Research Papers - Dept of Computer Systems Engineering

Permanent URI for this collection https://rda.sliit.lk/handle/123456789/1253

Browse

Filters

2020 - 202514

Advanced Search

Filter by

Settings

Author: search.filters.author.Jayakody, AStart date: 2020

Search Results

Now showing 1 - 10 of 14
  • Thumbnail Image
    PublicationOpen Access
    A novel application with explainable machine learning (SHAP and LIME) to predict soil N, P, and K nutrient content in cabbage cultivation
    (Elsevier B.V., 2025-03-06) Abekoon, T; Sajindra, H; Rathnayake, N; Ekanayake, I, U; Jayakody, A; Rathnayake, U
    Cabbage (Brassica oleracea var. capitata) is commonly cultivated in high altitudes and features dense, tightly packed leaves. The Green Coronet variety is well-known for its robust growth and culinary versatility. Maximizing yield is crucial for food sustainability. It is essential to predict the soil’s major nutrients (nitrogen, phosphorus, and potassium) to maximize the yield. Artificial intelligence is widely used for non-linear predictions with explainability. This research assessed the predictive capabilities of soil nitrogen, phosphorus, and potassium levels with explainable machine learning methods over an 85-day cabbage growth period. Experiments were conducted on cabbage plants grown in central hills of Sri Lanka. SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) were used to clarify the model’s predictions. SHAP analysis showed that high feature values of the number of days and plant average leaf area negatively impacted for nutrient predictions, while high feature values of leaf count and plant height had a positive effect on the nutrient predictions. To validate the results, 15 greenhouse-grown cabbage plants at various growth stages were selected. The nitrogen, phosphorus, and potassium levels were measured and compared with the predicted values. These insights help refine predictive models and optimize agricultural practices. A user-friendly application was developed to improve the accessibility and interpretation of predictions. This tool is a user-friendly platform for end-users, enabling effective use of the model’s predictive capabilities.
  • Thumbnail Image
    PublicationEmbargo
    Forecasting Model of Combining Mini Batch K Means and Kohonen Maps to Cluster and Evaluate Gait Kinematics Data
    (IEEE, 2022-10-04) Indumini, U; Jayakody, A
    When people are getting old, some gait abnormalities may have happened in their walking patterns. It means, there may be slight differences in their physical performance. Due to the complexity of that evaluation, a machine learning algorithm can be used to cluster the gait patterns. Kohonen Maps (KM) and mini-batch k-means (MBKM) have been combined to cluster the gait parameters according to the age groups to identify the principal gait characteristics which are affected to the walking pattern. Dataset is consisting of 180 gait data based on the data which have been gained through the inertial measurement unit (IMU). When analysing the results, the proposed algorithm is showing low computational cost and time which is more efficient. As well the results have been proved that the cadence is the most important and affected gait parameter when caused to a walking pattern of a person when he or she is getting older. These results provide clues for the health professionals to identify and evaluate the difficulties of walking patterns of patients according to age.
  • Thumbnail Image
    PublicationEmbargo
    Common Object Request Broker-based Publisher-Subscriber Middleware for Internet of Things - Edge Computing
    (IEEE, 2022-10-04) Perera, H; Jayakody, A
    The edge computing layer in IoT reduces the flow of a massive amount of data directly to the cloud by processing some data in the local network. The middleware in the layer enables this processing of data and the communication between heterogeneous devices and services in the nearby layers. CORBA, which uses as a powerful middleware technology in developing middleware solutions in enterprise-level distributed applications, has been abandoned in the current generation. The paper presents the design, and the performance evaluation of a publisher-subscriber middleware implemented using CORBA that was studied when exploring the applicability of CORBA as an IoT edge computing middleware. The evaluation was continued in two steps to analyse several parallel connections (Load test) and handle requests in a unit time (burst test) via simulating an IoT environment in a cloud environment.
  • Thumbnail Image
    PublicationEmbargo
    A Smart Aquaponic System for Enhancing The Revenue of Farmers in Sri Lanka
    (IEEE, 2022-10-19) Ekanayake, D; de Alwis, P; Harshana, P; Munasinghe, D; Jayakody, A; Gamage, N
