International Journal of Integrated Engineering

A Smart IoT-Based Prototype System for Rehabilitation Monitoring

Mad Kaidi H. — 2023-08-15

Smart healthcare is growing significantly in the healthcare sector due to the Internet of Things. A remote monitoring system is one of the smart healthcare implementations for rehabilitating stroke patients. Nowadays, as the COVID-19 pandemic continues to spread, patients undergoing home rehabilitation have difficulty meeting with their physicians due to movement constraints. In addition, the healthcare facilities are devoted to treating patients with COVID-19. As a result, physicians and patients could not frequently meet to gather their rehabilitation progress. This study involves developing a prototype to monitor a post-stroke patient's rehabilitation process using the Arduino Nano 33 Bluetooth Low Energy (BLE) and force-sensing resistor (FSR). The prototype analyzes critical aspects of the rehabilitation process based on handgrip, heart rate, sleep, and step tracking measurements. The results of the handgrip, heart rate, sleep, and step tracking measurements are evaluated for various types of subjects and six testing approaches showed an accurate and consistent results. However, experiments partially success with a small error is detected while tracking the steps of each subject. Several recommendations are highlighted to improve the prototype using other sensors such as force sensing resistor and flex sensor for handgrip force transducer, electromyogram (EMG) sensor for stroke-patients rehabilitation, and others.

Analysis On Drone Detection and Classification in LTE-Based Passive Forward Scattering Radar System

Noor Hafizah Abdul Aziz — 2023-08-15

Long-Term Evolution (LTE) is most commonly used in connection with 4G networks with high spectral efficiency, high peak data rates, flexible in frequency and bandwidth. By utilizing LTE signal in passive forward scattering radar as transmitter, this system is able to create a microwave domain at the radar's receiver part which generated a moving object's Doppler signature. The emergence of guided missiles, humans, airplanes, and drones that travel through between the forward scatter radar systems can really be spotted with this passive radar system. This study's primary goal is to employ passive forward scattering radar and an LTE signal to detect drones, which are commonly used by individuals to violate or invade private and secure places. In detail, a drone was detected at two distinct heights of two meters (lower) and three meters (higher) from the ground by utilizing passive forward scattering radar to generate Doppler signature of the flying drone. This experimental work is conducted at two locations which are Taman Suria (UiTM, Shah Alam) and Teluk Kemang (Port Dickson), due to the telecommunication transmitter antenna transmits Long-Term Evolution (LTE) signal with frequency of 1.8 GHz and 2.6 GHz. The results of drone detection at various heights were evaluated using Principal Component Analysis (PCA) on all the experimental data obtained. According to the evaluation, the lower height of the drone performed better in classification and confusion matrices analysis than the upper height due to a larger cross-sectional area for the lower height of the drone that travelled through the forward scatter zone. In summary, the overall study clearly demonstrates the effective categorization of flying drone detection at upper and lower positions in Principle Component Analysis (PCA). For future contribution of this research, it can be used at the airport to detect any unwanted drones trespassing the flight departure area, and important areas such as the Federal Administrative Centre of Malaysia, Putrajaya for spying purposes.

Recognition of Radar-Based Deaf Sign Language Using Convolution Neural Network

M. D. H. Dol Malik — 2023-08-15

The difficulties in the communication between the deaf and normal people through sign language can be overcome by implementing deep learning in the gestures signal recognition. The use of the Convolution Neural Network (CNN) in distinguishing radar-based gesture signals of deaf sign language has not been investigated. This paper describes the recognition of gestures of deaf sign language using radar and CNN. Six gestures of deaf sign language were acquired from normal subjects using a radar system and processed. Short-time Fourier Transform was performed to extract the gestures features and the classification was performed using CNN. The performance of CNN was examined using two types of inputs; segmented and non-segmented spectrograms. The accuracy of recognising the gestures is higher (92.31%) using the non-segmented spectrograms compared to the segmented spectrogram. The radar-based deaf sign language could be recognised accurately using CNN without segmentation.

Focal Region Ray Tracing of Conventional and Shaped Lens Antenna

M. Muhsin — 2023-08-15

Multibeam lens antennas are useful for 5G mobile communication systems. For multibeam design, feed position determination of specified lens becomes an important subject. The feed position data can be obtained from focal region ray tracing. In this paper, a simple and convenient ray tracing method by using the MATLAB function of “polyxpoly” is proposed. To evaluate the program, focal region ray tracing on conventional hyperbolic planar, spherical convex, and Abbe’s sine condition lens are presented in this paper. For focal region ray tracing, parallel incident ray is considered. The incident ray’s angle is changed from 0 degree to 30 degrees with an interval of 10 degrees. On hyperbolic planar and spherical convex lenses, ray concentration at the focal region is distorted at incident angle of 10 degrees and above. On Abbe’s sine condition lens, good concentration is maintained until 20 degrees. Concentration points agree well with the theoretical value. Therefore, the correctness of the program is ensured.

