Digital Communication Systems in Impulsive Atmospheric Radio Noise

Abstract

Analyses are presented of the performance of binary and M-ary coherent and noncoherent communication systems operating in the impulsive atmospheric radio noise environment. The receiver is usually a maximum likelihood detector for white Gaussian interference and therefore has the form of a parallel bank of matched filters followed by decision circuitry. By employing a Poisson or generalized Shot noise model for the impulsive noise with a suitable probability density function (pdf), closed-form expressions and bounds of error probabilities for M-ary noncoherent and coherent amplitude-shift keying (ASK), phase-shift keying (PSK), and frequency-shift keying (FSK) systems are obtained and the results discussed. ManuscriptIt is well kniown that muclh of the noise degrading the performance of radio communication is impulsive in nature.Tuie impulsive nioise nmay arise as a result of sparking in electric systemis oi-natural phenomena like lightning, the latter giving rise to atmospheric radio noise [1-3] iis type of noise is usually observed as a number of high amplitude, low duratioin impulses superposed on a relatively weak background [4-7] Suchi noise is enicountered friom very low frequency bands to even microwave bands [8]. Tihe noise may be predominantly impulsive or continuous depending upon the frequency range of interest, geophysical location, time of day, anid season of the year.At tropical latitudes, the noise is predominantly impulsive and is tlhe principal source of interference to radio conmmunications in the HiF band [3]. For evaluating the performance of any commllunication system, one requires a mathematically tractable model for impulsive noise wlhich is, of course, consistent witlh the observed clharacteristics.Thie amplitude probability distribution is reported to be non-Gaussian [5,15]. Studies of digital system performance in the face of impulsive noise are comparatively few [9 -18]. Zeimler [14] employed Poisson or generalized Shot noise model in obtaining the performance of M-ary colherent digital systems. The aimplitude of the noise has been assumed to follow a Gaussian distribution. Huynlh and Lecours [17] employed the samne noise model for the analysis of noncoherent systems, assuming the Gaussian distribution for noise amplitudes. By and large, very little explanation has been presented for the results obtained. Shinde [16] analyzed the performanice of binary systems by using an altogether different model of noise. However, the analysis is limited to binary systems alone.An analysis of the performance of digital systems. binary and M-ary, coherent and noncoherent, is presented for impulsive atmospheric noise. The noise model is assumed to be Poisson or generalized Shlot noise model with the amplitudes following the distribution proposed by Shinde [15] . Thiis probability density function (pdf) has been reported to be in excellent agreement with observations [16] While for binary systems it has been possible to evaluate precisely the performance values, only bounds could be worked out for highe...

Error Probability for Fading CPSK Signals in Gaussian and Impulsive Atmospheric Noise Environments

“…On this assumption and because in practice [1, 7] the atmospheric noise is highly impulsive (i.e., y < 1), from (5) and (6) Putting r = 1 in (9) (Gaussian noise alone) one obtains the well known [8] result for Gaussian noise and signal fading, for r = 0 (purely impulsive noise) and in the absence of signal fading, (9) is in agreement with the results given in [1] . Fig.…”

“…Recently [1,7] the performance of digital communication systems in the presence of purely impulsive atmospheric noise has been evaluated assuming the noise model proposed in [6] which is in excellent agreement with experimental results. However a more realistic approach in order to evaluate the performance of digital communication systems leads one to consider the noise at the receiver input as an additive combination of Gaussian (thermal) and impulsive noise, because in general both types of interference are present though sometimes only one may be predominant.…”

“…The bit error probability Pe has been evaluated considering the atmospheric noise model proposed in [6] (and already considered in [1]) which assumes that the areas of noise impulses are identically distributed according to the pdf: pa(a) = 2/ [7r(1 + a2 )2…”

“…The CPSK system is considered because, among the coherent and noncoherent conventional signaling schemes [on-off keying (OOK), frequency shift keying (FSK), CPSK, and differential phase-shift keying (DPSK)] , it is the best scheme for Gaussian [8] and impulsive [1,7] noise.…”

“…The signal at the receiver input in the signaling interval can be expressed as r(t) =+ F 2E/T cos w0t + n.(t) + n (t) (1) where E is the bit energy content, F is the Rayleigh distributed random variable which accounts for the slow and nonselective fading, n,(t) and ng(t) are the impulsive and Gaussian noise components, respectively, wo = n(2rr/T) with n some fixed integer and T the bit duration.…”