Transmit-receive beamforming for 60 GHz indoor wireless communications

Abstract

The vast unlicensed bandwidth available in the 60 GHz band is an attractive solution to provide multi-gigabit bit-rates over short distances in indoor environments. One of the crucial problems of the 60 GHz band is the limited link budget. In order to improve the link budget, antenna beam-forming techniques are employed at least at one end of the transceiver system.

This thesis studies the topic of transmit-receive (Tx-Rx) beam-forming, investigating the impact of the array size and the nature of the channel (LOS/NLOS) on the system performance. The scope of the investigation is limited to uniform rectangular arrays (URA) and to analog beam-forming with one scalar weight per antenna. In order to evaluate the Tx-Rx system, a multiple-input multiple-output (MIMO) Semi-Deterministic Channel Model (SDCM) is introduced, based on a combination of ray-tracing and the well-known Saleh-Valenzuela statistical model.

The MIMO channel is then applied to a beam-forming system based on beam-switching. With this technique, the Tx-Rx beam-vector pair that maximizes the average output SNR is selected within a codebook of pre-defined orthogonal beam-vectors spanning the whole 3-D space. The system performance is evaluated in terms of beam-forming gain, coherence bandwidth of the beam-formed channel, and average spectral efficiency in a band of 2 GHz.

The simulation results show that the beam-switching technique improves the system performance; the improvement is proportional to the array size and is observed both in LOS and NLOS cases (where the LOS path is obstructed). The average spectral efficiency is compared to that of an optimal beam-forming scheme, showing an acceptable performance penalty. Finally, alternative analog beam-forming techniques are investigated and compared against the beam-switching method. The investigation shows that within the class of analog beam-forming, and for the considered channel, beam-switching is a valid cost-performance trade-off.