Answers are grounded in this paper and cite their sources.
Partially coherent detection is studied as a problem of detection over diversity channels with unknown amplitudes, and known or perfectly estimated phases. The optimum composite hypothesis receiver is derived for binary, uniformly orthogonal signaling. Because of the high corn plexity of this optimum receiver, two suboptimal structures are considered a Linear equal gain combiner (LEGC) receiver and a quadratic coherent combiner, which is also a generalized likelihood ratio test (GLRT) receiver for this case. We examine the combiner output signal‐to‐noise ratio (SNR) for the two suboptimal receivers and show that this measure cannot be a performance comparison criterion for receivers of different orders. The exact performance of the GLRT receiver is evaluated in terms of the conditional, average and asymptotic bit error rates (BER), and graphical results are presented. An interesting result is that these BER expressions are obtained, for even and odd diversity orders, via different routes and that quadratic noncoherent detection represents a degenerate case of quadratic coherent detection with even diversity. The average BER of LEGC is exactly evaluated for dual diversity. LEGC and quadratic coherent combining are compared with respect to the asymptotic BEA and the exact BER for dual diversity. It is shown that LEGC is better in terms of the average BER in fading, while quadratic coherent combining is conditionally better whenever the signals received over different links have significantly different strengths.
No takes yet. Share an insight, caveat, or question.
Abdel-Ghaffar et al. (1996) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: