International Journal of Autonomous Robotics and Intelligent Systems (IJARIS)

Design and Development of a Smart Portable Water Purification System

Abdulkareem Bageri — 2025-08-05

Access to clean drinking water remains a critical challenge for individuals in remote, off-grid, or temporary living situations such as travelers, campers, and emergency responders. This project presents the design and development of a Smart Portable Water Purification System that addresses this need through a compact, solar-powered, and IoT-integrated solution. The system utilizes a multi-stage filtration process including sediment, pre-carbon, reverse osmosis, post-carbon, and silver carbon filters to effectively remove physical impurities, chemical contaminants, and biological hazards. Real-time water quality monitoring is achieved using TDS and pH sensors, while a Raspberry Pi Pico W microcontroller enables wireless control and monitoring via a custom-built graphical user interface accessible through a local Wi-Fi access point. Performance evaluations showed a significant improvement in water quality, with pH values normalized to safe ranges. Despite low to moderate filtration speeds and unoptimized data handling, the results validate the system's effectiveness and usability in field conditions. Recommendations for future enhancements include improving power efficiency, filtration speed, communication protocols, sensor integration, and mechanical design for better portability and environmental durability. The final prototype demonstrates a reliable, user-friendly, and sustainable approach to decentralized water purification.

Implementation of Music Emotion Classification using Deep Learning

Qing Xiang Sow — 2025-08-05

Music plays a crucial role in shaping emotions and experiences, making its classification an important area of research with applications in therapy, recommendation systems, and affective computing. This study develops a deep learning-based system to classify music into three emotional categories: "Angry," "Happy," and "Sad." The dataset, consisting of 22 audio files collected from YouTube, was manually labelled, segmented into 30-second clips, and augmented using pitch shifting and time stretching to enhance diversity. Features were extracted using Mel-Frequency Cepstral Coefficients (MFCC) and spectral contrast to analyse the harmonic and timbral characteristics of the audio. Three deep learning models, CNN, CNN-LSTM, and CNN-GRU, were evaluated. CNN-GRU achieved the highest weighted accuracy of 99.10%, demonstrating superior performance. Future work includes adding more emotion categories, diversifying the dataset, exploring advanced architectures like transformers, optimising hyperparameters, implementing real-time applications, and conducting user studies to assess effectiveness. This research successfully developed and evaluated a music emotion classification system, contributing to advancements in the field.

Design of an Autonomous Pipe-Cleaning Robot for Small-Scale Hydroponic Systems

Muhammad Nazrin Shah Bin Shahrol Aman — 2025-08-05

Hydroponic farming offers a sustainable solution to food insecurity, especially in developing countries, by enabling high-yield crop production with minimal land and water usage. However, one major challenge in hydroponic systems is the frequent clogging of PVC pipes due to debris from growing media and the buildup of algae, which thrives under high humidity and limited sunlight. These blockages not only reduce system efficiency but also increase the risk of plant diseases that can compromise entire harvests. To address this issue, this research focuses on the design and development of a compact mobile robot capable of cleaning hydroponic pipes. The robot is designed using SolidWorks and simulated in Tinkercad, targeting a 4-inch (110 mm) diameter PVC pipe with a length of 3 feet. It incorporates the HC-SR04 ultrasonic sensor to detect obstacles and uses acrylic as the structural material. The performance of the robot was tested against three types of common pipe obstructions: small stones, sand and grass roots. Experimental results show that the robot can navigate through the hydroponic pipe and partially remove debris, although complete cleaning remains challenging due to space constraints and the complexity of internal pipe conditions. This work demonstrates a low-cost, automated solution for maintaining hydroponic systems and highlights areas for further optimization in design and control.

Performance Evaluation of Human Facial Expression using Various Classification Methods

Ts. Dr. Abdul Halim Ismail — 2025-08-05

This study assesses and contrasts the efficacy of raw picture pixels and image vectors as features in face expression classification. The CKPLUS dataset is utilized, and the issue of class imbalance is tackled by data augmentation. The dataset is partitioned into a 70% training set and 30% validation set. The training set consists of 175 images for each class, while the validation set consists of 75 images. The features are displayed using Matplotlib for raw pixels and t-SNE for vector features, then categorized using Random Forest and CNN classifiers. The performance is evaluated by utilizing confusion matrices, accuracy, precision, recall, and F1-score. The findings indicate that the Random Forest algorithm, when combined with vector features, obtains the maximum level of accuracy (99.6190%). Additionally, CNNs using raw pixel features also demonstrate strong performance. The precision, recall, and F1-scores exhibit similarity among the different approaches, with Random Forest (vector feature) and 2D CNN (raw pixels) showing somewhat better performance compared to other methods. These findings suggest that vector features have superior performance when used in conjunction with Random Forest, whereas raw pixel features are more successful when utilized with CNN.

LIDAR-based Robot for Localization and Mapping

Shazmin Aniza — 2025-08-05

This paper presents an application of LIDAR sensor on a robot for 2D mapping construction of an environment and its capability to localize its own location. RP LIDAR is used and computational processing is done by using Robot Operating System (ROS). Mapping of the environment is done and with addition of localization of the robot with respect to static landmarks were also conducted. From the results, it can be seen that map of the environment is able to be reconstructed and several locations with respect to marking points are able to be identified with an average of 97% accuracy.

