(ReLBT): A Reinforcement learning-enabled listen before talk mechanism for LTE-LAA and Wi-Fi coexistence in IoT

摘要

The emergence of Internet of Things (IoT) has increased number of connected devices and consequently transmitted traffic over the Internet. In this regard, Long Term Evolution (LTE) is growing its utilization in unlicensed spectrum as well, and Licensed Assisted Access (LAA) technology is one of the examples. However, unlicensed spectrum is already occupied by other wireless technologies, such as Wi-Fi. The diverse and dissimilar physical layer and medium access control (MAC) layer configurations of LTE-LAA and Wi-Fi lead to coexistence challenges in the network. Currently, LTE-LAA uses a listen-before-talk (LBT) mechanism, and Wi-Fi uses a carrier sense multiple access with collision avoidance (CSMA/CA) as a channel access mechanism. LBT and CSMA/CA are moderately similar channel access mechanisms. However, there is an efficient coexistence issue when these two technologies coexist. Therefore, this paper proposes a Reinforcement Learning-enabled LBT (ReLBT) mechanism for efficient coexistence of LTE-LAA and Wi-Fi scenarios. Specifically, ReLBT utilizes a channel collision probability as a reward function to optimize its channel access parameters. Simulation results show that the proposed ReLBT mechanism efficiently enhances the coexistence of LTE-LAA and Wi-Fi as compared to the LBT, thus improves fairness performance.