Handover Prediction Based on Geometry Method in mmWave Communications - A Sensing Approach

Although mmWave communication offers many desirable properties for 5G, it faces unique challenges imposed by the special characteristics of a mmWave channel. In particular, frequent handover and the resultant high switching latency is an intrinsic problem faced by small and densified mmWave cellular networks. This issue is further exacerbated by the fact that a mmWave channel is vulnerable to the line-of-sight (LOS) blockage, which leads to a sudden drop or outage of the signal. Because a terminal's movement could frequently render the communication LOS be blocked by obstacles, a mmWave link has to handover even more frequently, causing higher handover time. To address this issue, in this paper we propose a LOS-blockage prediction mechanism for small and densified mmWave cellular networks, which can eliminate unnecessary handover and make necessary handover more smooth/seamless by exploiting accurate prediction on when the LOS component in a mmWave channel will be blocked and how long the blockage will last. The feasibility and effectiveness of the proposed mechanism are validated based on experiments in a real mmWave communication testbed.