Reinforcement Learning-Based Handover Scheme with Neighbor Beacon Frame Transmission

Mobility support to change the connection from one access point (AP) to the next (i.e., handover) becomes one of the important issues in IEEE 802.11 wireless local area networks (WLANs). During handover, the channel scanning procedure, which aims to collect neighbor AP (NAP) information on all available channels, accounts for most of the delay time. To reduce the channel scanning procedure, a neighbor beacon frame transmission scheme (N-BTS) was proposed for a seamless handover. N-BTS can provide a seamless handover by removing the channel scanning procedure. However, N-BTS always requires operating overhead even if there are few mobile stations (MSs) for the handover. Therefore, this paper proposes a reinforcement learning-based handover scheme with neighbor beacon frame transmission (MAN-BTS) to properly consider the use of N-BTS. The optimization equation is defined to maximize the expected reward to find the optimal policy and is solved using Q-learning. Simulation results show that the proposed scheme outperforms the comparison schemes in terms of the expected reward.