Method of cell coverage expansion and adjustment based on reinforcement learning and SDN

• 1、National Mobile Communications Research Lab., Southeast University, Nanjing 210096

Abstract：Cell-wide extension (CRE) is a method of extending the virtual base station range by increasing the offset value for the received power of the micro base station signal. Without the need to increase the transmission power of the micro base station to reach the effect of enhancing base station coverage, cell edge throughput and full network throughput. Currently, many studies focus on Intercell Interference Cancellation (ICIC) in CRE because the high transmit power of the macro base station (MBS) affects the signal of the user equipment (UE) in the extended area. The optimal offset value that minimizes UE outage probability depends on a number of factors, such as the ratio of MBS and PBS radio resources. And the optimal offset value of each UE is different. However, at present most of the research is to consider setting the unified offset value for all UEs through repeated experiments. In this paper, we propose a method that uses the SDN controller to find the optimal offset value for each UE and minimize the number of interrupts per UE using the Q-learning algorithm. Simulation results show that the Q-learning method can reduce the UE outage probability and improve the network throughput, compared with the traditional optimal offset value setting method.

Keywords： Information and communication systems Cell-wide extension Q-learning outage probability SDN