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
2598 results
Search Results
Item Embargo An Integrated Approach to Smart Criminal Judgment Analysis in Sri Lankan Courts(Institute of Electrical and Electronics Engineers, 2026-05-29) Sundaresan, K; Ekanayake, D; Thavachchelvam, N; Sivaanbu, A; Abeywardhana, L; Nawarathne, MCriminal justice practitioners in Sri Lanka face considerable challenges when accessing and analyzing past court judgments for case preparation, as most research tasks remain manual and time intensive. This study presents an integrated system comprising four interconnected components to address these challenges. A Legal Resource Extractor uses a Hybrid Neuro-Symbolic Architecture with domain-adapted Legal-BERT, fine-tuned on 133,338 legal text chunks, to extract structured legal knowledge from multilingual inputs including voice evidence submitted by lawyers. A Case Analysis and Argument Generation module employs LegalBERT embeddings over 1,601 Court of Appeal judgments with a Nearest Neighbors retrieval model achieving 94% perfect retrieval rate for precedent-based argument generation. The Appeal Outcome Prediction component has been optimized through a Hybrid Feature Engineering pipeline that integrates 1,000 TF-IDF lexical features, 49 traditional legaldomain indicators, and 768-dimensional Legal-BERT embeddings. By utilizing SMOTE and a Calibrated Voting Ensemble, the system predicts appeal outcomes with 79.75% accuracy across three outcome classes. A Public Legal Assistant system provides offline legal information in Sinhala, Tamil, and English through hybrid FAISS based retrieval using 1,146 legal documents and locally hosted language model generation. Together, these components offer practitioners an end-to-end platform for strengthening legal research, argument preparation, and access to justice.Item Embargo A Reinforcement Learning Approach with Human in the Loop to Explainable Insurance Risk Scoring and Intelligent Policy Portfolio Optimization(Institute of Electrical and Electronics Engineers, 2026-05-29) Gamage, C; Kasthuriarachchi, T; Denuwan, C; Mallawaarachchi, P; Abeywardhana, L; Nawarathne, MAssessing individual risk accurately and optimizing insurance portfolios in real time remain major challenges due to static actuarial tables, opaque models, and fragmented analytical pipelines. This paper proposes a two-part Explainable AI (XAI) framework addressing both issues. The first component, Artificial Intelligence-driven risk scoring with human-in-the-loop (HIL) weight adjustment, uses a Proximal Policy Optimization (PPO) agent to suggest feature-based changes to an insurer's risk-equation weights. Shapley Additive Explanations(SHAP) attributions and Generative AI reasoning make these changes interpretable, allowing human reviewers to approve modifications that are immediately applied to future customers, creating a self-improving loop. The second component, AI-driven policy optimization, leverages a PPO supported by an XGBoost expense regressor, SHAP/LIME explainability, PPO agent and a Retrieval-Augmented Generation (RAG) layer for rider assignment. Both components share a data backbone of 100,000 anonymized insurance records stored in MongoDB, enabling incremental updates without reprocessing. Experiments show the XGBoost regressor achieves Root Mean Square Error (RMSE) 0.4406 and Mean Absolute Error (MAE) 0.3600, the HIL guided agent increases average episodic reward by 10.3%, and the RAG layer reaches 91.7% rider-assignment accuracy. The framework significantly enhances predictive accuracy, interpretability, regulatory traceability, and portfolio adaptability compared to traditional actuarial and black-box approaches.Publication Embargo Spatial distribution and hydrological influences of microplastic contamination in the upstream reach of the ambathale water treatment plant intake, Sri Lanka(Springer Science and Business Media, 2026-08-07) Rathnayake, D; Miguntanna, N; Rathnayake, UMicroplastic contamination has emerged as a global environmental concern, with rivers serving as important pathways for the accumulation and transport of these particles. This study investigated the spatial distribution of microplastic contamination and the influence of hydrological and physicochemical factors on microplastic occurrence in the upstream reach of the Ambathale Water Treatment Plant intake along the Kelani River, Sri Lanka. Water samples were collected from four locations within the study area, including three upstream sites and one downstream site. Microplastics were detected at all sampling locations, with Kelaniya recorded the highest microplastic concentration (20.08 ± 10.19 mg/L), while Ambathale had the lowest (9.12 ± 3.89 mg/L). Three main microplastic types were identified: