Performance Analysis of Distributed Wireless Sensor Networks for Gaussian Source Estimation in the Presence of Impulsive Noise

Abstract

We address the distributed estimation of a scalar Gaussian source in wireless sensor networks (WSNs). The sensor nodes transmit their noisy observations, using the amplifyand-forward relaying strategy through coherent multiple access channel to the fusion center (FC) that reconstructs the source parameter. In this letter, we assume that the received signal at the FC is corrupted by impulsive noise and channel fading, as encountered for instance within power substations. Over Rayleigh fading channel and in presence of Middleton class-A impulsive noise, we derive the minimum mean square error (MMSE) optimal Bayesian estimator along with its mean square error (MSE) performance bounds. From the obtained results, we conclude that the proposed optimal MMSE estimator outperforms the linear MMSE estimator developed for Gaussian noise scenario.

Evaluation of Acoustic Noise Level and Impulsiveness Inside Vehicles in Different Traffic Conditions

“…Impulsive phenomena, when the noise varies subtly and greatly from the mean in a short period, can jeopardize the performance of solutions and strategies based on the traditional Gaussian approach [62][63][64]. Impulsive noise is present in several scenarios of communication systems, such as powerline communications [81], OFDM in wireless networks [82], and sensor networks [83]. Unlike the Gaussian, the alpha-stable distribution can have infinite variance, so it better represents data with heavy tails [61,84].…”

Statistical Inference-Based Distributed Blind Estimation in Wireless Sensor Networks

“…Cui et al [14] utilized Best Linear Unbiased Estimator (BLUE) to estimate the sensing value and discussed the power scheduling problem. Liner Minimum Mean Square Error (LMMSE) scheme was utilized in [11], [12], [15], [16]. Xiao et al [11] considered the joint estimation of a source vector by linear source-channel coding.…”

“…Considering the unknown fading channels, Senol and Tepedelenlioglu [15] studied the impact of imperfect channel estimation and presented the optimum number of sensors given the total power and noise statistics. Also, Alam et al [16] considered that the receiving signals were jointly corrupted by impulsive noise and channel fading, then LMMSE was used to recover the Gaussian source signal. Furthermore, distributed estimation schemes based on Maximum Likelihood Estimation (MLE) were proposed in [17]- [20].…”

“…Let the first Gaussian component denote the distribution of the signals from the inactive sensors, and the superposition of the other Gaussian components approximates the distribution of the signals from the active sensors. We can rewrite (16) as…”

EDB‐CHS‐BOF: energy and distance‐based cluster head selection with balanced objective function protocol

“…The advancements in wireless communication technologies have given rise to the development of sensors which are multifunctional smart devices of small size and low price that are capable of sensing, communicating, and conducting operations [1–4]. Environments are monitored and controlled by sensor nodes that gauge different physical phenomena like humidity, pollution levels, pressure, and so forth [5–7].…”