Research Publications
Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4194
This main community comprises five sub-communities, each representing the academic contribution made by SLIIT-affiliated personnel.
Browse
34 results
Search Results
Item Embargo AgriSense LK: Weekly Automated Machine Learning for Sri Lankan Produce Prices with Business Continuity Plan, Market Opportunity Ranking, Cultivation Targeting, and Yield Quality Valuation(Institute of Electrical and Electronics Engineers Inc., 2026-05-22) Matharaarachchi, Charaka J.; Samarasinghe, Ravindu T; Vidyasarani G.G.T.; Fasnas, M; Siriwardana, D; Wijesooriya, AIn Sri Lanka, agricultural decision-making remains largely traditional: decisions are often based on historical practices, informal consultation, and heuristic judgment. The primary barrier is that market price data is difficult to interpret without analytical expertise, resulting in unpredictable price volatility and suboptimal farmer income. AgriSense LK is a machine learning platform that converts historical price records into actionable recommendations for farmers and traders. The system comprises four components: business strategy classification, market opportunity ranking, cultivation targeting, and smartphone-based produce quality grading. The platform was trained on 123,985 real price records sourced from the Central Bank of Sri Lanka (CBSL), spanning 2017 to 2025. Key results include a MAPE of 0.7% and MAE of Rs. 1.86 on weekly price forecasting (a 98.1% improvement over the naive baseline), a ROC-AUC of 0.9056 on cultivation targeting, and 91.49% crop classification accuracy with 89.84% quality grade accuracy in the computer vision component. Direct price regression over a seven-day horizon proved unreliable; a binary profitability classifier was adopted instead and substantially outperformed the regression approach. While results are promising, further validation under real-world deployment conditions is required.Publication Open Access Development of a MEMS-based Earthquake Dataset using the Raspberry Shake Network in New Zealand(Sri Lanka Institute of Information Technology, 2026-05-21) Samiha, T.Z; Ravishan, DTraditional seismic monitoring is often limited by the high cost of instrumentation and logistical barriers, hindering the expansion of earthquake monitoring networks especially in under-resourced regions. Low-cost MEMS-based sensors offer a scalable alternative, but require specialized datasets to train machine learning models adapted to their unique noise characteristics and sensitivity profiles. To address this, we systematically collected waveforms from approximately 4,000 earthquakes (magnitude 2.7 to the highest recorded) recorded across 89 Raspberry Shake stations in New Zealand from 2020–2025. Events were matched to nearby stations based on epicentral distance criteria (100 km for M 2.7–5.5, 150 km for M>5.5). A staged filtering pipeline using PhaseNet, EQTransformer, and GPD models, cross-validated with theoretical TauP arrivals, was applied to ensure phase pick quality across three confidence tiers. The final curated dataset comprises 918 high-confidence waveforms with validated P and S wave arrivals, alongside approximately 16,035 total waveform records spanning all quality tiers. This dataset addresses the critical scarcity of labeled training data for low-cost seismic instrumentation, enabling the development of phase pickers specifically calibrated for MEMS sensors.Item Embargo HireGenius: Automated Interviewing System for Software Engineers(Springer Science and Business Media Deutschland GmbH, 2026-08-01) Hewamadduma N.A.; Nalinka G.K; Mahawaththa N.T.M.A.S.M; Rosa S.R.T.L; De Silva D.I.; Gunathilake P.Recruiting the right software engineers is a critical challenge, with traditional manual screening being time-consuming, subjective, and often inconsistent. Recruiters typically rely on Curriculum vitae reviews and interviews, which lack the depth needed for evaluating technical roles. For software engineers, it is essential to assess programming skills, academic performance, and personality traits. To overcome these limitations, this study developed an automated candidate selection and interview system using artificial intelligence, natural language processing, and deep learning. Ensemble learning and artificial intelligence models incorporating natural language processing were used to rank candidates and predict job match percentages. Top-ranked individuals were further evaluated through analysis of GitHub profiles, LinkedIn activity, and academic transcripts using machine learning and natural language processing techniques. Each candidate’s technical skills, experience, and education were assessed to generate accurate shortlists for technical interviews. These shortlisted candidates then participated in an automated interview process powered by advanced natural language processing and deep learning. A gamified human resource interview system was introduced, leveraging a machine learning model and structured scoring criteria to identify the best-fit candidates while streamlining and enhancing the hiring process.Item Embargo AloeGreen: Smart IOT-Based System for Aloe Vera(Institute of Electrical and Electronics Engineers Inc., 2026-05-21) Megasooriya, E; Rajapaksha, H; Rajapaksha, H; Bandara, A; Krishara, J; Wijendra, DAgriculture plays a vital role in food security, yet Aloe vera cultivation remains vulnerable to environmental variability, nutrient imbalance, disease occurrence, and unstable market conditions. This study presented AloeGreen, a crop-specific AI-IoT smart agriculture framework designed to support Aloe vera cultivation through integrated sensing, forecasting, and decision-support modules. The system combined real-time IoT-based field monitoring with machine learning models for yield prediction, environmental forecasting, disease detection, fertilizer recommendation, and price forecasting. A key contribution of the study was a forecast-informed yield prediction strategy in which short-term environmental forecasts were incorporated into the yield estimation pipeline to support future-aware decision-making. In addition, domain-specific agronomic features, including water stress and heat stress indices, were introduced to better represent Aloe vera growth conditions. For the yield prediction module, the cleaned hourly cultivation dataset contained 1,048,330 observations after removing missing critical fields and duplicates. Experimental results showed that XGBoost achieved the best yield prediction performance with an RMSE of $\mathbf{1 0. 