Channel Fading in Mobile Broadband Systems: Challenges and Opportunities

Abstract

High-speed data signals transmitted over mobile broadband channels are seriously distorted by both time-varying effect and frequency-selective fading (FSF). These distortions introduce challenges since channel variances in both time-domain and frequency-domain form a two-dimensional channel matrix which is hard to estimate, but meanwhile provide opportunities for information security since all signals are directly encrypted by the channels which are adequately random over time, frequency and space. These challenges and opportunities are studied in this thesis as two parts. In the first part, we propose a novel time-varying channel estimation (TVCE) algorithm named piece-wise time-invariant approximation (PITIA) to estimate a typical type of mobile broadband channels - the high-speed train (HST) channels. PITIA customizes general time-varying channel models according to HST channels' specific features, and outperforms conventional TVCE algorithms by about 3-dB in terms of estimation error. In the second part, we propose the first physical-layer challenge-response authentication mechanism (PHY-CRAM) which uses the mobile broadband channels to prevent eavesdropping during authentication. Since pilots and reference signals are eliminated, eavesdroppers cannot demodulate credential information, while legitimate receivers use the channels' reciprocal property to cancel FSF. PITIA is evaluated by computer based simulations, and the effectiveness of PHY-CRAM is validated by prototyping and real-world experiments. Both pieces of works are built upon a unified system model and orthogonal frequency-division multiplexing (OFDM) modulation.