作者: Syed Ali Jafar , Krishna Srikanth Gomadam
DOI:
关键词:
摘要: We explore the SNR-optimal relay functionality in a \emph{memoryless} network, i.e. network where, during each channel use, signal transmitted by depends only on last received symbol at that relay. develop generalized notion of SNR for class memoryless functions. The solution to optimization problem leads novel concept minimum mean square uncorrelated error estimation(MMSUEE). For elemental case single relay, we show MMSUEE is function regardless source and transmit power, modulation scheme. This scheme, call estimate forward (EF), also shown be with PSK parallel network. demonstrate EF performs better than best amplify (AF) demodulate (DF), both serial networks. determine AF near-optimal low power while DF high hybrid networks contain elements, when robust performance desired, advantage over found significant. Error probabilities are provided substantiate gain obtained through optimality. that, \emph{Gaussian} inputs, AF, become identical.