Article

Feasibility of Ultra-Wideband Channels at Millimeter

Wavelengths Faded by Rain in GeoSurf Satellite Constellations
Emilio Matricciani * , Maurizio Magarini and Carlo Riva

Dipartimento di Elettronica, Bioingegneria e Informazione, Politecnico di Milano, 20133 Milan, Italy;

[email protected] (M.M.); [email protected] (C.R.)
* Correspondence: [email protected]

Abstract: We have studied the interference caused by amplitude and phase distortions induced by
rain in ultra-wideband communication systems designed for using amplitude modulation in GeoSurf
future satellite constellations. The results concern radio links simulated with the synthetic storm
technique at Spino d’Adda (Italy), Madrid (Spain) and Tampa (Florida), which are sites located in
different climatic regions. The conclusions are (a) the three sites, although in different climatic zones,
are practically indistinguishable; (b) the channel signal-to-noise ratio can be increased or decreased by
interference with equal probability. Channel theoretical capacity loss, even in the worst case, is very
limited and rain, therefore, does not cause significant linear distortions in ultra-wideband channels at
millimeter waves; therefore, these channels could be used at millimeter waves.

Keywords: baseband channel; GeoSurf constellations; interference; linear distortions; millimeter

wavelengths; passband channel; rain attenuation; synthetic storm technique; time delay; ultra-
wideband channels

1. Satellite Constellations with Zenith Propagation Paths at Any Site

Among satellite constellations, the GeoSurf constellations are a good choice for future
Citation: Matricciani, E.; Magarini,
worldwide internet infrastructure because they share most of the advantages of current
M.; Riva, C. Feasibility of GEO (geostationary), MEO (medium Earth orbit) and LEO (low Earth orbit) satellite constel-
Ultra-Wideband Channels at lations without suffering most of their drawbacks because they emulate the geostationary
Millimeter Wavelengths Faded by orbit with zenith paths in sites located at any latitude [1].
Rain in GeoSurf Satellite In [2], we have compared the tropospheric attenuation of GeoSurf zenith paths with
Constellations. Telecom 2023, 4, the paths of GEO, MEO and LEO satellites.
732–745. https://doi.org/10.3390/ In [3], we have assessed how the annual average probability distribution P( A) of
telecom4040033 exceeding a given rain attenuation A (dB) depends on the carrier frequency f c (GHz) in the
Academic Editor: Petros Bithas
GeoSurf paths at sites in different climatic regions. All these studies refer to narrow-band
channels, i.e., channels with negligible linear (amplitude and phase) distortions.
Received: 17 July 2023 In [4], we have estimated the slowly time-varying transfer function and linear distor-
Revised: 18 September 2023 tions that are likely found in ultra-wideband radio links in GeoSurf constellations working
Accepted: 10 October 2023 at millimeter wavelengths. As a practical example, the bandwidth considered was 10 GHz
Published: 25 October 2023
wide, centered at 80 GHz (W–Band), because we think it might be used in future worldwide
internet radio links using spread spectrum modulation and code division multiple access
(CDMA) [5–10] with BPSK and QPSK modulation, once high-frequency large wideband
Copyright: © 2023 by the authors.
technology—now developed at lower frequencies [11]—will also be available at W-band.
Licensee MDPI, Basel, Switzerland. CDMA can provide a large processing gain and can be designed without considering issues
This article is an open access article related to frequency and access coordination which, together with the advantages of the
distributed under the terms and GeoSurf constellations mentioned in Reference [1], can be an effective choice to provide
conditions of the Creative Commons high bit rates to users.
Attribution (CC BY) license (https:// The literature on what today is defined as “wideband” communication does not refer
creativecommons.org/licenses/by/ to rain attenuation or to the ultra-wideband radio links studied here but instead refers only
4.0/). to radio links in clear-sky conditions (mainly multipath), both for terrestrial and satellite

Telecom 2023, 4, 732–745. https://doi.org/10.3390/telecom4040033 https://www.mdpi.com/journal/telecom

Telecom 2023, 4 733

systems [5,12–16]; therefore, we discuss these topics further, following work carried out
by [4,17].
Following our previous study [4], the purpose of this paper is to assess—with simula-
tions using onsite rain rate time series and the synthetic storm technique (SST) [18]—the
interference due to amplitude and phase distortions produced by rain attenuation in the
ultra-wideband communication systems mentioned above and designed for using double
sideband-suppressed carrier (DSB-SC) modulation in both quadrature channels (QPSK).
We assume that the total signal-to-noise ratio (SNR) of the Gaussian channel is greater
than the minimum required to guarantee a bit error probability smaller than the maximum
value tolerated by users. Under this hypothesis, we estimate how the in-band attenuation
and phase delay affect the baseband digital signal. We evaluate the sampler output of a
direct channel (e.g., the cosine channel, inter-symbol interference, ISI) and the quadrature
interference (QI) coming from the orthogonal channel (the sine channel, referred to also as
the quadrature channel).
For illustrating the characteristics of the interference, we report the results concerning
radio links simulated at Spino d’Adda (Italy), Madrid (Spain) and Tampa (Florida), which
are sites in different climatic regions (Table 1). We have considered these specific sites
because the rain rate time series R(t) (mm/h)—averaged in 1 min intervals—have been
continuously recorded onsite for several years, which is a sufficiently long period to provide
reliable experimental results when SST is applied.
The rain fade in GeoSurf radio links is independent of the particular GeoSurf de-
sign (e.g., altitude and number of satellites) because all paths to/from a satellite of the
constellation are always locally vertical (zenith). To determine the slowly time-varying
passband and baseband equivalent transfer functions distorted by rain, we must know
the rain attenuation A(t) (dB) and the phase delay Φ(t) (degrees) time series. Both are
calculated with the SST as shown in [4,18].
After this introduction, Section 2 shows examples of the results obtained with the
SST; Section 3 recalls how to calculate passband and baseband transfer functions; Section 4
reports histograms of interference; Section 5 models probability distributions of interfer-
ence; Section 6 calculates the theoretical channel capacity loss; Section 7 reports a general
conclusion and indicates future work.