Multibeam Array Antenna with Compact Size Butler Matrix for Millimeter-Wave Application

Noorlindawaty Md Jizat — 2023-08-15

New radio wave technologies of millimeter-wave (mmWave), compact cell size, and multi beam base station are introduced with the recent development of the 5G mobile system. The Butler Matrix (BM) feed circuit is the most preferable candidate for the 5G mobile system since it can achieve multi beam radiation patterns at the array antenna, provide structural compactness and produce good multi beams. The BM circuit is typically built on a single dielectric substrate. However, in this single-substrate structure, the micro strip line connecting several circuit elements in the BM spans over a large area, resulting in significant feeding loss in the millimeter frequency band. In this study, a compact size circuit configuration of BM is proposed, where the original single-substrate structure is modified into a two-substrate stacking structure. The via-hole is designed to connect the two substrates with minimal path loss. The BM is built for the 28 GHz band with four inputs and four outputs. The phase delay is optimized using via-hole to produce the phase difference of ±45º and ±135º. The coupling for the hybrid is -3 dB, while the transmission coefficient of -6 ± 3 is achieved from the BM structure and, the return loss (S_ii) for both input and output ports are less than -10 dB. The two-substrate BM is combined with the rectangular patch antenna and the via-hole patch antenna in a planar configuration of 0.5 λ₀ spacing to obtain the radiation patterns. When the Port 1 through Port 4 of the BM are fed, four beams are created, with peak gains of 11.2 dBi, 9.87 dBi, 10.2 dBi, and 11.7 dBi, respectively, towards +16°, -35°, +39°, and -12°. The analysis includes the radiation performance from the ideal value and from the BM input. Three-dimensional representations of good multibeam radiation patterns are obtained after each input signal of the BM is fed.

Textile Characterization for Wearable Antenna Application Using Transmission Line Method

Anis Fariza Md. Pazil — 2023-08-15

This proposed work introduces the textile characterisation analysis based on the effects of fabric thickness and dielectric properties using the transmission line method for a wearable textile antenna. The return loss (S₁₁) and transmission loss (S₂₁) were analysed for denim, felt and Tencel in correlation with the conductivity properties using Computer Simulation Technology with Rogers 5880 as the reference sample. By varying the fabric thickness from 0.5 mm to 4 mm, the optimum thickness and type of fabric can be identified from the S₁₁ and S₂₁ parameters. In the transmission line simulation, the sample-under-test is a stripline with the conductivity values from 10^-2 to 10⁸ S/m. Results are further plotted against the thicknesses to observe the behaviour of all three textiles samples. The results show that Felt substrate with 3 mm thickness give the best performance. The felt substrate demonstrates the best transmission performance judging from the lowest transmission loss due to the low tangent loss followed by Tencel and denim. For accuracy analysis, the actual and calculated conductivity were also presented to show the textiles performance at low and high region of conductivity. The study reveals the importance of choosing the correct substrate with suitable dielectric and electrical properties before being implemented into the antenna design for good efficiency.

Investigation on Dielectric Properties of Sludge Waste from Water Treatment Using Microwave Non-Destructive Testing (MNDT)

M. A. H. Zable — 2023-08-15

The demand for water cleanup rises in tandem with a country's requirements and development. Recovery of purified water containing nutrients and other beneficial materials is a critical opportunity that must be taken advantage of. A challenge that needs to be tackled is the necessity for large capacity and high-value management of sludge waste following the water treatment process. The pH level and microwave frequencies influence were used as a starting point for assessing the content of the sludge waste. Microwave non-destructive testing (MNDT) is a microwave measurement that can be used to determine the dielectric characteristics of materials without destroying or modifying the sample's content. The methodology employs a free-space measurement technique with a frequency range of 8 to 12 GHz (X-band). Through S-parameters acquired, a correlation analysis was done to analyze the effect of frequencies with the sludge waste. A comparative investigation with peat soil samples in establishing if the sludge has similar attributes to normal soil is used to ensure the accuracy of the sludge waste data. It can be determined that the sludge waste has a high signal correlation towards the frequency band 8 GHz to 12 GHz, which is compatible with the MNDT approach. All of the sludge samples had a pH range that is appropriate for agricultural use.

Characterizing Passively Q-switched Fiber Laser in LiDAR Application

Mahroof Mohamed Mafroos — 2023-08-15

A LiDAR system consists of a Q-switched fiber laser that emits light pulses to measure the distance from the target. We have experimentally demonstrated a passively Q-switched erbium-doped fiber laser (EDFL) by employing graphene saturable absorber (SA). The SA was prepared by dipping a polyvinyl alcohol (PVA) thin film into the graphene solution. Once the SA was fabricated, it can be placed in the cavity to perform pulses and it is operating at 1558.92 nm. The shortest pulse received is 3.9 µs and generated at the repetition rate of 115 kHz. The pulses are stable between pump powers of 59.6mW and 127.1 mW. At the maximum pump power value of 127.1 mW, The laser cavity produced pulses with a 1.4mW output power and a 1.1nJ pulse energy.