    Sri Lanka's agricultural sector confronts serious challenges from fertilizer shortages and agriculture-related chemical scarcity. Innovations comparable to aquaponic systems may be offered to Sri Lankan farmers to overcome these difficulties using IoT and ML technology. This research scope is to implement a smart and secure aquaponic environment monitoring system to forecast plant and fish growth factors, provide Sri Lankan farmers with insights into the environment's behaviors, and take measures according to the predictions utilizing control mechanisms. In this research, more exact predictions have been generated by the Random Forest algorithm model rather than the LSTM model, and most of the investigated parameters given good accuracy according to the absolute mean error (Media TDS-1.95, Media pH-0.06, Media Temperature-0.49, Env. Temperature- 0.94, Env. Humidity-2.70) except the environment light intensity (64.11). The ML solution studied in this research paper would increase the quality of traditional agriculture in Sri Lanka for greater productivity and economic benefit.
  • Thumbnail Image
    PublicationEmbargo
    Guide-Me:Voice authenticated indoor user guidance system
    (IEEE, 2021-12-01) Dissanayake, D. M. L. V; Rajapaksha, R. G. M. D. R. P; Prabhashawara, U. P; Solanga, S. A. D. S.P; Jayakody, A
    Due to a lack of knowledge about the building structure and possible impediments, the majority of blind persons require assistance when traveling through unknown regions. To solve this issue, this paper provides "Guide-Me" as a strategy for indoor navigation with optimum accessibility, usability, and security, decreasing obstacles that the user may meet when traveling through indoor surroundings. Because the intended audience for this research is blind or visually impaired persons, "Guide-Me" makes use of the user’s voice-based inputs. This paper also includes Bluetooth beacon integration for localization, a Smart stick with sensors for obstacle detection, a machine learning model for voice authentication, and an algorithm protocol for a secure connection between server and application Integration driven architecture to assist vision impaired in navigating the known and unknown indoor environment.
  • Thumbnail Image
    PublicationEmbargo
    Converting high resolution multi-lingual printed document images in to editable text using image processing and artificial intelligence
    (IEEE, 2022-06-21) Jayakody, A; Premachandra, H. W. H; Kawanaka, H
    The optical character recognition technique is used to convert information, mainly printed or handwritten text in paper materials, into an electronic format that the computers can edit. According to the literature, there are few competent OCR systems for recognizing multilingual characters in the form of Sinhala and English characters together. The lack of an appropriate technology to recognize multilingual text still remains as a problem that the current research community must address, and it has been designated as the key problem for this study. The main goal of this research is to develop a multilingual character recognition system that uses character image geometry features and Artificial Neural Networks to recognize printed Sinhala and English scripts together. It is intended that the solution would be improved to cover three Sri Lanka’s most commonly spoken languages, with the addition of Tamil as a later upgrade. The primary technologies for this study were character geometry features and Artificial Neural Networks. At the moment almost an 85% of success rate has been achieved with a database containing around 800 images, which are divided into 46 characters (20 Sinhala and 26 English), and each character is represented in 20 different forms of character images. Recognition of text from printed bi-lingual documents is experimented by extracting individual character data from such printed text documents and feeding them to the system.
  • Thumbnail Image
    PublicationEmbargo
    Agro-Mate: A Virtual Assister to Maximize Crop Yield in Agriculture Sector
    (IEEE, 2021-12-09) Dayalini, S; Sathana, M; Navodya, P. R. N; Weerakkodi, R. W. A. I. M. N; Jayakody, A; Gamage, N
    Information Technology plays a vital role in the agriculture industry. The main goal of the project is to develop a mobile application to support farmers to take accurate decisions and help them with activities such as soil quality determination, best crop selection, rice disease prediction, and disaster prediction for the wet zone of Sri Lanka. To achieve the main goal the project has incorporated advanced technologies such as Deep Learning, Image Processing (IP), Internet of Things (IoT), and Machine Learning that can support farmers or investors in a way to maximize yield. ‘Agro-Mate’ application is developed in a way to facilitate the agriculture industry. ‘Agro-Mate’ consists of four components such as soil quality determination and fertilizer recommendation, best crop selection, rice disease prediction and recommendation, and natural disaster prediction and providing the recommendation. Also, the application suggests fertilizer when soil is lacking quality and provides recommendations whenever rice diseases or natural disasters are identified. The usage of android mobile devices in agriculture is one of the key components of the sector's growth, which facilitates the farmer's inaccurate decision-making to gain more quality and quantity of crops. Agro-mate’ is more likely to increase the productivity of crops and indirectly increase the GDP of Sri Lanka.