Table 1. Geographical coordinates, altitude (km), rain height HR (km), number of years of continuous
rain rate time series measurements at the indicated sites.

Precipitation Rain Rate Data

Site Latitude N (◦ ) Longitude E (◦ ) Altitude HS (km)
Height HR (km) Bank (Years)
Spino d’Adda (Italy) 45.4 9.5 0.084 3.341 8
Madrid (Spain) 40.4 356.3 0.630 3.001 8
Tampa (Florida) 28.1 277.6 0.050 4.528 4

2. Rain Attenuation and Phase Delay Due to Rain in Zenith Propagation Paths
To calculate the complex passband (radio frequency) and complex baseband equivalent
transfer functions of direct and orthogonal channels in DSB-SC modulation, defined and
discussed in [4] and based on classical linear modulation theory (e.g., [19,20]), we must
know the time series A(t) (dB) and Φ(t) (degrees). Both are calculated with the SST [18],
as shown in [4], with the modelling of [21–23].
Figure 1 shows the average annual probability distribution P( R) of exceeding R
(mm/h, averaged in 1 min) measured at Spino d’Adda, Madrid and Tampa. The different
climatic rain conditions at these sites are clearly evident if the rain rates exceeded at equal
probability are compared.
Telecom 2023,44, FOR PEER REVIEW
Telecom2023, 7343

Telecom 2023, 4, FOR PEER REVIEW 3

Figure 1. Annual probability distribution (%) P(R) of exceeding the value indicated in abscissa at
Figure 1.1.Annual
Figure Annualprobability distribution
probability distribution (%)
(%) 𝑃(R)
𝑃(R) of exceeding
of exceeding the value
the value indicated
indicated in abscissa
in abscissa at at
Spino
Spino d’Adda,
d’Adda, Madrid
Madrid and
and Tampa.
Tampa.
Spino d’Adda, Madrid and Tampa.

Figure 2a shows the annual probability distribution P( A) of exceeding A (dB) calcu-

Figure2a
Figure 2ashows
shows the
the annual
annualprobability
probabilitydistribution 𝑃(𝐴)𝑃(𝐴)
distribution of exceeding 𝐴 (dB)𝐴calcu-
of exceeding (dB) calcu-
lated with
lated withEquation
Equation (8)
(8) of Reference
of Reference[4],
[4],and
and Figure
Figure 2b 2b shows
shows the the corresponding
corresponding annual annual
lated with Equation (8) of Reference [4], and Figure 2b shows the corresponding annual
probability distribution P (
distribution 𝑃(Φ)
probabilitydistribution Φ ) of exceeding ΦΦ Φ (degrees) calculated with Equation
(9) of (9) of
probability 𝑃(Φ) ofofexceeding
exceeding (degrees) calculated
(degrees) with with
calculated Equation
Equation (9) of
Reference [4].
Reference [4].
Reference [4].

(a) (b)
Figure 2. (a) Annual probability distribution (%) 𝑃(A) of exceeding the value indicated in abscissa
Figure 2. (a) Annual probability distribution (%) P(A) of exceeding the value indicated in abscissa
at Spino d’Adda, Madrid and Tampa; (b) annual probability distribution (%) 𝑃(Φ) of exceeding the
(a)
at Spino d’Adda, Madrid and Tampa; (b) annual probability distribution(b) (%) P(Φ) of exceeding
the value indicated in abscissa at Spino d’Adda, Madrid and Tampa,
Figure 2. (a) Annual probability distribution (%) 𝑃(A) of exceeding the valueat 80 GHz and with
indicated circular
in abscissa
polarization.
at Spino d’Adda, Madrid and Tampa; (b) annual probability distribution (%) 𝑃(Φ) of exceeding the
Telecom 2023, 4 735

In the following, we consider the relative phase delay (degrees) at radio frequency ψ:

ϕ(ψ) = Φ(ψ) − Φ(ψ = f c ) (1)

We defined the time delay (picoseconds):

1000
T (t) = Φ(t) (2)
360ψ

where ψ is measured in GHz and the relative time delay is given by

τ ( f ) = T (ψ) − T (ψ = f c ) (3)

We defined the normalized magnitude | H (ψ)| of the passband transfer function H (ψ),
at time t, as
A( f c )− A(ψ)
| H (ψ)| = 10 20 (4)
The real and imaginary parts of H (ψ) are given by

Hr (ψ) = | H (ψ)| cos( ϕ(ψ)) (5)

Hi (ψ) = | H (ψ)| sin( ϕ(ψ))

From Equation (5), according to Reference [4] (see Equations (5)–(7) of [4]), we obtain
the equivalent baseband complex transfer function of the direct channel as

