DEVELOPMENT AND DEPLOYMENT OF WEB APPLICATION FOR USER CLUSTERING USING MODIFIED K-MEANS ON PD-NOMA

Power Domain Non-Orthogonal Multiple Access (PD-NOMA) is a multiple access technique that supports the system to achieve ultra-low latency and ultra-high connectivity in 5G cellular communications. NOMA has a drawback, if imperfect successive interference cancellation (SIC) occurs, error happened...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Rafii Pradana Santoso, Putra
التنسيق: Final Project
اللغة:Indonesia
الوصول للمادة أونلاين:https://digilib.itb.ac.id/gdl/view/79654
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
المؤسسة: Institut Teknologi Bandung
اللغة: Indonesia
الوصف
الملخص:Power Domain Non-Orthogonal Multiple Access (PD-NOMA) is a multiple access technique that supports the system to achieve ultra-low latency and ultra-high connectivity in 5G cellular communications. NOMA has a drawback, if imperfect successive interference cancellation (SIC) occurs, error happened when deciding to reduce the wrong modulated signal, it will produce residual interference. A user clustering system is needed so that it can prevent errors and interference that ultimately increase the sum-rate value. Modified k-means clustering is used as the clustering model. In the experiment, two modified k-means models are proposed by using silhouette coefficient as a determinant of k value, the first model uses optimum distance and the second model uses a combination of k-means clustering with near-far pairing. Furthermore, it will be compared with Time Division Multiple Access (TDMA) and near-far pairing. The results show that the first modified k-means model, with optimum distance, produces a better sum-rate than other clustering models. In line with this, a web application was also built which is used as a place to visualize and deliver information on the results of the analysis of the user clustering process using modified k-means. Users can provide input in the form of user coordinate positions to find out how the process works and the analysis results generated from the program. The web application uses the streamlit framework and is deployed using the streamlit community cloud and google cloud. There are two main features that can be used, file uploads and random generators, from both of which user position data will be obtained which will later be processed in the program. Users can also download reports on the output of the clustering process.