Adaptive beamforming for 60 GHz radios

Abstract

Beamforming is a critical and natural solution component for 60 GHz radios due to the increased signal attenuation that occurs as a consequence of the very small wavelength (mm-length) at these frequencies. In this paper, we discuss two practical challenges for 60 GHz beamforming that are likely to arise in most real-world antenna implementations, namely the infeasibility of (a) explicit channel estimation, and (b) generating isotropic antenna patterns. To overcome these challenges, we design a novel iterative training algorithm for antenna weight vectors (AWVs) that uses relatively coarse-grained feedback to obtain the optimal transmit and receive AWVs. We also propose a simple, but effective technique to identify the best transmit and receive sectors that serve to initialize our AWV training algorithm - without these well-chosen initial AWVs, the convergence time of our algorithm can be significantly longer.