DEVELOPMENT OF WEB DASHBOARD FOR VISUALIZING MACHINE LEARNING-BASED BEAM TRACKING ON 5G HIGH SPEED TRAIN

• Main Author: Tinarta, David
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/82253
• Institution: Institut Teknologi Bandung

At the command center of PT Kereta Api Indonesia, high-capacity services are required to ensure smooth and efficient operations. The performance of beam tracking on the 5G network is crucial to monitor in order to maintain optimal service quality. Therefore, a dashboard capable of visualizing beam tracking performance data is needed to allow operators to monitor and assess network performance. This way, potential issues can be identified and addressed to reduce the risk of service disruptions. Additionally, the dashboard is expected to assist in analyzing historical data to evaluate long-term performance and make decisions based on observed trends. The ability to customize the data display according to needs also allows operators to focus on the metrics and parameters most relevant to current operational conditions. The goal of this final project is to develop a front- end web dashboard system capable of visualizing beam tracking performance data generated by machine learning algorithms for high-speed 5G trains. With this visualization, the predicted performance of beam tracking can be presented in a way that is easier for users to read and understand. The web dashboard is developed using Streamlit, a Python framework that allows for the rapid and easy creation of interactive web applications. This dashboard displays various data visualizations that enable users to analyze beam tracking performance in real-time and historically. The interactive dashboard also provides users with the flexibility to customize visualizations as needed, such as filtering data based on specific parameters. The results of this final project show that the developed front-end web dashboard system can be an effective tool for visualizing beam tracking prediction performance on 5G networks, is interactive for users, easily customized and can be easily read and understood.