The University of British Columbia

Distributed space-time filtering (DSTF) has been recently proposed as a means for node cooperation in wireless networks. In this paper, we derive a new optimization criterion for the node signature filter vectors (SFVs) used in DSTF and we provide two practical design methods for sets of SFVs. We show that for the special case of a frequency-nonselective channel and two active nodes the SFV optimization problem is related to the design of Grassmannian frames. This relation allows us to derive a lower bound for the performance penalty incurred by the distributed implementation. We show that SFV sets designed with the proposed methods perform close to this lower bound and achieve a significant performance gain over previously proposed designs. Simulation results confirm the excellent performance of the proposed SFV sets in frequency-nonselective and frequency-selective channels.