Design of Normal Mode Helical Antenna for Seawater Application

Siti Harliza Mohd Razali — 2023-08-15

To overcome the substantial propagation loss in the ocean, a low frequency band should be chosen when contemplating radio wave communication. Then, the antenna used at the portable radio equipment should be very small compare to the wavelength. For the miniaturized antenna, the normal mode helical antenna (NMHA) is suitable because of achieving high efficiency. In this paper, design method of NMHA in the seawater condition (_r=81, =4) is explained. The frequency bands of 100MHz and 15MHz are selected. The self-resonant structures are determined based on the theoretical equation. Electrical characteristics such as resonant frequency, input impedance, VWSR and antenna efficiency are obtained through simulation by FEKO simulator. Electromagnetic simulation model is formed by taking into account the practical experimental condition. At 100MHz and 15MHz, resonances are ensured for antenna diameter lengths of 6.83 cm and 12.2 cm respectively. Input resistance of 27.15 and 94.14, antenna efficiencies of -14.88dB and -26.27dB are achieved at 100MHz and 15MHz respectively.

Metallic Cylinder Reflected Power Measurement For 93.1GHz Frequency Modulated Continuous Wave Radar Calibration

P. N. Ja’afar — 2023-08-15

A metallic cylinder is one of the best materials and shapes to calibrate a radar system performance. The measurement of a 4cm diameter and 3cm-height metallic cylinder as a target has been presented for the analysis at the millimeter-wave (mm-wave) spectrum. This experiment was conducted in a real airport environment at Kuala Lumpur International Airport considering clear sky conditions. The measurement was carried out at 93.1 GHz which uses Frequency Modulated Continuous Wave (FMCW) radar to consistently detect the target. The radar cross-section (RCS) of the metallic cylinder is measured with respect to the angle of runway pavement. It is found that the measurements exhibited smaller RCS value with an average of –43.47 dBsm at a longer range compared to –30.16 dBsm at a shorter range with a total change of 13.31 dBsm. The reflectivity characteristics of the radar target, theoretical measurement of the metallic cylinder, its incident angle from the radar target, and measurement evaluation are presented in this paper.

Convertible Bandstop to Allpass Filter using Defected Ground Structure with Ideal Switch for Millimeter-Wave Band in 5G Application

Adib Othman — 2023-10-02

According to this study, a defective ground structure (DGS) with an ideal switch can be used to create a bandstop to allpass filter for 5G applications. The redesigned Hairpin DGS's bandstop and allpass responses are mathematically investigated in this paper. Utilising an ideal switch via open circuit and short circuit conditions on DGS, the convertible filter is operated. Therefore, the filter's performance in terms of return loss, attenuation, and insertion loss is simulated. As a result, the filter operates at 25.875 GHz in open circuit condition with a narrowband (2.16 GHz) bandstop response at 10 dB and a maximum attenuation of 29.5 dB, and at 26 GHz with a wideband allpass response and return loss greater than 10 dB. As a result, the filter is appropriate for 5G applications that use millimeter-wave RF front-end systems

Design of Complex Multiplier Using Vedic Mathematics

Hasliza Hassan — 2023-10-02

In this project, a 4x4 multiplier is implemented that utilizes the Urdhava Tiryakbhyam sutra method in Vedic mathematics. This method is applicable in all two decimal number multiplications which offers high speed calculation and improved efficiency. Thus, the design of a 4x4 Vedic-based multiplier is solely aimed at performing faster multiplications and achieving quicker processing speeds than the traditional multipliers. The architecture of the Vedic multiplier consists of four 2x2 multipliers and three adders of different bit sizes that are assembled using the Wallace tree implementation. The coding for the multipliers and adders is written in Verilog Hardware Description Language (HDL) in the Quartus Prime 17 Software. Functional simulation is then carried out to ensure that the Vedic multiplier performs the accurate multiplication operations, while the Verilog Compiled Simulator is employed to compile and simulate the multiplier design. Following this, the Design Compiler (DC) and Integrated Circuit Compiler (ICC) command scripts are then composed to allow the logic and physical synthesis to be performed on the Vedic and traditional multipliers. From there, the performance level of both these multipliers are assessed through reference to several key parameters such as timing, area, power consumption, overflow percentage and congestion statistics. Based on the results obtained in the synthesis process, the Vedic multiplier possesses faster operational speed than the traditional multiplier (due to a shorter processing time), but ultimately exhibits a greater power consumption and wider area coverage.

