Near-Optimal Detection of Monobit Digitized UWB Signals in the Presence of Noise and Strong Intersymbol Interference

Abstract

Ultra-wideband (UWB) channels include many multipath components and thereby introduce a strong intersymbol interference (ISI) for high-rate communications. Hence, in such cases, a high-performance low-complexity detection is required to cancel ISI. Since monobit analog-to-digital converters (ADCs) are simple and consume less energy, they are increasingly used in UWB communication systems. However, since monobit digitization is a nonlinear operation, ISI removal after monobit digitization becomes a difficult task. In this paper, a near-optimal two-stage algorithm is proposed to iteratively detect the noise-and-ISI-contaminated monobit digitized UWB signals. Moreover, a novel theoretical framework for bit-error-rate (BER) evaluation is developed which guarantees a real-valued BER and doubles the speed of the numerical calculations by using the conjugate (Hermitian) symmetric property of the characteristic function. Computer simulations have been employed to validate the theoretical analysis and the approximations used therein. Both simulation and theoretical results show that the monobit receiver employing the proposed algorithm only incurs about 0.9 dB performance loss compared to the optimal monobit receiver. Moreover, the resistance of the proposed algorithm to error propagation is validated.