Optimization of Adaptive Frequency Hopping Strategies for Satellite Communications Applying LSTM Neural Networks

With the increase of communication bandwidth and the number of communication terminals, frequency hopping collision issues become more and more serious. Aiming at this issue, this article intended to optimize the adaptive frequency hopping strategy of satellite communication using LSTM (Long Short-term Memory) neural network. Furthermore, the dynamic performance model of satellite communication system based on LSTM neural network was studied. Simulation results showed that the algorithm had good adaptive ability, could effectively reduce the bit error rate and packet loss rate, and improved its maximum throughput to 898 Gbps. The research results provide a new idea for the optimization of frequency hopping strategy in the future satellite communication system, and it is also a valuable research topic.