Spatial-Temporal Learning-Based Distributed Routing for Dynamic LEO Satellite Networks

TL;DR

This paper introduces a novel distributed routing framework for dynamic LEO satellite networks using spatial-temporal learning, integrating GAT, LSTM, and deep Q-networks for adaptive decision-making.

Contribution

It presents a new deep learning-based routing approach formulated as a POMDP, outperforming traditional methods in throughput, delay, and congestion management.

Findings

Significantly improves throughput and reduces packet loss.

Achieves up to 23.26% queue reduction with proactive congestion avoidance.

Maintains low computational overhead and negligible carbon emissions.

Abstract

In this paper, we propose a spatial-temporal learning-based distributed routing framework for dynamic Low Earth Orbit (LEO) satellite networks, where graph attention networks (GAT) and long short-term memory (LSTM) are integrated within a deep Q-network (DQN)-based architecture to enable distributed and adaptive routing decisions based on local observations. The routing problem is formulated as a partially observable Markov decision process (POMDP) to address partial observability under dynamic topology and time-varying traffic. Simulation results show that the proposed method significantly outperforms conventional and learning-based routing schemes in terms of throughput, packet loss, queue length, and end-to-end delay, while achieving proactive congestion avoidance with up to 23.26% queue reduction. In addition, the proposed approach maintains low computational overhead with negligible carbon emissions, demonstrating its efficiency from a Green AI perspective.