An improved UWB receiver employing generalized normal-Laplacian distribution model

Abstract

Abstract:The generalized normal-Laplace (GNL) distribution is employed to reflect the heavy-tailed and impulsive nature of the multiple access interference (MAI) plus AWGN noise in ultra-wideband (UWB) systems. To accurately represent the impulsive feature of the MAI-plus-noise while keeping longer tails, the kurtosis matching (KM) method combined with the method of moments estimation (MME) is proposed for the parameter estimation for time-hopping UWB multiple access communications in AWGN channels. The GNL based UWB receiver using the KM approach outperforms the conventional matched filter receiver, the soft-limiting receiver, the Gaussian-Laplace mixture receiver, and the MME-based GNL receiver in high SNR ranges.

Novel multiuser ultra‐wideband receiver based on adaptive modelling of generalised normal‐Laplace distribution

“…Thus, the GNL distribution has been recently studied as a good candidate for modelling the MAI-plus-noise environments in the non-Gaussian UWB systems. However, sometimes the GNL PDF based on the MME algorithm does not accurately represent the actual distribution of the non-Gaussian noise (MAI or MAI-plus-noise) in the high SNR range with relative small users [12]. To obtain the more impulsive PDF with longer tails, the authors have modified the MME approach for GNL parameter estimation by using the KM method.…”

“…The conventional matched filter (CMF) receiver based on a Gaussian approximation (GA), which has been widely employed for UWB signal detection, underestimates its bit error rate (BER) performance under multiple access interference (MAI)-plus-noise environments [2][3][4][5]. In the recent contributions, better statistical probability models for the MAI-plus-noise than the GA have been introduced for developing enhanced multiuser UWB receivers [6][7][8][9][10][11][12]. In [6], a Laplace distribution has been used to model the MAI in the time-hopping (TH) UWB systems.…”

“…In [10,11], a new statistical probability distribution, which is called a generalised normal-Laplace (GNL) distribution, has been employed for better modelling the MAI-plus-AWGN in the multiuser UWB systems. Recently, an improved GNL receiver has been presented for TH UWB signal detection on AWGN channels in [12]. The GNL parameters for the GNL receiver have been estimated by a method of moments estimation (MME) combined with a kurtosis matching (KM) method (called MME-KM).…”

“…The GNL parameters for the GNL receiver have been estimated by a method of moments estimation (MME) combined with a kurtosis matching (KM) method (called MME-KM). It has been shown in [12] that the GNL receiver using the MME-KM algorithm achieves better performance than the CMF UWB receiver, SL UWB receiver and GLM UWB receiver.…”

“…In this paper, we develop a novel GNL receiver that can enhance the BER performance of the GNL receiver based on the MME-KM algorithm in [12]. It is shown that the Laplace, GG, GLM and GNL-MME-KM PDFs could not precisely represent the non-Gaussian feature of the multiuser UWB system at high SNR ranges.…”