Zanco Journal of Pure and Applied Sciences

Accurate Estimation of Robotic Arm Movements for Effective Motion Control: Utilizing Multiple Sensors and Data Fusion

Dler Salih Hasan — Mon, 15 Apr 2024 00:00:00 +0000

The robotic manipulators are highly complex coupling dynamic systems, which require a mathematical model for planning and controlling the robotic motions. It is imperative to calculate the kinematic parameters such as rotational matrix, joint angles, angular velocity, and angular acceleration, which determines the control performance of the models. For this purpose, a multiple-sensor-based Mathematical approach that utilizes inertial measurement unit (IMU) and triple-axis accelerometer is presented in this paper. A combination of one IMU and three triple-axis accelerometers is affixed to each of the two rigid bodies for real-time determination of parameters and the robotic arm orientation. Additionally, the model incorporates an Extended Kalman filter (EKF) fusion technique to combine data from various sensors, mitigate measurement noise, and adapt in real-time to changing environments. To implement this approach, a MATLAB code is developed to read, preprocess sensors data, and simulation of the proposed model. All the results are presented graphically and indicate that the motion parameters and pose measurements are calculated accurately and effectively.

Masked Face Recognition using deep learning models

Omer T. Hamajan , Abbas M. Ali — Mon, 15 Apr 2024 00:00:00 +0000

Face recognition has become indispensable in our daily lives as a quick and painless technique of confirming our identities since in the era of wearing face masks the traditional face recognition system may not effectively recognize the concerned person as an important part of the face (mouth, nose, and chin) which makes a substantial contribution to the face recognition process are occluded and partially hidden. The objective of our research is to tackle the challenges of the partially occluded face with a mask by training our custom dataset using those powerful pre-trained deep learning models like YOLO and Mask R-CNN which have not been used before for this purpose and to compare which one is outperforming the better results. To this end, models like (YOLOv5, YOLOv7, YOLOv8, and Mask R-CNN) have been employed and trained on the created dataset to check the accuracy and robustness of the occluded face recognition process. In addition, an online dataset such as (mfr2) which contains celebrities, and politicians masked and unmasked faces after expansion with more images has been used. The experimental findings demonstrate that the proposed algorithms give an accurate result, we achieve an accuracy of 97.5% using YOLOv8s, an accuracy of 89.7% using YOLOv7, and an accuracy of 89% using YOLOv5x, while an accuracy of 94.5% using Mask R-CNN. The study concludes that YOLOv8s outperforms the other models in masked face recognition.

Ocular Disease Classification Using Different Kinds of Machine Learning Algorithms

Mardin Abdullah Anwer, Ghassan Akram Qattan, Abbas Mohamad Ali — Mon, 15 Apr 2024 00:00:00 +0000

Ocular disease is a term used to describe a wide range of illnesses that affect the eyes and visual system. These diseases can affect one or both eyes and can range from mild to severe. The use of machine learning algorithms to categorize ocular diseases has become an area of interest in the ophthalmology community.

This study is to compare the performance of different machine learning algorithms in classifying ocular diseases based on fundus images. The dataset of fundus images of patients diagnosed with different ocular diseases like Cataracts, pathological myopia, glaucoma, age-related macular degeneration, and abnormalities are considered. Ocular Disease Intelligent Recognition (ODIR) has been used. The SeequzeNet and GoogleNet deep learning models with different machine learning algorithms employed in experimental work includes KNN, random forest, support vector machines, logistic regression, and gradient boosting. The performance of each algorithm is evaluated using accuracy, sensitivity, and specificity metrics. The results show that logistic regression outperforms the other algorithms in terms of accuracy, sensitivity, and specificity. The findings of this study suggest that machine learning algorithms, particularly Logistic Regression, can be useful in accurately classifying ocular diseases based on fundus images. Feature extraction using SeequzeNet achieved an accuracy of 71.6%, outperforming GoogleNet's accuracy of 68.2%.

Documentation of Historical City Using Oblique UAV Imagery: Case Study of the Traditional Terrace of the Jewish Neighbourhood in Akre.

