is interrupted or becomes unreliable, a Kalman filter is used to predict the location of the UAV. To study the ability of the filter to track the UAV, an actual video capture of flying drone is used, a section of the video is removed to simulate an interruption of the feed. The results show the ability of the filter to accurately predict the location of UAV.

This paper investigates the impact of measurements availability on the accuracy of position prediction using a mobile device equipped with GPS, acceleration, and gyro sensors. The paper also discusses the effects of the sampling factor for the available data. The experiment results emphasize the importance of sampling factor in achieving robust and reliable location estimation, particularly in challenging environments where GPS signals may be intermittent or compromised. The insights gained from this study may contribute to the advancement of location estimation-based applications, fostering improved autonomous systems in real-world scenarios.

This work presents a solution to monitor breathing patterns to detect any signs of abnormalities and ensure properly ventilating pulmonary system. The solution includes the ability to track and detect coughing. The system can be used by individuals to monitor breathing or athletes to monitor performance while exercising. The solution utilizes machine learning algorithms implemented through Edge Impulse to classify and analyze breathing patterns. It also features a user mobile application to record and transmit data and receive the classification results.

Optical Wireless (OW) provides an alternative to RF for wireless connectivity. In applications where Radio Frequency (RF) might interfere with system operations, OW can be used to provide connectivity. This work studies the applicability of OW to connect two agents in a swarm. The work discusses transmitter and receiver designs and shares the results of simulation and lab experiments for different circuits configurations. The work shows the results of a reliable 100Kbps link between agents operating at a distance of approximately equal to 3 meters.

Thermography combined with experimental psychology and image processing has great potential in many applications including marketing, social media, and security. One of the challenges of adopting this tool is the large amount of data needed to get accurate results. Currently, an experimental psychologist manually detects and classifies each thermograph, resulting in time consumption and inefficiency that prevents large data analysis, leading to unreliable results. The Pinocchio effect refers to changes in nasal temperature when someone lies. This is a result of changes of blood flow in the area. This work develops a tool that helps detect and track this effect. The tool allows for large data to be collected and analyzed.

Road strips are used worldwide to alert drivers to changing road conditions and potential danger. These markers are currently passive and work by causing vehicle to vibrate, alerting drivers to pay attention to the road and road signs. In this work, we propose encoding the markers so that a message can be exchanged to help drivers and enhance autonomous vehicle performance. The instants vibrations occur can carry information that can be used by the vehicle. This information can be a warning message or information about the location. The proposed solution can be used by autonomous vehicles or ordinary vehicles equipped with motion sensing capabilities. Utilizing motion gyroscopic sensors, we illustrate how a message can be transferred and decoded from these markers.

This paper aims to develop a low-cost energy storage system by evaluating reference generation algorithms for dispatching solar power for 1 MW photovoltaic (PV) arrays. Based on battery state of charge (SOC), rule-based algorithms are developed to adjust the grid reference power for each one-hour dispatching period. In this paper, several rule-based algorithms are used to control the SOC of the battery that plays a significant role to design cost-effective energy storage system. The price comparison is made between two kinds of energy storage system (i) Battery only (ii) Battery+ Supercapacitor (SC), where a low pass filter is used to allocate power between battery and SC. The most economical energy storage system is developed through extensive simulations in MATLAB/Simulink environment. The results show that the hybrid energy storage system (HESS), combination of battery and SC, outperforms a battery-only operation.

This paper studies the optimum (most economical) scaling of a battery and supercapacitor hybrid storage for 1 MW photovoltaic (PV) arrays for a one hour dispatching period for an entire day. The optimization is based on the time constant of a low pass filter (LPF) that is used to allocate the power between a battery and a supercapacitor (SC). This paper also presents the price comparison between lead-acid and li-ion battery frameworks. Extensive simulations were conducted for thorough analysis of various hybrid energy storage system (HESS) combinations. According to the results, HESS outperforms battery or SC-only operation.

In this work, a novel model of a wave energy converter that utilizes a slider crank power take-off system (PTOS) has been studied to determine maximum energy extraction. This system includes a buoy, slider crank linkage, and generator with a gear box. With a regular sinusoidal wave excitation force, effects of three parameters on the system are analyzed. These include the slider crank radius, connecting rod length, and phase lock offset. It is found that the slider crank radius has the largest effect on total energy extracted. Under irregular wave conditions only the effect of phase lock offset is analyzed. The results show that the control system can exhibit small percentages of performance degradation without large adverse effects on the overall energy production.

Road conditions play a critical role in ensuring traffic safety and reducing traffic jams and congestions. Ensuring healthy conditions require constant monitoring to detect and predict potential road deterioration. This work proposes a low-cost solution that takes advantage of sensory capabilities of smartphones. By recording Gyro rotation sensor data, we show that abnormalities can be detected by calculating the second moment of sensor data. Our work is validated by drive tests that show results are consistent and repeatable. The work also proposed a dynamic time warping technique to measure similarity between drive results and to obtain accurate representation of multiple drives data.