Hed ( f ) = Hedr ( f ) + jHedi ( f ) (6)

and the equivalent baseband complex transfer function that couples it to the orthogonal
channel:
Heq ( f ) = Heqr ( f ) + jHeqi ( f ) (7)
Now, from A(t) and Φ(t), the transfer function can be calculated at any time t, as we
will recall in Section 3. Notice that, since R(t) is averaged in 1 min, A(t) and Φ(t) are also
“averaged” in 1 min, so that the time-varying transfer function changes every minute and,
therefore, very slowly compared to any digital signal so that, in every minute, the Gaussian
channel is affected by the same in-band rain attenuation and phase delay (see [4]).
Moreover, we assume that at any time t, the channel SNR is larger than the minimum
required by the bit error probability constraint; therefore, the rain fade, in order to be
compensated with power control or with other techniques, is that which is measured at
the highest frequency of the upper side-band, i.e., at 85 GHz in our exercise. Because our
purpose is to study transfer functions, we assume the channel is still working as required; in
other words, outages are not considered as they are not directly related to transfer functions.
An example illustrates how the SST transforms R(t) into A(t) and T (t). Figure 3
shows R(t) recorded in Tampa and Figures 4 and 5 show the corresponding A(t) and the
time delay T (t) at 80 GHz and the relative values at the extremes of the 10 GHz bandwidth
centered at 80 GHz. Notice that since A(t) and T (t) increase with frequency non–linearly
in decibels and in picoseconds, the relative differences are not odd functions.
Telecom 2023, 4 Telecom
Telecom2023,
2023,4,4,FOR
FORPEER
PEERREVIEW
REVIEW 736

Figure
Figure 3. Rain rate time3.3.
Figure Rainrate
Rain
series Rrate time
(t)time series
recorded 𝑅(𝑡)
seriesat𝑅(𝑡)
Tamparecorded at Tampa on
on 27 September on 27
27 September
1997. According1997.
September 1997. According
According
to the rain totothth
gauge
gauge collectinggauge collecting
collecting
rainfall, rainfall,
the rainrainfall, the rain
the
event started rain
atevent AM, at
started
7:34:27 7:34:27
time.AM,
7:34:27
local AM, local
local time.
Samples time. Samples
Samplesare
are averaged areaverage
in average
1 min. min.
min.

Figure 4. SST simulated event at Tampa, 27 September 1997, started at 7:34:27 AM local time. (U
Figure
panel):4.rain
Figure 4. SST simulated SST simulated
event at Tampa,
attenuation event
𝐴(𝑡)27at Tampa,
September
(dB) 27 September
at 80 GHz 1997, 1997,
started
(circular at started
7:34:27atAM
polarization). 7:34:27
local
(Lower AM local
time.
panel): time.att
relative (
panel):
(Upper panel): rain rain attenuation
attenuation A ( t ) (dB)𝐴(𝑡)
at 80(dB)
GHzat 80 GHz
(circular (circular polarization).
polarization).
ation at the extremes of a 10 GHz bandwidth. Sampling time is 1 min. (Lower (Lower
panel): panel):
relative relative a
ation
attenuation at the at the of
extremes extremes
a 10 GHz of abandwidth.
10 GHz bandwidth.
SamplingSampling time is 1 min.
time is 1 min.
From these time series, we can calculate the passband and equivalent baseband t
From thesefer
timeFrom these
series,
functions, we time series,
can calculate
according we
the
to the can calculate
passband
theory theequivalent
and
discussed passband
in [4]. and equivalent
baseband baseband
transfer
fer functions,
functions, according according
to the theory to the in
discussed theory
[4]. discussed in [4].
Telecom 2023, 4, FOR PEER REVIEW 6
Telecom 2023, 4 Telecom 2023, 4, FOR PEER REVIEW 737

Figure
Figure5.5.SST Figure event
SSTsimulated
simulated 5.event
SSTatsimulated
atTampa, event
Tampa,27 at Tampa,
27September
September 27 September
1997,
1997, started
startedat 1997, started
at7:34:27
7:34:27 AM, at 7:34:27
AM,local
local time.AM,
time. (Upperlocal time. (U
(Upper
panel):
panel): phase
phasedelaypanel):
delay T (tphase
𝑇(𝑡) delay 𝑇(𝑡)at
) (picoseconds)
(picoseconds) at(picoseconds)
80
80 GHz at 80polarization).
GHz (circular
(circular GHz (circular(Lower
polarization). polarization).
(Lower panel):(Lower
panel): relativepanel): rel
relative
phase
phasedelay
delayat atthephase
the delayof
extremes
extremes ataathe
of 10extremes
10 GHz of a 10 GHz
GHz bandwidth.
bandwidth. bandwidth.
Sampling
Sampling time Sampling
time is
is min. time is 1 min.
11 min.

