A Learning-Based Coexistence Mechanism for LAA-LTE Based HetNets

TL;DR

This paper proposes a novel learning-based coexistence mechanism for LAA-LTE heterogenous networks, optimizing spectrum sharing with WiFi users through a two-level framework combining Q-learning and game theory.

Contribution

It introduces a joint resource allocation and access framework using a two-level learning approach to improve spectrum efficiency and coexistence in LAA-LTE HetNets.

Findings

Enhanced spectrum utilization demonstrated in simulations

Effective coexistence with WiFi users achieved

Adaptive learning methods outperform traditional approaches

Abstract

License-assisted access LTE (LAA-LTE) has been proposed to deal with the intense contradiction between tremendous mobile traffic demands and crowded licensed spectrums. In this paper, we investigate the coexistence mechanism for LAA-LTE based heterogenous networks (HetNets). A joint resource allocation and network access problem is considered to maximize the normalized throughput of the unlicensed band while guaranteeing the quality-of-service requirements of incumbent WiFi users. A two-level learning-based framework is proposed to solve the problem by decomposing it into two subproblems. In the master level, a Q-learning based method is developed for the LAA-LTE system to determine the proper transmission time. In the slave one, a game-theory based learning method is adopted by each user to autonomously perform network access. Simulation results demonstrate the effectiveness of the proposed solution.