Automatic Generation Control System: The Impact of Battery Energy Storage in Multi Area Network

Norhafiz Salim — 2023-10-02

Renewable energy sources (RES) are currently experiencing significant expansion, and the integration of these sources into power systems necessitates more complex auxiliary facilities. Battery energy storage systems (BESS) have been widely recognized in recent literature as an effective means of enhancing control capabilities. This study focuses on the implementation of an Automatic Generation Control (AGC) system with the integration of BESS in a multi-area network. Maintaining system frequency, especially during peak loads, poses challenges for AGC systems. The objective of this study is to investigate the utilization of BESS to enhance AGC for frequency control in power system networks. Additionally, the effectiveness of BESS in improving frequency control in multi-area networks is demonstrated through several case studies. The AGC and BESS simulations were conducted using MATLAB Simulink to evaluate the proposed frequency control method's effectiveness.

Development of Pesticide Sprayer Robot Prototype for Chilli Farm Agricultural Application

Sumaiya Mashori — 2023-10-02

Accidental spills of chemicals, leakages, poor spraying equipment, or a lack of knowledge about personal protection equipment usage can expose farm workers to pesticides either in open fields or greenhouse. Pesticides may affect their health, particularly on dermal contact and respiratory system. Besides that, traditional pest management for massive agriculture industry requires a big number of employees, large quantities of pesticides, and the risk of inefficient and uneven pesticides coating. Therefore, a pesticide spraying robot, which is equipped with two ultrasonic sensors is developed to spray pesticide independently and allows the workers to maximize pesticide usage on the agricultural application such as chilli farm. The prototype is designed to automatically moves and operate in two modes, either automatic or manual. For automatic mode, the robot travel and operated independently without human control or intervention. While for manual mode, the Blynk application is used to remotely control the robot’s movement and spraying insecticide. In this research, the garlic extract and dishwashing liquid detergent were used as organic insecticides since they are effective and safe for plants and the environment. The study discovered that for a 50 cm two-month chilli plant, a three-seconds pesticide setting was appropriate, with 30 ml pesticide applied per chilli plant.

Downlink Massive MIMO Systems: Reduction of Pilot Contamination for Channel Estimation with Perfect Knowledge of Large-Scale Fading

Qazwan Abdullah — 2023-10-02

Massive multiple-input multiple-output (MIMO) technology is considered crucial for the development of future fifth-generation (5G) systems. However, a limitation of massive MIMO systems arises from the lack of orthogonality in the pilot sequences transmitted by users from a single cell to neighboring cells. To address this constraint, a proposed solution involves utilizing orthogonal pilot reuse sequences (PRS) and zero forced (ZF) pre-coding techniques. The primary objective of these techniques is to eradicate channel interference and improve the experience of end users who are afflicted by low-quality channels. The assessment of the channel involves evaluating its quality through channel assessment, conducting comprehensive evaluations of large-scale shutdowns, and analyzing the maximum transmission efficiency. By assigning PRS to a group of users, the proposed approach establishes lower bounds for the achievable downlink data rate (DR) and signal-to-interference noise ratio (SINR). These bounds are derived by considering the number of antennas approaches infinity which helps mitigate interference. Simulation results demonstrate that the utilization of improved channel evaluation and reduced loss leads to higher DR. When comparing different precoding techniques, the ZF method outperforms maximum ratio transmission (MRT) precoders in achieving a higher DR, particularly when the number of cells reaches .

Hydrogen Peroxide (H2O2) Biosensor Based on the Conducting Polymer Using Self-Assembly Technique

Siti Amira Othman — 2023-10-02

Biosensors are in principle fabricated by immobilized biomaterials on a detector membrane and combining them with electrochemical equipment. The applications of enzyme-based biosensors can be explored such as in the process of gas detection, medicine, pathogen detection and detection of toxic levels of substances before and after bioremediation. In this study, H₂O₂ detection was performed using HRP/PANI, HRP/PPY, HRP/PT and HRP/PT/PPY/PANI layers. The HRP/PANI layer from Variable Pressure Scanning Electron Microscopy (VPSEM) image exhibited a dry surface. The HRP/PPY layer exhibited a surface with agglomerate molecules. The HRP/PT layer, on the hand, exhibited a layer surface with almost the same molecular size. This is confirmed by the higher surface roughness value for HRP/PPY compared to other layers obtained via characterization with Atomic Force Microscopy (AFM). The increasing current response for all three layers was arranged in HRP/PANI> HRP/PT> HRP/PPY. VPSEM and AFM images exhibited surfaces with molecules being in an agglomeration state after the H₂O₂ detection process. In terms of current response, the response rate of H₂O₂ on the surface of the HRP/PT/PPY/PANI electrode caused the current response obtained to be fast. The roughness value increased with time due to the reaction that took place between the surface of the HRP/PT/PPY/PANI layer with H₂O₂. The day-based current response showed that day 1 to day 14 exhibited a uniform graph pattern but from day 21 to day 30 there was a change in the graph pattern due to the HRP/PT/PPY/PANI layer undergoing degradation. The activity of the HRP enzyme was studied by looking at its absorption effect for 30 days. From day 1 to day 14, there was a difference in the overall rate of absorption. However, from day 21 to day 30, the rate of absorption remained constant which explains the slowing down of HRP activity