Development of a Floating Beacon for Real-Time Water Quality Measurement

Hassrizal — 2025-08-05

Water quality is a critical determinant of agricultural productivity and environmental sustainability. This study addresses the challenge of real-time water quality monitoring in agricultural settings, specifically focusing on paddy fields in Perlis, Malaysia. The Jabatan Pengairan dan Saliran of Perlis (JPS) currently faces difficulties in continuously monitoring water quality parameters after water discharge from the Timah Tasoh Dam into the irrigation system. To overcome this, this research focuses on developing a Floating Beacon (FB), an Internet of Things (IoT)-based system for autonomous water quality assessment. The system integrates sensors to measure key water quality parameters, including pH, turbidity and temperature. Real-time data collected by these sensors are wirelessly transmitted to the Blynk cloud platform, enabling continuous data recording and visualization on a customizable dashboard. Furthermore, the system incorporates a feature for tracking the geographical location of the monitoring devices, providing crucial context for water quality measurements across the agricultural landscape. This research is anticipated to significantly enhance the efficiency of water quality management for the JPS of Perlis, thereby supporting optimal water supply to paddy fields and contributing to improved agricultural practices. Based on the result, the FB is capable provides real-time water quality data and display on the web and mobile devices using Blynk 2.0. Furthermore, the FB shows high accuracy for the location measurement with the maximum error at 0.000743% for latitude and 0.000037% for longitude.

Machine Learning Techniques in Credit Card Fraud Detection: A Hybrid Supervised and Unsupervised Approach

Li Phng Yeoh — 2025-08-05

In the dynamic landscape of financial transactions, the escalating threat of fraudulent activities necessitates cutting-edge solutions for real-time detection. This research introduces an innovative approach utilizing the Kaggle credit card dataset, focusing on comparing the effectiveness of hybrid models versus purely supervised learning models. While traditional models rely solely on supervised learning, this study explores the potential performance gains of integrating unsupervised learning (USL) into supervised learning (SL) frameworks. The core investigation centers on whether unsupervised clustering can enhance pattern recognition in unlabeled data and subsequently improve the performance of supervised models. This research not only evaluates the practical benefits of hybrid methodologies in fraud detection, but also advances real-time analytics through Power BI, aiming to provide a more comprehensive and adaptive solution to emerging financial threats. The algorithms yield an accuracy of 99.75% and a remarkably low underkill rate of 0.20%, demonstrating the effectiveness of integrating human oversight with advanced machine learning techniques.

Development of an Intelligent Sumo Robot based on Embedded Fuzzy Logic

Mohd Shakirin Mohd Shukri — 2025-08-05

This paper presents the design and development of an intelligent sumo robot utilizing embedded fuzzy logic for real-time opponent detection and motion control. The system integrates analog infrared (IR) distance sensors and a fuzzy inference engine deployed on an Arduino Mega microcontroller to improve the robot’s responsiveness and adaptability in dynamic environments. Through sensor calibration and optimized placement, the robot accurately detects opponent positions across multiple zones. A fuzzy rule base interprets the continuous sensor inputs and generates smooth, context-aware motor responses. The system was evaluated through simulation in MATLAB and embedded implementation, with test scenarios involving five fixed opponent positions and gradually varying inputs. Comparative results show that the fuzzy controller significantly outperforms traditional if-else logic, producing smoother motion, better coverage, and more reliable decision-making, with error values between MATLAB and Arduino outputs as low as 0.22%. The embedded fuzzy logic system closely replicates simulated behavior, validating its real-time feasibility on low-cost hardware. This approach demonstrates a scalable and effective solution for intelligent mobile robotics.

Real-Time Classification of Chilli Ripeness using Convolutional Neural Network (CNN)

Nurul Syahirah Binti Khalid — 2025-08-05

Chilli harvesting plays an important role in Malaysia’s economy as it is one of the crops with high demand in the country. Normally, farmers harvest and categorise the ripeness of chillies by using the naked eye which can lead to errors and human fatigue. To overcome the limitations of this manual harvesting, an automated real-time chilli vision system that can classify between ripe and unripe chillies were developed. This research involved with a diverse dataset of chilli images using various chilli varieties and growth stages. The YOLOv8 model was trained using Google Colab's GPU-accelerated environment to optimize the performance. The model's deployment for real-time inference and classification was facilitated through Visual Studio Code, with HSV colour analysis used to differentiate between ripe and unripe chillies. CNN was used to validate and analyse the accuracy of the proposed system. As a result, the system achieved an accuracy of 88% for chilli classification. These findings proved the potential of Artificial Intelligence (AI)-driven systems in supporting precision agriculture.

Development of 3-Phase Power Measurement Instrument

Mohd Nasir Ayob — 2025-08-05

This paper presents a real-time 3-phase power measurement instrument development using the ESP32 microcontroller and analog sensors. The instrument measures key electrical parameters including RMS voltage and current, real power, apparent power, reactive power and power factor for each phase of a 3-phase system. Key components used include the ZMPT101B voltage sensor, ACS712 current sensor, a 20x4 I2C LCD display and push buttons for manual data navigation. The system implements time-domain sampling and cross-correlation techniques for accurate phase angle and power factor calculations. The performance evaluation demonstrates its accuracy suitable for educational applications, as the measurement error compared to reference instruments was less than 2%.