3.3.Ultra-Wideband
Ultra-Wideband3. Ultra-Wideband
ChannelsDistorted
Channels Channels
Distorted Distorted by Rain
byRain
by Rain
In this section, we recall the theoretical the
In this section, In
we this section,
recall the we recall
theoretical theoretical
flowchart
flowchart of the
of flowchart
the of the digital
digital receivers.
digital receivers. receivers.
Figures
Figures 66 Figu
and77show
and showthetheand 7 showreceiver
baseband
baseband the baseband
receiver inideal
in receiver
idealp conditions
conditions in ideal conditions
and
and ininrain and in rain
rainconditions.
conditions. Sconditions.
𝑆(𝑓) the 𝑆(𝑓) i
( f )isisthe
two-sidedspectrum
two-sided spectrum ofofthe
two-sided the Nyquistpulse,
spectrum
Nyquist ofpulse, S
the Nyquist
√𝑆(𝑓) ( f ) is the
pulse,
is the matched
√𝑆(𝑓)
matched is filter,
the
filter, n (
matched
𝑛(𝑡) t ) is
is the receiver
filter,
the 𝑛(𝑡)
receiver is the rece
total additive Gaussian
total white
additive
total additive Gaussian white noise. noise.
Gaussian white noise.

Figure 6. Flowchart of the baseband receiver in ideal conditions. 𝑆(𝑓) is the two-sided spectru
Figure
Figure6.6.Flowchart
Flowchartofofthe
thebaseband
basebandreceiver
receiverininideal conditions.𝑆(𝑓)
idealconditions. S( f ) isisthe
thetwo-sided
two-sidedspectrum
spectrumof of
pNyquist pulse, √𝑆(𝑓) is its matched filter, 𝑛(𝑡) is the receiver total additive Gaussian w
the
the
the Nyquist pulse, √𝑆(𝑓)
Nyquist pulse, S( f ) isisits
itsmatched
matchedfilter, n(t𝑛(𝑡)
filter, is the
) is the receiver
receiver totaltotal
additiveadditive Gaussian
Gaussian whitewhite
noise.
noise.
noise.
In the ideal case (Figure 6), the amplitude at the sampling time is normalized to unity.
In the ideal case (Figure 6), the amplitude at the sampling time is normalized to u
In the
In other ideal case
words, a hard(Figure
decision6), the amplitude
gives at the sampling
∓1 by sampling the pulsetime is normalized
peak at the sampling to unity.
time
In other words, a hard decision gives ∓1 by sampling the pulse peak at the sampling
In
t= 0. words, a hard decision gives ∓1 by sampling the pulse peak at the sampling time
other
𝑡 = 0.
𝑡 = 0.In the rain fade case (Figure 7), we suppose that, at any time t, the received signal is
In the rain fade A( fcase
c) (Figure 7), we suppose that, at any time 𝑡, the received sign
In the by
amplified rain fade case (Figure
multiplying it by 107), 20 we tosuppose
remove the that,
𝐴(𝑓 at any time 𝑡, the received signal is
𝑐 )variable rain attenuation at the carrier
amplified by multiplying 𝐴(𝑓𝑐 ) it by 10 20 to remove the variable rain attenuation at the
frequencyby
amplified = 80 GHz (this
f c multiplying it by 10 20
is not related to the outage
to remove condition
the variable raindiscussed
attenuation above,
at thewhich
car-
rier frequency 𝑓𝑐 = 80 GHz (this is not related to the outage condition discussed ab
must
rier consider 𝑓the
frequency 𝑐 = 80
rain attenuation
GHz (this is at
not85 GHz).
related After
to the this multiplication,
outage condition only
discussedthe in-band
above,
which must consider the rain attenuation at 85 GHz). After this multiplication, only
signal must
which variations of thethe
consider carrier
rain frequency
attenuation areatleft,
85 which
GHz). is the effect
After we want to study.
this multiplication, onlyAtthe
the
in-band signal variations of the carrier frequency are left, which is the effect we wa
output of the sampler, before the hard decision, we find
in-band signal variations of the carrier frequency are left, which is the effect the algebraic sum s +
sd we s of direct
sq want to
study. At the output of the sampler, before the hard decision, we find the algebraic
and orthogonal
study. At the output channel
of thesamples
sampler, at before
the reference
the hard time t = 0, we
decision, corresponding to the sum
find the algebraic peak
𝑠𝑠𝑑 + 𝑠𝑠𝑞 of direct and orthogonal channel samples at the reference time 𝑡 = 0 , co
𝑠value of the direct (cosine)
𝑠𝑑 + 𝑠𝑠𝑞 of direct
and orthogonal
and orthogonal channelchannelsamples (sine) pulses
at the presenttime
reference simultaneously
𝑡 = 0 , corre- in
sponding to the peak value of the direct (cosine) and orthogonal channel (sine) pu
the two orthogonal
sponding to thepresent
peakchannels.
value ofInthe
general,
direct however,
(cosine) both
and values are affected
orthogonal channel by(sine)
self-channel
pulses
simultaneously in the two orthogonal channels. In general, however, both va
(ISI) interference
present due to the
simultaneously tails
in the twoof the previouschannels.
orthogonal and following pulses however,
In general, and to theboth
interference
values
coming from the orthogonal channel (QI).
Telecom 2023, 4, FOR PEER REVIEW 7

Telecom 2023, 4 738

are affected by self-channel (ISI) interference due to the tails of the previous and following
pulses and to the interference coming from the orthogonal channel (QI).

Figure 7.
Figure 7. Flowchart
Flowchart ofof the
the quadrature
quadrature baseband
baseband receiver
receiver in
in rain attenuation. 𝑆(𝑓)
rain attenuation. is the
S( f ) is the two-sided
two-sided
spectrum of the Nyquist reference pulse assumed to be positive, √𝑆(𝑓)
spectrum of the Nyquist reference pulse assumed to be positive, S( f ) is the matched filterfilter
is the matched and
p
and n (t)
𝑛(𝑡) is the receiver total additive Gaussian white noise for each channel of equal
is the receiver total additive Gaussian white noise for each channel of equal power. power.