fibers, fragments, and films. Fibers and fragments were the most abundant microplastic types across all sampling locations, while films occurred at comparatively lower concentrations. Statistical analyses revealed a significant positive correlation between river water level and microplastic concentration (r = 0.883, p = 0.001), indicating the importance of hydrological conditions in microplastic transport. Fiber and film concentrations exhibited particularly strong positive relationships with water-level fluctuations (r = 0.940 and r = 0.904, respectively; p < 0.01). Furthermore, positive associations between microplastic abundance and suspended solids suggest that particulate matter plays an important role in the transport and distribution of microplastics within the river system. The findings demonstrate significant spatial variability in microplastic contamination and highlight the influence of hydrological processes on microplastic occurrence and transport in the Kelani River. This study provides baseline information for future monitoring and contributes to a better understanding of microplastic dynamics in tropical freshwater environments.Publication Open Access CALCOM: An integrated techno-economic and life-cycle environmental framework for electric vehicle assessment in Sri Lanka(Elsevier B.V., 2026-08-01) Abeygunawardena, N; Wijayapala, A; Jathunga, TElectric vehicle (EV) adoption is accelerating worldwide, creating a need for comprehensive frameworks that assess economic feasibility and environmental performance. This study presents a context-adaptive levelized cost of mileage (CALCOM) framework to evaluate the economic and environmental performance of battery electric vehicles (BEVs), hybrid electric vehicles (HEVs), and internal combustion engine vehicles (ICEVs) under Sri Lankan conditions. The framework integrates discounted life-cycle cost (LCC), net present value (NPV), and greenhouse gas (GHG) emissions into a unified assessment model. Real-world operational ad cost data were collected from owners representing Nissan Leaf (BEV), Toyota Aqua (HEV), and Toyota Vitz (ICEV). The analysis included purchase cost, energy consumption, maintenance, battery replacement, salvage value, and environmental costs over a 10-year ownership period. The BEV achieved the lowest levelized cost of mileage of 0.080 USD/km which further decreased to 0.050 USD/km under renewable charging. Life-cycle GHG emissions were 59% lower than those of ICEV. Sensitivity analysis identified electricity price, annual distance travelled, and charging efficiency as the primary determinants of BEV competitiveness. Threshold analysis indicated that BEVs remain economically attractive when domestic electricity tariffs are maintained below 0.32 USD/kWh. The findings demonstrate that BEVs offer the greatest economic and environmental benefits and offer evidence-based guidance for policies supporting electric mobility in developing countries.Publication Embargo Fusion of Lean Six Sigma-DMAIC Tools and Techniques into Construction Variation Management: Barriers and Strategies(American Society of Civil Engineers (ASCE), 2026-11-01) Shashini, T.W.K.N; Jayanetti, J.K.D.D.T; Disaratna, V; Ranadewa, T; Perera, B.A.K.SVariations in construction projects pose a significant challenge to achieving successful project outcomes. Lean Six Sigma has the potential to manage variations in the construction industry. This study aims to explore the fusion of lean Six Sigma (LSS)-Define, Measure, Analyze, Improve, and Control (DMAIC) tools and techniques into construction practice to manage variations. Adopting a pragmatic stance and a qualitative approach, empirical data were collected through a three-round Delphi expert survey using semistructured interviews, with the data analyzed through content analysis. The findings reveal that 25 LSS-DMAIC techniques and tools can be effectively applied to manage variations in the construction sector. In addition, 17 barriers and 21 strategies were identified, and a mapping was undertaken to align each barrier with suitable strategies. This research addresses a gap in existing literature by extending investigations of LSS-DMAIC in construction to the specific context of managing variations, identifying barriers to implementation, and proposing strategies, an area that has not been explicitly examined previously. By investigating the barriers to implementing LSS-DMAIC and by suggesting tailor-made strategies for overcoming each barrier, the study makes both theoretical and practical contributions. It provides a framework for on the applicability of LSS-DMAIC in developing countries while offering construction managers context-specific solutions to enhance project outcomes and improve stakeholder satisfaction.Publication Open Access