0 2}$ and an $\mathbf{R}^{\mathbf{2}}$ of $\mathbf{0. 8 9 2}$, while the environmental forecasting module achieved strong performance for temperature and humidity prediction, although rainfall prediction remained comparatively weaker. The disease detection module achieved balanced classification performance of approximately 77% accuracy, and Random Forest performed best in both price forecasting and fertilizer recommendation tasks. Overall, the findings showed that integrating IoT sensing with intelligent analytics in a unified Aloe vera cultivation platform can improve decision support, reduce uncertainty, and contribute to more sustainable smart agriculture practices.Publication Open Access An Intelligent Risk Aware Navigation Framework for Accident Hotspot Prediction, Real-Time Traffic Analysis, and Safety-Oriented Route Planning(Sri Lanka Institute of Information Technology, 2026-05-21) Rathnapala, A; Seneviratne, ORoad traffic accidents remain a major public safety issue, particularly in urban regions where increasing traffic density and complex road environments contribute to higher accident risk. Existing navigation systems primarily optimize routes based on travel time or distance, without considering accident risk, which can expose drivers to unsafe road segments. This study proposes an intelligent risk-aware navigation framework that integrates accident hotspot prediction with real-time traffic analysis for safety-oriented route planning. A supervised machine learning model was trained using historical accident records obtained from Sri Lanka Police data to estimate accident risk levels across road segments. These risk scores are combined with real-time traffic information retrieved from mapping services to evaluate alternative routes based on both safety and travel efficiency. Experimental results show that the proposed model achieves an accuracy of 0.93 in predicting accident risk levels. Furthermore, the system is able to recommend routes that reduce exposure to accident-prone areas while maintaining acceptable travel time. A mobile prototype was developed to visualize accident hotspots and provide safer route commendations. The results demonstrate that integrating predictive accident analytics with real-time traffic information can significantly enhance navigation systems by enabling safety-aware decision-making and improving overall road safetyPublication Open Access Smart-Camp Box: An Integrated IoT and Artificial Intelligence Framework for Safety, Communication, Learning Assistance, and Resource Optimization in Disaster Relief Camps(Sri Lanka Institute of Information Technology, 2026-05-21) Kashmira S S; Hansana G P; Deerasinghe A D S N S; Jayakody, A; Lokuliyana, SNatural disasters continue to inflict devastating consequences on communities across South Asia, with Sri Lanka ranking among the most frequently affected nations in the region. When disasters strike, temporary relief camps serve as critical shelters for displaced populations; however, existing systems fail to address three persistent operational challenges simultaneously: camp-level flood and landslide prediction, psychological and attentional readiness assessment for displaced children, and resilient communication under degraded network conditions. This paper presents the Smart Camp Box, an integrated, portable IoT-based framework that addresses all three dimensions through tightly coupled sub-systems. The first sub-system provides location-aware environmental risk monitoring using GPS/GNSS, IoT sensors, DEM integration, and a machine learning prediction model. The second introduces the Psycho-Attentional Gated Educational System (PAGES), an offline-first application for assessing and supporting displaced children's cognitive readiness. The third proposes a Semantic-Aware Adaptive Message Prioritization (SAAMP) framework for reliable MQTT communication under constrained network environments. Currently in active development, the Smart Camp Box represents a paradigm shift toward proacPublication Open Access Development Of An Ai-Based Model With Low Computational Complexity For Accurate Solar Energy Forecasting(Faculty of Engineering, 2025-09-09) Chandrasinghe, S; Fernando, NThis paper introduces a short-term solar energy forecasting model that is designed with a focus on low computational complexity and addresses the challenges posed by fluctuations in solar energy generation, which are significantly influenced by environmental factors. These fluctuations can lead to instability when solar power generation systems are integrated into national energy grids, creating difficulties in maintaining a balanced supply and demand. If solar energy generation can be accurately forecasted before fluctuations occur, potential issues can be identified in advance, allowing for better management of the energy system, including optimizing storage facilities when energy generation is high. Current solar energy forecasting systems face significant challenges due to their high computational