From these equations, as shown in [4], [4], wewe can

can calculate
calculate the
the baseband
baseband equivalent
equivalent
transfer functions at frequency 𝑓, with 𝐻
at frequency f , with H𝑒𝑑 (𝑓) for
ed ( f ) for the direct channel and 𝐻
and H𝑒𝑞 (𝑓) for the
eq ( f ) for
orthogonal channel (Figure 7), whose averages for Spino d’Adda, Madrid and Tampa are
reported inin [4].
section, we report the statistics of interference due to the direct channel
In the next section, channel
to the
(ISI) and to the quadrature
quadrature channel
channel (QI).
(QI). These
These results are are fundamental
fundamental for assessing the
feasibility of the ultra-wideband
ultra-wideband radio
radio link.
link.

4.
4. Experimental
Experimental Interference
Interference
Numerical
Numerical results
results have
have been
been obtained
obtained byby means
means of of time-domain
time-domain simulations
simulations of
of the
the
system described in Figure 7 in MATLAB/Simulink® ® . We have simulated the transmission
system described in Figure 7 in MATLAB/Simulink . We have simulated the transmission
of a sequence of 104 4symbols drawn from a quadrature phase-shift keying (QPSK) modula-
of a sequence of 10 symbols drawn from a quadrature phase-shift keying (QPSK) mod-
tion of long sequences of independent bits at the bit rate ρ (bits per second, bps) according
ulation of long sequences of independent bits at the bit rate 𝜌 (bits per second, bps) ac-
to the value of the roll-off factor r. Since the baseband width is fixed to B = 5 GHz, the bit
cording to the value of the roll-off factor 𝑟. Since the baseband width is fixed to 𝐵 = 5
rate is a function of r, given by
GHz, the bit rate is a function of 𝑟, given by 2B
ρ= (8)
1 +2𝐵r
𝜌= (8)
We have simulated bit rates corresponding 1 +to𝑟 r = 0.25, i.e., ρ = 8 Gbps (gigabits per
second); r = 0.50,
We have ρ = 6.67
simulated Gbps;
bit rates r = 1, ρ =
correspondingp5 Gbps.
to 𝑟 =A numerical
0.25, i.e., 𝜌 =approximation
8 Gbps (gigabitsof per
the
continuous-time
second); 𝑟 = 0.50, square-root Nyquist
𝜌 = 6.67 Gbps; 𝑟= 1, 𝜌 =S5
filter ( f Gbps.
) has been implemented
A numerical by considering
approximation of the
acontinuous-time
windowing of square-root
16 symbol intervals and an oversampling factor of 4 compared
Nyquist filter √𝑆(𝑓) has been implemented by considering to the
symbol rate.
a windowing of 16 symbol intervals and an oversampling factor of 4 compared to the
rate. we refer to normalized values, the algebraic sum of total interference ssd + ssq
Because
symbol
givesBecause
the factor
weby which
refer the SNR Γvalues,
to normalized obtainable in the ideal
the algebraic sum channel
of total(Figure 6) must
interference 𝑠𝑠𝑑be+
multiplied to obtain the SNR in the presence of ISI and QI. Therefore,
𝑠𝑠𝑞 gives the factor by which the SNR Γ obtainable in the ideal channel (Figure 6) must we defined the
channel factor µ:
be multiplied to obtain the SNR in the presence of ISI and QI. Therefore, we defined the
µ = ssd + ssq (9)
channel factor 𝜇:
The SNR obtainable with interference is given by
𝜇 = 𝑠𝑠𝑑 + 𝑠𝑠𝑞 (9)
2
γ=
The SNR obtainable with interference Γ by
is µgiven (10)
2 (10)
𝛾 =µ𝜇2 +
γ(dB) = 10 × log Γ 10 × log Γ
10 10

In applying these equations,

𝛾(dB) we assume
= 10 × log10that
𝜇2 +the
10 channel
× log10 Γ Gaussian noise power is
constant, although it should slightly increase with rain attenuation.
Our main results are the histograms (relative frequency) of the sampler output ampli-
tude (dB) first but only with ISI and then only with ISI and QI, i.e., with the total interference.
To show the dependence on the rain attenuation calculated at the center frequency (i.e.,
 In applying these equations, we assume that the channel Gaussian noise power is
constant, although it should slightly increase with rain attenuation.
Our main results are the histograms (relative frequency) of the sampler output am-
Telecom 2023, 4 plitude (dB) first but only with ISI and then only with ISI and QI, i.e., with the total 739 inter-
ference. To show the dependence on the rain attenuation calculated at the center fre-
quency (i.e., 80 GHz), we have considered the cases in which rain attenuation (and corre-
80 GHz),time
sponding we have considered
delay) was in the
thecases in which
following rain attenuation
ranges: 9 ≤ 𝐴80 GHz(and≤ corresponding
11 dB (referred timeto, for
delay) was in the following ranges: 9 ≤ A ≤ 11 dB (referred to, for
short, as “10” dB), 18 ≤ 𝐴80 GHz ≤ 22 dB (“20” dB) and 27 ≤ 𝐴80 GHz ≤ 33 dB (“30”
80 GHz short, as “10” dB), dB).
≤ A80 GHz
18 Figure ≤ 22 dB (“20” dB) and 27 ≤ A ≤ 33 dB (“30” dB).
8 shows the relative frequency histogram of the channel factor 𝜇 (dB) for 𝑟 =
80 GHz
Figure 8 shows the relative frequency histogram of the channel factor µ (dB) for
0.25, Figure 9 for 𝑟 = 0.50 and Figure 10 with 𝑟 = 1.
r = 0.25, Figure 9 for r = 0.50 and Figure 10 with r = 1.

