IEEE Conference on Computer Communications (INFOCOM)
Poster: Mutual Information Analysis for Antenna Subset Modulation in Line-of-Sight Channel
Hsiang-Yu Wang, Ooi Yuan Hao, Chia-Yi Yeh
Abstract
Antenna subset modulation (ASM) is a physical-layer security technique that, without affecting the constellation at the legitimate user Bob, scrambles the constellation at the potential eavesdropper Eve located at a different angle than Bob, by randomly switching the transmit antenna subset at the symbol rate. Conventionally, ASM is evaluated using the bit error rate (BER), with a BER of 0.5 representing secrecy. However, BER analysis offers weaker secrecy guarantees compared to information-theoretic-based analysis. That is, a BER of 0.5 does not indicate zero information leakage. Yet, to date, the mutual information (MI) characteristics of ASM have not been explored. In this poster, we study the mutual information of ASM by first deriving its mutual information expression in the line-of-sight (LoS) scenario with AWGN noise. We then employ the mutual information neural estimator (MINE) as a practical method to quantify ASM information-theoretic secrecy when only datasets are available.