Beamforming Algorithms for Information Relaying in Wireless Sensor Networks

We develop beamforming algorithms for information relaying over shared slowly nonselective fading channels in wireless sensor networks. We assume that, prior to beamforming their received data to a destination, the relays preprocess them by either data amplifying or decoding. The beamforming weights are broadcasted by the destination to the relays and are formed based on the individual relay-destination channel coefficients and an
m
-bit description of the quality of each source-relay channel. For both relay data-preprocessing models, we present methods for optimizing the
m
-bit quantizer employed at each relay for encoding its source-relay channel quality level, and for choosing the beamforming weights at the destination, so as optimize the destination uncoded bit error rates. As our simulations and analysis reveal, a coarse single-bit description of each source-relay channel coefficient at the destination may suffice, as it results in only a small increase in uncoded bit error rates with respect to the case where full knowledge of the source-relay channel coefficients are exploited at the destination.