Interference-Aware Emergent Random Access Protocol for Downlink LEO Satellite Networks

TL;DR

This work tackles interference-aware random access in downlink LEO satellite networks using a multi-agent deep reinforcement learning framework to learn emergent protocols. It introduces Ce2RACH, a centralized and compressed emergent signaling scheme that aggregates per-user activations through a shared relay and autoencoder-based signaling compression to mitigate inter-satellite interference. Compared with eRACH and De2RACH, Ce2RACH delivers substantial throughput gains and collision reductions while keeping signaling overhead scalable, highlighting downlink LEO suitability for MADRL-based protocol learning. The approach advances scalable, interference-aware protocol learning for 6G non-terrestrial networks with practical signaling efficiency improvements.

Abstract

In this article, we propose a multi-agent deep reinforcement learning (MADRL) framework to train a multiple access protocol for downlink low earth orbit (LEO) satellite networks. By improving the existing learned protocol, emergent random access channel (eRACH), our proposed method, coined centralized and compressed emergent signaling for eRACH (Ce2RACH), can mitigate inter-satellite interference by exchanging additional signaling messages jointly learned through the MADRL training process. Simulations demonstrate that Ce2RACH achieves up to 36.65% higher network throughput compared to eRACH, while the cost of signaling messages increase linearly with the number of users.

Figures (4)

• Figure 1: Downlink communication after random access with two LEO satellites in the presence of inter-satellite interference.
• Figure 2: Architectures of eRACH, De2RACH, and Ce2RACH.
• Figure 3: Resource utilization snapshots with respect to collision.
• Figure 4: Signaling communication costs of De2RACH and Ce2RACH.