Drivers of carbon emissions in G7 economies: Evidence on energy use, globalisation, urbanisation, industrialisation and innovation(Elsevier Ltd, 2026-08-26) Weerasinghe, L; Vithanage, N; Rupasinghe, D; Keesha, C; Jayathilaka, RRising CO2 emissions remain a major sustainability challenge, particularly in advanced economies that account for a considerable share of historical emissions. This study examines the key determinants of CO₂ emissions in G7 countries by jointly considering globalisation, energy consumption, urbanisation, industrialisation, and technological innovation. Using a balanced panel dataset for Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States over the period 1997–2023, the analysis applies panel regression techniques and country-specific multiple linear regressions to capture both common effects and heterogeneity across economies. The results indicate that energy consumption is a robust driver of CO₂ emissions across the G7. Technological innovation contributes to emission reduction at the panel level; however, its effect becomes statistically insignificant in country-specific estimations, reflecting cross-country differences in innovation structures and policy environments. Globalisation significantly increases emissions in Canada, while its influence is negligible elsewhere. Urbanisation shows a mitigating effect only in the United States, and industrialisation increases emissions in Canada and Italy but reduces emissions in Japan. Overall, the findings highlight that decarbonisation strategies in advanced economies should prioritise the transition to clean energy while strengthening innovation-oriented climate policies tailored to country-specific contexts.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 A Novel Sample-Splitting Receiver Architecture for SWIPT: Design and Resource Allocation(Institute of Electrical and Electronics Engineers, 2026-06-08) Vithanage, G. S; Jayakody, D.N.K; Muthuchidambaranathan P.; Dinis, RThis paper presents a novel sample splitting (SS) technique for simultaneous wireless information and power transfer (SWIPT), addressing the inefficiencies of conventional power splitting (PS) and time switching (TS) methods. Unlike traditional approaches, SS directly samples the received signal after it is captured by the antenna. Subsequently, the sampled signal is used for information decoding (ID), and the residual component is redirected towards energy harvesting (EH) by means of a single-pole double-throw (SPDT) switching mechanism. A digital receiver architecture is designed to implement SS, and its performance is evaluated against PS and TS through Monte Carlo simulations over Rayleigh fast fading channels. The results demonstrate an improved bit error rate (BER) and harvested power trade-off, with SS achieving maximum EH gains of approximately 189% over PS and 106% over TS under the considered system parameterization, providing its greatest advantage in ID-prioritized scenarios where the EH branch is most constrainedItem 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 Identify Dyscalculia, Dysgraphia Learning Disabilities in Deaf and Mute Primary Students and Help to Improve Learning Abilities(Institute of Electrical and Electronics Engineers, 2026-01-22) Perera, G; Neththasinghe, H; Rasanjana, D; Thalakotunna, Y; Krishara, J; Rajendran, KLearning disabilities, particularly Dyscalculia and Dysgraphia, significantly hinder students' academic, social, and future occupational outcomes. Deaf and mute primary students face amplified challenges due to limited availability of specialized educational resources and tools that cater to their unique communication needs. This research presents an innovative, intelligent learning environment designed specifically to identify and mitigate Dyscalculia and Dysgraphia among deaf and mute primary students. Leveraging advanced artificial intelligence, computer vision, and machine learning (ML) technologies, the developed system incorporates Sinhala Sign Language (SLSL) and interactive, adaptive learning methodologies. The identification process employs Convolutional Neural Networks (CNNs) to analyze handwritten numerical inputs and sign language gestures, categorizing students based on the severity of their conditions. Subsequently, personalized, engaging instructional activities facilitate gradual skill development and continuous improvement. Robust data privacy measures ensure ethical standards, while automated tracking provides actionable insights for educators and parents. Initial findings indicate significant improvements in student performance and engagement, demonstrating the system's efficacy in delivering inclusive, culturally relevant, and accessible educational interventions tailored for deaf and mute students.