Figure
Figure 8. Relativefrequency
8. Relative frequencyhistogram
histogram of
ofthe
thechannel
channelfactor µ (dB)
factor withwith
𝜇 (dB) ISI (upper panel)
ISI (upper and with
panel) and with
ISI and QI (lower panel). Color key: blue = Spino d’Adda; green = Madrid; red = Tampa.
ISI and QI (lower panel). Color key: blue = Spino d’Adda; green = Madrid; red = Tampa. Curve Curve key: key:
continuous99≤≤𝐴A80
continuous GHz≤
80GHz ≤11
11 dB;
dB;dashed
dashed ≤ A≤80𝐴GHz
18 18 ≤ 22 dB; “+” 27 ≤ A80 GHz
80 GHz ≤ 22 dB; “+” 27
≤ GHz
≤ 𝐴80 33 dB.
≤ Roll-off
33 dB. Roll-
factorr =
off factor 𝑟= 0.25.
0.25.

From these results, we can draw the following general conclusions:

From these results, we can draw the following general conclusions:
(1) The three sites, although in different climatic zones, are practically indistinguishable
(1) The in three sites, although in different climatic zones, are practically indistinguishable
all cases.
(2)in All
all cases.
histograms show even symmetry; therefore, indicating that for about 50% of the
(2) Alltime,
histograms show even
we can consider symmetry;
the channel factortherefore, indicating
to be either that
µ < 0 (dB), for about
therefore, Γ; orof the
γ <50%
time,
µ >we can therefore,
0 (dB), consider γ the Γ.
> channel factor to be either 𝜇 < 0 (dB), therefore, 𝛾 < Γ; or
(3)𝜇 >Histograms with only 𝛾ISI>and
0 (dB), therefore, Γ. with ISI + QI are distinguishable.
(3)(4)Histograms
With ISI only,withtheonly
lowest
ISI range of attenuation
and with ISI + QI are(10distinguishable.
dB) shows more marked peaks at
µ = 0 dB and µ ≈ ±0.3 dB. These peaks are largely smoothed when QI is also added.
Telecom 2023, 4, FOR PEER REVIEW 9

(4) With ISI only, the lowest range of attenuation (10 dB) shows more marked peaks at
Telecom 2023, 4 𝜇 = 0 dB and 𝜇 ≈ ±0.3 dB. These peaks are largely smoothed when QI 740 is also
added.
(5) The roll-off factor 𝑟 plays a small role because the values of the peaks change a little,
especially
(5) 𝑟 = 1.r plays a small role because the values of the peaks change a little,
for factor
The roll-off
especially
Because of these = 1.
for r general findings, in Figure 11, we have drawn the average relative
frequencyBecause of theseby
histograms general findings, in
distinguishing Figure
only the 11, weattenuation
three have drawn ranges.
the average relative
In other words,
frequency histograms by distinguishing only the three attenuation ranges. In other
this figure should provide a “global” frequency distribution of the channel factor in a 10 words,
GHzthis figure should provide a “global” frequency distribution of the channel factor in a
bandwidth centered at 80 GHz. In the next section, we will propose a theoretical
10 GHz bandwidth centered at 80 GHz. In the next section, we will propose a theoretical
approach which can justify these findings.
approach which can justify these findings.

Figure
Figure 9. Relative
9. Relative frequency
frequency histogramof
histogram ofthe
the channel
channel factor
factorµ 𝜇(dB) with
(dB) ISI ISI
with (upper panel)
(upper and with
panel) and with
ISI and QI (lower panel). Color key: blue = Spino d’Adda; green = Madrid; red =
ISI and QI (lower panel). Color key: blue = Spino d’Adda; green = Madrid; red = Tampa. CurveTampa. Curve key: key:
continuous 9≤ 9≤ A80 GHz
𝐴80 ≤ 11 dB; dashed 18 ≤ A 𝐴80 ≤ 22 dB; “+” 27 ≤ A 𝐴80≤GHz
33 dB. Roll-off
continuous GHz ≤ 11 dB; dashed 18 ≤ 80 GHz ≤ 22 dB; “+” 2780≤GHz
GHz ≤ 33 dB. Roll-
factor r𝑟 =
off factor =0.50.
0.50.
Telecom
Telecom 2023, 44,
Telecom 2023,
2023, 4, FOR
FOR PEER
PEER REVIEW
REVIEW 741
10
10