  • Thumbnail Image
    PublicationEmbargo
    Assistive Learning Platform for Children with Down Syndrome
    (IEEE, 2020-11-04) Wellala, S; Thathsarani, S. A; Senaratne, D; Samaranayake, P; Jayakody, A
    Down Syndrome (DS) is a genetic disorder. Researchers believe that persons with DS generally have poor logical knowledge, communication, motor skills, and skills needed for everyday life. The proposed system is a web-based assistive learning platform for children with DS to address those problems. It provides an excellent opportunity for learning educational subjects, including math and language. The authors also created interactive modules improving their health habits, social skills, and motor skills. Since the authors deeply consider their requirements, the system was developed by providing excellent features with those modules. Users' faces can be recognized to keep their attention with the system, thereby suggesting and referring their most interesting content according to their emotions while using the system. Also, the dashboard can analyze user data. Most importantly, it capable of assessing the users through the system. Here the research team has assessed 50 children with DS, and 31 showed improvement after using the system. Therefore, the proposed system with all these modules and features can be introduced as a very productive assistive learning platform in Sri Lanka.
  • Thumbnail Image
    PublicationEmbargo
    HemoSmart: A Non-invasive, Machine Learning Based Device and Mobile App for Anemia Detection
    (IEEE, 2020-12-22) Jayakody, A; Edirisinghe, E. A. G. A.
    This paper presents a non-invasive method to detect Anemia (a low level of Hemoglobin) easily. The Hemoglobin concentration in human blood is an important substance to health condition determination. With the results which are obtained from Hemoglobin test, a condition which is called as Anemia can be revealed. Traditionally the Hemoglobin test is done using blood samples which are taken using needles. The non-invasive Hemoglobin measurement system, discussed in this paper, describes a better idea about the hemoglobin concentration in the human blood. The images of the finger- tip of the different hemoglobin level patients which are taken using a camera is used to develop the neural network-based algorithm. The pre-mentioned algorithm is used in the developed noninvasive device to display the Hemoglobin level. Before doing the above procedure, an account is created in the mobile app and a questionnaire is given to answer by the patient. Finally, both the results which are obtained from the mobile app and the device are run through a machine learning algorithm to get the final output. According to the result patient would be able to detect anemia at an early stage.
  • Thumbnail Image
    PublicationEmbargo
    i-Police-An Intelligent Policing System Through Public Area Surveillance
    (IEEE, 2021-10-27) Jayakody, A; Lokuliyana, S; Dasanayaka, K; Iddamalgoda, A; Ganepola, I; Dissanayake, A
    Technology and law enforcement are now commonly used hand in hand to improve public safety. Most police departments only use CCTV cameras at a few major intersections for remote surveillance. The public is waiting too long for emergency response lines, therefore using new technologies to improve the current policing system has become one of the police's main goals. The paper presents a coordinated framework that could identify the subtleties of violations via an automated public area surveillance system, specifically the weapon-related crimes and vehicle accidents, which are then disassembled, analyzed, and stored for future inspections. The trained models are aimed to reduce the false positives of incident detection. The weapon detection system had the best average precision (93.8%) by using YOLOv5 while the vehicle accident detection system resulted in the best average precision (94.9%) by using YOLOv4. The system is tested against the collected set of CCTV footage and tested how long it takes to create a notification which is the main goal of this system. Notification is generated in less than 5 seconds after an incident is detected. The evidence collection engine developed with MATLAB delivered the expected with an accuracy of 97% making the extracted evidence reliable to both vehicle accidents and crime scenes. Additionally, the framework provides an effective and efficient communication channel through which the residents can report crimes to regular parties.