Publication Date

Spring 2024

Degree Type

Master's Project

Degree Name

Master of Science in Computer Science (MSCS)

Department

Computer Science

First Advisor

Navrati Saxena

Second Advisor

Robert Chun

Third Advisor

Abhishek Roy

Keywords

Handover, Long Short Term Memory Neural networks, Machine Learning, Predictive Proximity, Satellite Networks.

Abstract

Modern telecommunications heavily rely on Satellite communication networks to provide global coverage, especially in remote areas which link the whole world in a loop. Conventional handover algorithms methods rely on fixed and predefined rules and thresholds predefined statically to make a handover decision. However, these static handover algorithms may become inefficient under the changing conditions of the network. Therefore, it would be useful to measure the proximity order of satellites to the specific base station. Consequently, the assessed relative proximity helps in optimizing the handovers proactively inside related coverage areas. This results in the service quality and the delays in communications significantly decrease. This project formulates a viable way of enhancing satellite handovers using distance-based predictive models and adding Satellite bandwidth, path loss, etc. This will significantly improve the performance of the satellite handover and ensure that the user receives adequate signal strength to maintain a network connection. Further, simulations are run using real satellite data to verify the efficacy of the technique. This version is expected to guarantee high handover fulfillment, low latency, and acceptable network performance than standard handover approaches. Furthermore, the capability to adapt to any network’s conditions makes this approach suitable for active satellite handover.

Download

Available for download on Thursday, May 22, 2025

Share

COinS