Figure 10.
Figure 10. Relative
Relative frequency
frequencyhistogram
frequency histogramofof
histogram ofthe
the
the channel
channel
channel factor
factor
factor 𝜇𝜇 (dB)
µ (dB)(dB) with
withwith ISI (upper
ISI (upper
ISI (upper panel)
panel) and
and with
panel) and
with
with ISIQI
ISI
ISI and and
and QI (lower
QI
(lower(lower panel).
panel).
panel). ColorColor
Color key:
key: key: blue
blue blue == Spino
= Spino Spino d’Adda;
d’Adda;
d’Adda; green
greengreen == Madrid;
Madrid;
= Madrid; red
red =red == Tampa.
Tampa.Tampa. Curve
Curve
Curve key:
key:
key: continuous
continuous
continuous 9≤
9 ≤ A980 ≤ 𝐴𝐴80
GHz 80≤GHz
GHz ≤
11 ≤ 11dashed
11
dB; dB; dashed
dB; dashed
18 ≤ A18 18 ≤≤ 𝐴𝐴80
80 GHz ≤ ≤≤ 22
22 dB;
80 GHz
GHz 22 dB;
“+”dB;
27 ≤“+”
“+”A8027GHz
27 ≤ 𝐴𝐴≤
≤ 8033
80 ≤ 33
≤
dB.
GHz
GHz 33 dB.
dB.
Roll-off
Roll-off
Roll-off factor
factor
factor r = 1. 𝑟𝑟 =
= 1.
1.

Figure
Figure 11. Average
Figure 11.
11. Average relative
Averagerelative frequency
relativefrequency histogram
frequencyhistogram
histogramofof the
ofthe channel
thechannel factor𝜇𝜇µ (dB)
factor
channelfactor (dB) in
(dB) in the
in the presence
the presence of
presence of ISI
of ISI
ISI
and
and QI.
and QI. Curve
QI. Curve key:
Curve key: black
key: black 9
black 99≤ ≤
≤A 𝐴
𝐴80
80GHz ≤
GHz ≤≤ 11;
11;
11; blue
blue
blue 18
18≤
18 ≤
≤A 𝐴
𝐴80 ≤
GHz ≤
80 GHz ≤ 22;
22;
22; red
red
red 27
27≤
27 ≤
≤A 𝐴
𝐴80 ≤
GHz ≤
80 GHz ≤ 33.
33.
33. Curves
Curves
Curves
80 GHz 80 GHz 80 GHz
are averaged
are averaged over the
the sites and
and roll-off factors.
factors.
are averaged over
over the sites
sites and roll-off
roll-off factors.
Telecom 2023, 4 742

5. Modeling Interference
In this section, we propose a theoretical approach which can justify the relative fre-
quency histogram of the channel factor µ found in Section 4.
The channel factor µ at the sampling time tk is given by

k =+ N k =+ N
µ ( t k = 0) = ∑ ssd (tk ) + ∑ ssq (tk )
k =− N k =− N

µ = id + iq (11)
Now, since both id and iq are stochastic variables given by the sum of many independent
values, their probability density function—according to the central limit theorem—should
approach the Gaussian distribution; therefore, the probability density function of µ should
be given by the convolution of two Gaussian models, i.e., another Gaussian probability
distribution with a mean value given by the sum of the two mean values and variance given
by the sum of the two variances; this hypothesis holds if all variables display the same order
of magnitude [24,25].
Let us carry out some useful transformations. According to Figures 8–11, the channel
factor µ ranges between −1 dB (i.e., 10−1/20 = 0.891) and +1 dB (10+1/20 = 1.122). Now,
by recalling that ln( x ) ≈ ( x − 1) if x ≈ 1, we can write

µ ≈ ln(µ) + 1 (12)

Therefore, the relationship between µ in decibels and µ in linear units is given by:

10
µ(dB) = ( µ − 1) (13)
ln(10)

Hence, if µ is Gaussian, then µ(dB) is also approximately Gaussian. However,

Figures 8–10 seem to suggest two different overlapping Gaussian distributions, especially
for r = 1. In other words, µ < 1 and µ > 1 seem due to two similar sets of data of equal
variance but with mean values set approximately at about ±0.3 dB; therefore, the central
limit theorem applies to these two separate distributions. The overall relative frequency
histograms shown in Figure 11 confirm this general finding, especially at low fade (∼10 dB).
The peak at about µ = 0 dB can be explained by the Dirac pulse produced by the constant
− ln10
(10)
of Equation (13). In the experimental histograms, it is due to the very small spread
of the sampler output amplitude due only to the reference pulse and not to the ISI or QI.
Now, the question is: how “bad” is the ultra-wideband channel discussed above? In
the next section, we attempt a quantitative analysis by using Shannon’s channel capacity.

6. Channel Capacity Loss

In this section, we discuss a quantitative analysis of the results obtained in the ultra-
wideband channels studied in the previous sections by using Shannon’s channel capacity.
For this purpose, let us define the capacity factor c, independent of the roll-factor and the
logarithm used in its definition, which is given by

ln 1 + µ2 Γ


c= (14)
ln(1 + Γ)

In Figure 12, we show c versus Γ (dB), calculated at the extreme values of µ, and which
namely approximately ±1 dB.
Let us discuss only the worst case: the red curve referring to the negative values of µ
(dB). We can notice that, as the signal-to-noise ratio Γ (dB) increases in ideal conditions (no
rain), the capacity factor increases from c = 0.825 at Γ = 3 dB to c = 0.950 at Γ = 20 dB;
therefore, the capacity loss ranges from 17.5% to 5%.
Telecom2023,
Telecom 2023,44, FOR PEER REVIEW 12
743

In conclusion,
conclusion,wewethink
thinkthat
thatthe
thecapacity
capacity loss
loss in in
thethe worst
worst casecase is very
is very limited
limited and
and that
that rain,
rain, therefore,
therefore, does does not cause
not cause significant
significant linearlinear distortions
distortions in ultra-wideband
in ultra-wideband chan-
channels at
nels at millimeter
millimeter waves. waves.