Shireen Younus Ismael — Mon, 15 Apr 2024 00:00:00 +0000

UAV photogrammetry is a crucial technique for conserving cultural heritage sites, particularly oblique images. However, studies on this method are limited. The Jewish Neighbourhood in Akre, with its significant historical and architectural value, needs urgent documentation to prevent further deformation. This article aims to present the methods and techniques of data acquisition and processing to document the traditional terrace form of the Jewish Neighbourhood. UAV was used to capture 337 images at an altitude of 40 meters with a tilted camera angle of 45 degrees. Agisoft Metashape photogrammetry software was used to process the images and produce a 3D digital surface model of the buildings, quantitative and qualitative methods were carried out to assess the accuracy of the deliverables. In addition, Global Mapper and ArcGIS software were used to drive more documentation details that would serve to set the required operational guidelines to preserve the traditional terrace form of this site. Results show that the accuracy of the produced 3D model is high and further proves that UAV photogrammetry using oblique images is an effective technique for documenting such complex historical sites.

Factors affecting Bologna process course registration in Architectural Engineering Department-Salahaddin University

Salahaddin Yasin Baper — Mon, 15 Apr 2024 00:00:00 +0000

Developing the traditional higher educational system towards bologna process is considered as the instruction for reshaping in the higher education system in Kurdistan. This transformation faced crucial difficulties in term of course selection in Kurdistan universities. This study aims to explore thoughts, obstacles and difficulties that face students in term of course selection. Further, it tries to formulate an equation of operative factors that affection course selection in bologna process system. The rationale behind applying bologna process in Kurdistan is related to create common higher education area which intended to increase mobility, and inclusive comparability of programs within similar higher education institutes in the region. More precisely, the intention is to indicate how bologna process course selection have affected what goes on in the classroom by comparing select aspects of studying in both pre-and post-Bologna times at Salahaddin university-college of engineering. The methodology of this study is questionnaire-based cross-sectional survey of a random selection of architectural engineering students plus intensive interviews with experts in the field of architectural engineering. Regression analysis is applied to indicate the most operative factors among predictors (course factors, social factors and individual factors). Results specify that these factors have equivalent consequence on course selection and positively affected the students’ academic performance, in other hands the bologna process is one of the effective factors in improving Kurdistan higher education as comprehensive accreditation system.

Generating 3D City Mesh Model Based on Aerial Imagery: A Case Study, College of Engineering Campus

Haval AbdulJabbar Sadeq — Mon, 15 Apr 2024 00:00:00 +0000

3D city modeling is considered one of the most prominent products in urban planning, simulation, and visualization. Although manually producing 3D models is considered to be very time-consuming and costly, this paper presents the pipeline for generating and sharing a 3D city model from aerial imagery based on the photogrammetric technique. For the 3D city mesh modeling, aerial images with a resolution of 10 cm have been used. Initially, aerial triangulation was achieved to obtain the exterior orientation parameters of the images using ground control points that were measured at the site. Later, the obtained exterior orientation parameters are used in SURE photogrammetric software to produce a dense point cloud and mesh model using the image-matching technique. The produced point cloud and mesh model have been shared using the Cesium website, which has been embedded into the website https://3d-erbil.weebly.com/ for easy access. The published model allows the user to better understand the buildings in the area, which is useful for urban planning. Furthermore, it is possible to make measurements from the shared model and obtain object coordinates for any point on the ground.

Driver Behavior during All Red Interval at Signalized Intersections in Erbil City

Aso Faiz Saeed Talabany — Mon, 15 Apr 2024 00:00:00 +0000

This case study investigated red light violations at urban signalized intersections in Erbil city-Iraqi Kurdistan Region. The results obtained in the study are important to improve traffic safety at signalized intersections. Field observations were conducted at six at-grade and grade separated signalized intersections which were selected in different locations in the city. Drivers’ behavior at these signalized intersections based on direct field data collection and video recording technique for the whole observed time is examined. The video recording technique is used to abstract the driver behavior traffic data during the onset of all red light. Also, some camera pictures are taken to assist studying the driver behavior while crossing the stop line after the onset of a red light. The analysis is conducted to identify the driver behavior during red light gathered from the data collected of a total of 328 violation records which are lower than the red light violation rates recorded by other studies.