Development of a Wireless Monitoring System to Monitor River Water Levels in Real Time

Siti Norshafika Mohd Zain — 2023-10-02

This wireless monitoring system is developed to alert and warn the resident regarding the upcoming flood. Due to the floods that occur every year, especially during the monsoon season, the loss and damage done by this disaster are uncontrollable. When the water level rise-up above the safe ground, it may hit the communication substation, as well as the power tower. In most cases, many residents are not alert with their surroundings because there is no alert system that can remind them regarding to the issues that will arise. By not preparing themselves for the upcoming nature disaster, it could endanger themselves and destroy their property. Thus, this project focus on monitoring a real-time river water level system for flood prediction purposes which utilize three devices connected together in a centralized system. By having this system, people will be notified about the current situation and the water level of the river near to them via Internet of Things (IoT) such as Blynk application. ESP32 and ESP32-Cam are used as the main microcontroller to control the system. The sensor used to detect the water level at the river is TF Mini Lidar and ultrasonic sensor for the water level detection of the rain gauge. The ESP32-cam are used to monitor the surrounding area at the river. The combination of these three systems is used to monitor and predict the flood occurrence.

Time Monitoring System for Assembly Line Operators Using RFID

Maisarah Hassan — 2023-10-02

This project discussed monitoring system in an assembly line to monitor and track the performance of the operators in industries. Industry is an important workplace around the world including research and development, design, process, production, quality assurance and so on. A large number of jobs involve humans, and there are also parts that require humans to work using machines, especially operational workers or operators. There are many systems developed in the industry that involve the operator, some of them are to monitor the machining process, and some are to monitor and control the machine. Assembly line operator performance is a critical aspect that severely affects the overall production performance in any production nature. Non-productive time contributes to inequality in task distribution that often results in bottleneck points on a production line. Therefore, the non-productive time is considered as a waste of resources that has a relative impact on the production yield of an industry. In this project, Radio Frequency Identification (RFID) is used on the production line as a monitoring system so that it can help organizations to increase the performance of the working environment among all the employees as well as to achieve better production yield by reducing the non-productive time of assembly line operator.

High Isolated 10-MIMO Antenna Elements for 5G Mobile Applications

Fayad Ghawbar — 2023-10-02

The enormous increase in gadgets has resulted in a data rate shortage insufficient to satisfy the user's needs. The multiple input multiple output (MIMO) technique is substantially deployed in the 5G wireless communication system to increase channel capacity and provide sufficient throughput. However, MIMO antennas are associated with mutual coupling, especially between closely spaced antenna elements, resulting in a low MIMO performance. Therefore, effective isolation techniques are essential to reduce the mutual coupling between the adjacent MIMO antenna elements. A hybrid decoupling technique of self-isolation and an orthogonal mode approach has been proposed to provide significant isolation for high MIMO order 5G mobile applications. A compact self-isolated 10 × 10 MIMO antenna system has been proposed for 5G mobile phone applications operating at the 3.5 GHz frequency band. The antennas act as radiating and isolating elements simultaneously, providing significant isolation. Furthermore, the self-isolated 10-MIMO antenna elements are printed on double side edges of FR-4 small substrates orthogonal to the system substrates, forming an orthogonal mode that enhances the self-decoupling approach. The s-parameters results indicate significant isolation of less than -19 dB between the adjacent 10-MIMO antenna elements. Likewise, the evaluation results of the MIMO performance metrics such as envelope correlation coefficient (ECC), diversity gain (DG), total active reflection coefficient (TARC), and channel capacity Loss (CCL), are less than 0.006, 9.97 dB, -10 dB, and 0.08 bits/s/Hz respectively. The isolation result and the evaluated MIMO performance metrics demonstrate that the proposed 10-MIMO antenna system is sufficient for 5G mobile applications.

An Investigation of Microwave Tomography Technique to Image Brain Tumour Through Cross-Section Imaging with Different Number of Electrode

Anisah Abd Wahab — 2023-10-02

Brain tumours resulted from the irregular growth and cell division within the skull, indicating a high risk for malignancies to develop and can lead to brain injury or even death. The brain tumour can affect nervous system’s function based on the tumour’s growth rate and location. Early detection of brain tumour is essential to improve patients’ survival rates through appropriate medical care. As the current clinical imaging has a few impediments e.g. radiation-based and expensive, tomography technique is seen possible to provide safe and inexpensive technology. The aim of this research is to investigate the feasibility of brain tumour detection using microwave tomography technique with different numbers of electrodes. The 2D finite element modelling approach is applied, and the images are reconstructed using a linear back projection (LBP) algorithm in MATLAB. A different number of rectangular sensing electrodes are arranged around the head phantom in an elliptical array, working in pairs as transmitters and receivers. The simulation shows that the system is able to detect the permittivity difference, thus detecting the existence of the tumour in the head phantom.Theimage reconstruction presented promising tumour images with an 8-antenna microwave tomography system at all locations, i.e. left, right, top, centre, and bottom, in comparison to 4-antenna and 12-antenna systems.

