Intelligent NLOS Backhaul for 5G Small Cells

Millimeter wave (mmW) technologies are currently being proposed as backhaul for outdoor urban small cells. However, non-line-of-sight (NLOS) operation must be addressed before solutions can be fully realised. This letter proposes a reinforcement learning algorithm for improving the reliability of an mmW NLOS small cell backhaul system based on propagation by diffraction. Simulation results show that the algorithm achieves the desired data rate and can detect malfunctioning on a path within a predetermined time, and make a decision to switch to an alternative path, adjust transmission power, or change the operating mode in order to improve system performance.