Fast beam discovery for mmWave cellular networks

Pencil beamforming is used to compensate for channel impairments in millimeter wave cellular networks. This method requires testing of all beam combinations at the base and mobile stations in order to detect the unique beam pair with the optimum signal level for use in the control and data planes. However exhaustive search results in high computational complexity and prolonged discovery times, and therefore has an implicit incremental impact on control plane latency. Furthermore, power usage also increases, leading to reduced energy efficiency. Hence this paper presents fast and adaptive beam search algorithms derived from the heuristic Nelder Mead method, the Levy flight random walk, and the Von Neumann and Chebyshev distance searches. These innovations give a notable reduction in complexity (25-99%), reduced discovery time and lower energy consumption versus existing solutions. Success rates are also high.