Coexistence of LTE-Unlicensed and WiFi: A Reinforcement Learning Framework

K. P. Naveen, Chaitanya Amballa

Abstract: The proposal of LTE-unlicensed, whereby LTE operators are allowed to transmit in the unlicensed WiFi spectrum, is seen as a promising solution to meet the increasing demand for mobile data. Implementing LTE-unlicensed however requires novel techniques to enable LTE to coexist with the already existing WiFi technology. In this work we propose a detailed queueing theory based framework to model the problem of achieving coexistence between the LTE and WiFi technologies using a time-division duplexing method referred to as CSAT (Carrier Sense Adaptive Transmission). Then, using techniques from Markov decision processes and Q-learning we propose an online algorithm that can enable LTE and WiFi to share the unlicensed spectrum in an efficient manner. Finally, we conduct an extensive simulation study to demonstrate the efficacy of the proposed solution.

External IDs:dblp:conf/comsnets/NaveenA21