The calculated red light violation rates for the 6 studied intersections ranged from 0.58 violation per 10,000 vehicles-cycle to 1.52 violation per 10,000 vehicles-cycle. The results show that average aged male and female drivers are more probable to run red lights. In addition, the male drivers have much higher tendency to run the red light than female drivers. Most red light violating drivers in Zanko and Mashtal signalized intersections are young because they are near the college town. Most of red light violations occurred during the peak hours between 8:00 a.m. to 9:00 a.m when most urban driving is done. About 95.34 % of the red light violation records are passenger cars (private and taxi). Approximately 17.38% of drivers tended to stop after onset of the red light. The drivers crossed the stop line after the onset of red light and not stopped are 76.22% within the all red interval and only 6.4% after the end of all red interval. About 51.0 % of the violators ran the red light within 2.0 seconds of its onset, while over 95.0 percent of the violators ran the red light within 4.0 seconds of its onset. The average speed of the red light violator vehicles is 34.95 kph. The most frequent vehicle speed at the time of violation is 32.14 kph. For all studied signalized intersections, 8.0 – 26 % of violating vehicles ran the red light at speeds ≤ PSL, while 74.0 – 92.0 % of violating vehicles ran the red light at speeds > posted speed limit (PSL). The number of red light violations (RLVs) are high at low and high traffic volumes but they are low for medium traffic volumes. The red light violation increases as the signalized intersection control delay value increases and it decreases as the control delay decreases. 56.3% of red light violators are straight ahead, while (43.7%) are left turn vehicles. Simple regression analysis is carried out using IBM SPSS to model the effect of elapsed time since red light onset, violating vehicle speed, traffic volume, signalized intersection control delay, and average approach speed on red light violation (RLV).

The most suitable relationship between the RLVs and elapsed time since red light onset is the fourth-degree polynomial model with a coefficient of determination R² = 0.9931, the third-degree polynomial model for vehicle speed since red light onset with a coefficient of determination R² = 0.9519, second-degree polynomial for the signalized intersection traffic volume counts, positive second-degree polynomial for signalized intersection control delay with a coefficient of determination R² = 0.8493, the third-degree polynomial model for the observed mean approach speed with a coefficient of determination R² = 0.773.

Seasonal variation, group/ solitary behavior occurrence and reproduction time of the predator leech Helobdella stagnalis (Hirudenia: Glossiphoniidae) in Sarchnar Stream/ Sulaymaniah Province, Kurdistan Region- Iraq.

Shayan J. Hamad Ali Hallaq, Younis S. Abdullah, Samir J. Bilal — Mon, 15 Apr 2024 00:00:00 +0000

A total of 194 specimens of the predatory leech Helobdella stagnalis were collected from Sarchnar Stream, in Sulaimaniyah City along year of 2022 from January to the end of December. The seasonal distribution and occurrence, time of egg laying and hatching in relation with some physico-chemical factors were studied, including air and water temperatures, dissolved oxygen, pH, FCO₂, HCO₃^-, Cl^-, Ca⁺² and Mg⁺. The temperature and dissolved oxygen were the most effective factors on leech’s population density as well as time of egg lying and juvenile hatching. The highest density was in Autumn, while the lowest was recorded during Summer. Cocoons were seen to laid in late winter and maximum juvenile numbers were recorded during Spring.

Existence Canard Solutions For Four Dimensional Hindmarsh-Rose Model with Respect to Infinitesimal Parameter

Chiman Qadir, Ibrahim Hamad, Waleed Aziz — Mon, 15 Apr 2024 00:00:00 +0000

This research endeavor seeks to investigate the presence and viability of canard solutions within the context of the generalized Hindmarsh-Rose model, specifically when extended to four dimensions. To achieve this, nonstandard analysis is employed as a powerful tool for identifying and characterizing canard solutions within the four-dimensional singularly perturbed system, wherein two fast variables are considered in the folded saddle case. By undertaking this rigorous approach, we aim to contribute valuable insights to the understanding of canard phenomena in complex dynamical systems.

Identification and quantification of the main organic components in artisanal apple vinegars from Iraq Kurdistan region by 1H NMR spectroscopy

Lana H. Chawishli — Mon, 15 Apr 2024 00:00:00 +0000

The proton NMR spectra of apple vinegar samples obtained from the farms of Kurdistan region-Iraq, have been thoroughly examined and reported. Many organic molecules from various classes were assigned, including organic acids, alcohols, volatile compounds, and amino acids. Without extraction or pre-concentration processes, the possibility of quantifying the compounds that were present in the whole vinegar sample was also investigated. The results showed that ¹H NMR with water suppression allows a quick simultaneous determination of acetic, formic, lactic, malic, succinic, tartaric acids, ethanol, methanol, acetoin, 2,3-butanediol, glucose and fructose, by using dimethyl sulfone (DMSO₂) or potassium hydrogen phthalate (KHP) as internal standards. The ¹H NMR method was applied to differentiation of the various samples of apple vinegar.