Enhanced FSK Demodulation with Accurate Log-Likelihood Ratio for Plasma Sheath Channels

The plasma sheath channel (PSC) is very challenging due to its fast time-varying characteristics. Thereby, communication systems under PSCs should be elaborately designed, so as to deal with the radio blackout problem. In the existing studies, Mary FSK (MFSK) modulation has been shown more appropriate than other schemes, however, the accurate log-likelihood ratio (LLR) for MFSK under PSCs is still unknown. In this paper, we propose an accurate LLR to further enhance the performance of MFSK demodulation under PSCs. In particular, we use correlators instead of a frequency discriminator. Also, we make use of both the amplitude information and the phase information, which is the key difference between this work and the existing papers. This point however renders the derivation of LLR computationally difficult. To solve this problem, we approximately model the outputs of the correlators as the Rayleigh and log Normal distribution, and then the LLR for MFSK modulation is derived in a compact form. The proposed adaptive detection scheme does not need numerical integration and has lower computational complexity compared to the existing frequency detection scheme. Finally, simulations show that the proposed method can dramatically improve the BER performance under highly dynamic PSCs.