CNN-LSTM Based Prediction of LoS Duration in Satellite Communication Network

Abstract

Satellite communication system is crucial to the advancement of sixth-generation (6G) networks, providing global connectivity and supporting applications that demand ultraspeed high and low latency. However, the high mobility of low Earth orbit (LEO) satellites poses significant challenges to maintaining consistent communication links, primarily due to the frequent handover required by user equipment. Traditional signal strength-based handover technique, which is effective in terrestrial networks, struggles in the satellite network due to the minimal signal strength variation across overlapping cells. This leads to frequent and unnecessary handovers, resulting in reduced network performance. To address these challenges, we propose an CNN-LSTM based predictive line-of-sight (LoS) duration estimation method that allows user terminals to anticipate satellite connection stability, minimizing unnecessary handovers and improving network efficiency in dynamic satellite environments.