complexity, which results in increased power consumption and lower accuracy. To address these issues, this study focuses on the development of an artificial intelligence (AI)-based forecasting model using an Artificial Neural Network (ANN). The goal is to reduce the computational complexity of the model while maintaining high accuracy. To achieve this, various data analysis and complexity reduction techniques, such as variable reduction, pruning, and quantization, were applied. The performance of the optimized AI model was evaluated by comparing the forecasted values to actual solar energy generation data. The results demonstrate that the proposed model successfully reduces computational complexity while maintaining a satisfactory level of accuracy. This optimization makes the model more suitable for real-time forecasting, particularly in resource-constrained environments, and provides a more efficient approach to solar energy management. The findings of this study suggest that AI-based forecasting models can play a critical role in enhancing the integration of solar energy into national grids, ensuring a more reliable and sustainable energy supply. Further research could explore additional optimization techniques and the introduction of generalization techniques to improve transferability of the model and applicability across diverse geographical regions. Additionally, focus on utilizing AI techniques that minimize computational complexity without compromising the accuracy of the model, aiming to maintain high forecasting precision while optimizing the efficiency of the system.Item Embargo Bovitrack:Animal behavior monitoring using Machine learning and IoT(Institute of Electrical and Electronics Engineers Inc., 2025) Viraj, H; Wijesekara, S; Tharuka, K; Fernando, S; Jayakody, A; Wijesiri, PAnalyzing dairy cattle behavior and anomalies is a critical component of precision livestock farming, allowing farmers to remotely monitor animals for health and behavior. In order to accomplish this task better, the use of IoT technology and machine learning algorithms is more appropriate as per the time. The YOLO (you only look once) object recognition algorithm is more suitable for that, and the use of this algorithm allows these processes to be performed automatically and in real time with high accuracy. YOLO's ability to recognize multiple objects in images or videos makes Yolo ideal for cattle detection and tracking.Item Embargo Hybrid Model-Based Automated Exterior Vehicle Damage Assessment and Severity Estimation for Insurance Operations(Institute of Electrical and Electronics Engineers Inc., 2025) Jayagoda, N.M; Kasthurirathna, DAfter a vehicle accident, insurance companies face the critical task of assessing the damage sustained by the involved vehicles, a process essential for maintaining the insurer's credibility, building consumer trust, and meeting legal and ethical obligations. This assessment is crucial for ensuring clients' financial protection and proper compensation, upholding the integrity of the insurance process. Traditionally, evaluations have been conducted through manual inspections by experienced professionals who meticulously document vehicle damage. Despite its thoroughness, this approach suffers from significant inefficiencies, high costs, and extended time requirements. Moreover, the method is vulnerable to human errors and subjective biases, which can result in inflated valuations. To overcome these challenges, this research introduces an innovative system designed to leverage technology for analyzing images of damaged vehicles uploaded by the user. This system aims to accurately identify the damaged external components, assess the severity of the damage, and determine the repair needs based on the compromised sections of the vehicle. The findings reveal that the hybrid model used in this research is capable of determining vehicle damage severity with an overall accuracy of 73.3%. This level of accuracy demonstrates the model's robust capability to effectively navigate and analyze complex damage patterns, underscoring its practical applications. By accurately determining damage levels on the first assessment, the model reduces the need for further assessments and disagreements, which frequently cause claim delays. This enhancement increases productivity, reduces administrative costs, and improves the customer experience, resulting in a more efficient, transparent, and satisfactory resolution of vehicle insurance claims.Item Embargo MindBridge: Early Identification of Learning Difficulties in Children as a Supporting Tool for Teachers(Institute of Electrical and Electronics Engineers Inc., 2025) Mapa, N; Deshapriya, M; Premathilake, M; Samarakoon, S; Thelijjagoda, S; Vidanaralage, A.JLearning difficulties in children significantly impede academic success by affecting information processing, mathematical performance, and the learning of proper reading and writing. This paper proposes a Progressive Web Application (PWA) based on artificial intelligence (AI) and machine learning (ML) for identifying potential learning barriers. In contrast with standard diagnostic instruments, the proposed system is designed as a prediction tool with the potential for teachers to conduct timely and focused interventions. By automating feature extraction and reducing manual processing, the system overcomes the limitations of existing learning systems and improves early detection accuracy. Preliminary evaluations indicate that the PWA can effectively identify at-risk students and improve intervention methods and overall academic performance. This research contributes to the integration of computational methods and pedagogy, offering a scalable and low-cost solution for helping slow learners overcome their learning challenges.