Research Opportunities for Application of Gamification Models with VR for Crop Cultivation: A Systematic Literature Review

Suhendi — 2023-10-02

The learning delivery model in the increasingly developing information technology era and the era of teaching and learning between students and lecturers during the Covid-19 pandemic requires lecturers' creativity. The application of gamification using virtual reality is an alternative solution that can be applied to fill the saturation of online learning using Google Meeting or Zoom meeting. This review of the literature aims to see how many gamification applications using virtual reality are specifically applied to plant cultivation. The process of obtaining this literature review is based on the systematic literature review (SLR) method. The assessment process is carried out through four online databases based on the keywords gamification and virtual reality. The results obtained are 32 relevant literature on gamification using virtual reality although gamification using virtual reality for plant cultivation is not found and only one paper on forestry only discusses the concept. However, the results of this process become the basic literature for further research on the application of virtual reality gamification in plant cultivation.

COVID-19: Symptoms Clustering and Severity Classification Using Machine Learning Approach

Nurul Fathia Mohamand Noor — 2023-07-31

COVID-19 is an extremely contagious illness that causes illnesses varying from either the common cold to more chronic illnesses or even death. The constant mutation of a new variant of COVID-19 makes it important to identify the symptom of COVID-19 in order to contain the infection. The use of clustering and classification in machine learning is in mainstream use in different aspects of research, especially in recent years to generate useful knowledge on COVID-19 outbreak. Many researchers have shared their COVID-19 data on public database and a lot of studies have been carried out. However, the merit of the dataset is unknown and analysis need to be carried by the researchers to check on its reliability. The dataset that is used in this work was sourced from the Kaggle website. The data was obtained through a survey collected from participants of various gender and age who had been to at least ten countries. There are four levels of severity based on the COVID-19 symptom, which was developed in accordance to World Health Organization (WHO) and the Indian Ministry of Health and Family Welfare recommendations. This paper presented an inquiry on the dataset utilising supervised and unsupervised machine learning approaches in order to better comprehend the dataset. In this study, the analysis of the severity group based on the COVID-19 symptoms using supervised learning techniques employed a total of seven classifiers, namely the K-NN, Linear SVM, Naive Bayes, Decision Tree (J48), Ada Boost, Bagging, and Stacking. For the unsupervised learning techniques, the clustering algorithm utilized in this work are Simple K-Means and Expectation-Maximization. From the result obtained from both supervised and unsupervised learning techniques, we observed that the result analysis yielded relatively poor classification and clustering results. The findings for the dataset analysed in this study do not appear to be providing the correct result for the symptoms categorized against the severity level which raises concerns about the validity and reliability of the dataset.

Development of an IoT Embedded Wearable Device with Non-Contact Temperature Detector for Early Detection of Fever

Nayli Nabila Azman — 2023-07-31

One of the most common and early symptoms of any viral infections is fever which is the reaction to a disease-specific stimulus causing the increase of human body temperature. The current common method of monitoring the human body temperature uses the application of non-contact infrared thermometer (NCIT) and is only limited for stationary conditions within short distances and mostly suitable for indoor premises. The available technology to detect human body temperature for longer distances uses the thermal camera which is costly and large. Thus, it is challenging to detect anyone with high body temperature in non-stationary conditions, at longer distances, especially outdoor. The paper proposes an innovation to the current practice, for a wearable non-contact temperature detector device which is portable. The wearable non-contact temperature detector embeds a thermal sensor and a microcontroller to a normal hat. It is able to detect objects with higher temperature (37.5 °C) within 1 meter radius of 60° angle view in stationary and non-stationary conditions. The wearable device communicates via Bluetooth to a mobile device to display the detected temperature and notifies the user via alert message and alarm for high temperature detection. Display of the object’s thermal image is also available with a resolution of 8 8 pixel. The wearable non-contact temperature detector is able to achieve 99% accuracy of temperature measurement for detection distance of up to 70 cm for indoor and within 20 cm for outdoor when tested with normal temperature subject and high temperature objects and compared with the actual temperature detected via a commercial NCIT device.

