On The Capacity of Low-Rank Dyadic Fading Channels in the Low-SNR Regime

Kamal Singh

2024-02-10

Abstract:We characterize the capacity of a low-rank wireless channel with varying fading severity at low signal-to-noise ratios (SNRs). The channel rank deficiency is achieved by incorporating pinhole condition. The capacity degradation with fading severity at high SNRs is well known: the probability of deep fades increases significantly with higher fading severity resulting in poor performance. Our analysis of the dyadic pinhole channel at low-SNR shows a very counter-intuitive result that - \emph{higher fading severity enables higher capacity at sufficiently low SNR}. The underlying reason is that at low SNRs, ergodic capacity depends crucially on the probability distribution of channel peaks (tail distribution); for the pinhole channel, the tail distribution improves with fading severity. This allows a transmitter operating at low SNR to exploit channel peaks `more efficiently' and hence improves spectral efficiency. We derive a new key result quantifying the above dependence for the double-Nakagami-$m$ fading pinhole channel - the capacity ${C} \propto (m_T m_R)^{-1}$ at low SNR, where $m_T m_R$ is the severity parameters (product) of the fadings involved.

Information Theory