The Knowledge and Use of Speech Therapy Mobile Applications: Speech-Language Pathologists’ Perspectives in Malaysia

Afiqah Iylia Kamaruddin — 2023-07-31

Technology incorporation in speech therapy has been growing over the years. Mobile applications are among the adoptions that facilitate delivering speech therapy services. The situation in Malaysia is discouraging because there are not enough speech-language pathologists (SLPs) to serve the growing number of populations. Despite the abundance of available speech therapy mobile applications in the market, there is a lack of information focusing on the SLP’s knowledge and usage perspectives, especially in Malaysia. The objectives of this study are to describe the knowledge and usage perspectives of speech therapy mobile applications among SLPs in Malaysia and to analyze the instructional features and functional features relationships within the perspectives of SLPs. Surveys are established in three parts, with demographic questions in Part A, Likert scale responses for statements in Part B, and open-ended questions in Part C. This study is co-designed to relate to the results from an initial study that adopted PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and features analysis. The data from the initial study includes a review of 161 apps out of 1797 that have been identified. Five instructional features and nine functional features are presented. There are 35 SLPs participating in the survey. Their responses demonstrate evidence of SLPs’ knowledge and usage of speech therapy mobile applications. We will propose a conceptual framework for the features of speech therapy mobile applications, using people with aphasia as a point of reference for users with speech and language disorders.

Relationship Between Various Cement Mixture, Cement Fixation and Gait Study for Total Hip Replacement Via Finite Element Analysis (FEA)

Solehuddin Shuib — 2023-07-31

To secure the total hip replacement (THR) components, introduced in the 1960s, polymethyl methacrylate (PMMA) bone cement was used as a fixation. The cement polymerizes and becomes firm to hold the implant in place. However, the failure of cement in total hip replacement may lead to hip fractures and dislocations which is detrimental to the patient’s well-being whether in the short-term or long-term. Hence, the aim of this study is to find suitable cement mixtures for total hip replacement compromising of Young Modulus of 2.24 GPa, 0.3129 GPa, 0.03394 GPa and 0.07961 GPa, as reported from prior research. Three separate sorts of proximal cemented techniques were used to deposit the PMMA cement: 40 mm cement reduction, 80 mm cement reduction and full cement (datum). The Titanium Ti-6A1-4V (Ti-41) Charnley hip implant stem model with a Young Modulus of 100 GPa and a Poisson’s ratio of 0.3 was applied in the ANSYS Workbench 2020 R2 software to be analyzed with the three different proximal cemented approaches for each cement mixtures. Subsequently, the total deformation and von Mises stress were simulated under various loading circumstances, including standing, walking, stair climbing and falling. Nevertheless, as shown in the results obtained, all the hip implants consider safe because their von Mises stress does not exceed the yield strength of Titanium Ti-6A1-4V, which is 0.88 GPa. Finally, it may be concluded that, in comparison to the full cement (datum) and 80 mm cement reduction with Young Modulus of 2.24 GPa, 0.3129 GPa, 0.03394 GPa and 0.07961 GPa, the most improvement in the context of total deformation and von Mises stress is the 40 mm cement reduction with Young Modulus of 2.24 GPa.

Investigation of ConViT on COVID-19 Lung Image Classification and the Effects of Image Resolution and Number of Attention Heads

Pun Liang Thon — 2023-07-31

COVID-19 has been one of the popular foci in the research community since its first outbreak in China, 2019. Radiological patterns such as ground glass opacity (GGO) and consolidations are often found in CT scan images of moderate to severe COVID-19 patients. Therefore, a deep learning model can be trained to distinguish COVID-19 patients using their CT scan images. Convolutional Neural Networks (CNNs) has been a popular choice for this type of classification task. Another potential method is the use of vision transformer with convolution, resulting in Convolutional Vision Transformer (ConViT), to possibly produce on par performance using less computational resources. In this study, ConViT is applied to diagnose COVID-19 cases from lung CT scan images. Particularly, we investigated the relationship of the input image pixel resolutions and the number of attention heads used in ConViT and their effects on the model’s performance. Specifically, we used 512x512, 224x224 and 128x128 pixels resolution to train the model with 4 (tiny), 9 (small) and 16 (base) number of attention heads used. An open access dataset consisting of 2282 COVID-19 CT images and 9776 Normal CT images from Iran is used in this study. By using 128x128 image pixels resolution, training using 16 attention heads, the ConViT model has achieved an accuracy of 98.01%, sensitivity of 90.83%, specificity of 99.69%, positive predictive value (PPV) of 95.58%, negative predictive value (NPV) of 97.89% and F1-score of 94.55%. The model has also achieved improved performance over other recent studies that used the same dataset. In conclusion, this study has shown that the ConViT model can play a meaningful role to complement RT-PCR test on COVID-19 close contacts and patients.

COVID-19 Confirmed Cases Forecasting in Malaysia Using Linear Regression and Holt's Winter Algorithm

Hudzaifah Hasri — 2023-07-31

The 2019 coronavirus disease pandemic (COVID-19) has emerged and is spreading rapidly over the world. Therefore, it may be highly significant to have the general population tested for COVID-19. There has been a rapid surge in the use of machine learning to combat COVID-19 in the past few years, owing to its ability to scale up quickly, its higher processing power, and the fact that it is more trustworthy than people in certain medical tasks. In this study, we compared between two different models: the Holt’s Winter (HW) model and the Linear Regression (LR) model. To obtain the data set of COVID-19, we accessed the website of the Malaysian Ministry of Health. From January 24^th, 2020, through July 31^st, 2021, daily confirmed instances were documented and saved in Microsoft Excel. Case forecasts for the next 14 days were generated in the Waikato Environment for Knowledge Analysis (WEKA), and the accuracy of the forecasting models was measured by means of the Mean Absolute Percentage Error (MAPE). According to the lowest value of performance indicators, the best model is picked. The results of the comparison demonstrate that Holt's Winter showed better forecasting outcome than the Linear Regression model. The obtained result depicted the forecasted model can be further analyzed for the purpose of COVID-19 preparation and control.

Subcutaneous Vein Recognition System Using Deep Learning for Intravenous (IV) Access Procedure

Chan Xiao Jing — 2023-07-31

Intravenous (IV) access is an important daily clinical procedure that delivers fluids or medication into a patient’s vein. However, IV insertion is very challenging where clinicians are suffering in locating the subcutaneous vein due to patients’ physiological factors such as hairy forearm and thick dermis fat, and also medical staff’s level of fatigue. To resolve this issue, researchers have proposed autonomous machines to be used for IV access, but such equipment are lacking capability in detecting the vein accurately. Therefore, this project proposes an automatic vein detection algorithm using deep learning for IV access purpose. U-Net, a fully connected network (FCN) architecture is employed in this project due to its capability in detecting the near-infrared (NIR) subcutaneous vein. Data augmentation is applied to increase the dataset size and reduce the bias from overfitting. The original U-Net architecture is optimized by replacing up-sampling with transpose convolution as well as the additional implementation of batch normalization besides reducing the number of layers to diminish the risk of overfitting. After fine-tuning and retraining the hypermodel, an unsupervised dataset is used to evaluate the hypermodel by selecting 10 checkpoints for each forearm image and comparing the checkpoints on predicted outputs to determine true positive vein pixels. The proposed lightweight U-Net has achieved slightly lower accuracy (0.8871) than the original U-Net architecture. Even so, the sensitivity, specificity, and precision are greatly improved by achieving 0.7806, 0.9935, and 0.9918 respectively. This result indicates that the proposed algorithm can be applied into the venipuncture machine to accurately locate the subcutaneous vein for intravenous (IV) procedures.

A Study on the Correlation Between Hand Grip and Age Using Statistical and Machine Learning Analysis

Sahnius Usman — 2023-07-31

Handgrip strength (HGS) is an easy-to-use instrument for monitoring people's health status. Numerous researchers in many countries have done a study on handgrip disease or demographic data. This study focused on classifying aged groups referring to handgrip value using machine learning. A total of fifty-four participants had involved in this study, ages ranging from 24 years to 57 years old. Digital Pinch Grip Analyzer had been used to measure the handgrip measurement three times to get more accurate results. The result is then recorded by Clinical Analysis Software (CAS) that is built into the analyzer. An independent t-test is used to investigate the significant factor for age group classification. The data were then classified using machine learning analysis which are Support Vector Machine (SVM), Random Forest (RF), and Naïve Bayes. The overall dataset shows that the Support Vector Machine is the most suitable classification technique with average accuracy between 5 groups of age is 98%, specificity of 0.79, the sensitivity of 0.9814 and 0.0185 of mean absolute error. SVM also give the lowest mean absolute error compared to RF and Naïve Bayes. This study is consistent with the previous work that there is a relationship between handgrip and age.

Identification of Risk Factors for Scoliosis in Elementary School Children Using Machine Learning

Ahmad Aizat Che Rahmat — 2023-07-31

Scoliosis is an abnormal curvature of the spine and often diagnosed in childhood or early adolescence. In this study, the risk factors for scoliosis in elementary school children is investigate based on age, backpack weight and gender. There are 260 children participated in this study from aged 7 up to 12 years old. Scoliometer is used to measure the angle of trunk rotation (ATR) on Adam Forward Bending Test. Statistical analysis of analysis of variance (ANOVA) is used to determine the characteristic difference of ATR readings on the risk factors for scoliosis. Significant results with P-value less than 0.001 are found among ATR readings on a linear combination of risk factors for scoliosis of age and backpack weight. Then, the risk factors for scoliosis are classified among elementary school children using Decision Tree and K-Nearest Neighbor. The classification results shown that both Decision Tree method produced highest classification percentage up to 98.08%. This finding indicates that age and backpack weight are significant as the risk factors for scoliosis.