12. Capacity
Figure 12. Capacityfactor versusΓ Γ(dB)
factorversus (dB)
at at
thethe extreme
extreme values
values 𝜇; namely,
of µ;ofnamely, ±1The
±1 dB. dB.experimental
The experi-
mentalof
values values of Figures
Figures 10–12
10–12 are arethese
within within these bounds.
bounds.

7. Conclusions
Following
Following our our previous
previous studies
studies concerning
concerning GeoSurf
GeoSurf satellite
satellite constellations, in the
constellations, in the
present paper, we have studied the interference due to amplitude and
present paper, we have studied the interference due to amplitude and phase distortionsphase distortions
produced
produced by by rain
rain in
in ultra-wideband
ultra-wideband communication
communication systems
systems designed
designed for
for using
using double
double
sideband-suppressed
sideband-suppressed carrier carrier (DSB-SC)
(DSB-SC) modulation.
modulation.
For
For illustrating
illustrating thethe general
general characteristics
characteristics of
of linear
linear interference
interference due
due to
to rain
rain in
in ultra-
ultra-
wideband
wideband channels, we have reported and discussed the results concerning radio links
channels, we have reported and discussed the results concerning radio links
simulated
simulated with with SST
SST at at Spino
Spino d’Adda
d’Adda (Italy),
(Italy), Madrid
Madrid (Spain)
(Spain) and
and Tampa
Tampa (Florida),
(Florida), which
which
are
are sites
sites located
located inin different
different climatic
climatic regions.
regions. Numerical results have
Numerical results have been
been obtained
obtained byby
means of time-domain simulations of
means of time-domain simulations of the system. the system.
After
After reviewing
reviewing the the classical
classical theory
theory ofof passband
passband (radio
(radio frequency)
frequency) and
and baseband
baseband
equivalent transfer functions of direct and orthogonal channels in DSB-SC
equivalent transfer functions of direct and orthogonal channels in DSB-SC modulation, modulation, we
have
we havecalculated
calculatedthem when
them radio
when links
radio are faded
links by rain.
are faded by rain.
We defined the self-channel (ISI) interference due to the tails of the previous and
We defined the self-channel (ISI) interference due to the tails of the previous and fol-
following pulses and the interference coming from the orthogonal channel (QI). We have
lowing pulses and the interference coming from the orthogonal channel (QI). We have
reached the following general conclusions:
reached the following general conclusions:
(1) The three sites considered, although in different climatic zones, are practically indis-
(1) The three sites considered, although in different climatic zones, are practically indis-
tinguishable in all cases.
tinguishable in all cases.
(2) The channel factor µ can be either µ < 0 (dB), therefore, γ < Γ; or µ > 0 (dB),
(2) The channel factor 𝜇 can be either 𝜇 < 0 (dB), therefore, 𝛾 < Γ ; or 𝜇 > 0 (dB),
therefore, γ > Γ with equal probability.
therefore, 𝛾 > Γ with equal probability.
(3) Histograms with only ISI or with ISI + QI are diverse.
(3) Histograms with only ISI or with ISI + QI are diverse.
(4) With ISI only, the lowest range of attenuation (10 dB) shows more marked peaks in the
(4) With ISI only, the lowest range of attenuation (10 dB) shows more marked peaks in
relative frequency histograms at µ = 0 dB and µ ≈ ±0.3 dB. These peaks are largely
the relative frequency histograms at 𝜇 = 0 dB and 𝜇 ≈ ±0.3 dB. These peaks are
smoothed when QI is added.
largely
(5) The roll-off smoothed
factor when
r playsQIaissmall
added.role because the peaks change only for r = 1.
(5) The roll-off factor 𝑟 plays a small role because the peaks change only for 𝑟 = 1.
We have proposed a theoretical approach which can justify these findings.
We have proposed a theoretical approach which can justify these findings.
Telecom 2023, 4 744

In conclusion, the channel theoretical capacity loss, even in the worst case, is very
limited and rain, therefore, does not cause significant linear distortions in ultra-wideband
channels at millimeter waves in QPSK schemes.
Future work should be conducted in simulating higher-order QAM ultra-wideband
channels to assess whether these channels are also affected a little by linear distortions due
to rain.

Author Contributions: Conceptualization, E.M., M.M. and C.R.; methodology, E.M., M.M. and C.R.;
software, E.M., M.M. and C.R.; investigation, E.M., M.M. and C.R.; data curation, E.M., M.M. and
C.R.; writing—original draft preparation, E.M., M.M. and C.R.; writing—review and editing, E.M.,
M.M. and C.R.; visualization, E.M., M.M. and C.R. All authors have read and agreed to the published
version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Not applicable.
Acknowledgments: We wish to thank Roberto Acosta, at NASA years ago, for providing the rain
rate data of Tampa and we also wish to thank José Manuel Riera, at the Universidad Politécnica de
Madrid, for providing the rain rate data of Madrid.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Matricciani, E. Geocentric Spherical Surfaces Emulating the Geostationary Orbit at Any Latitude with Zenith Links. Future
Internet 2020, 12, 16. [CrossRef]
2. Matricciani, E.; Riva, C.; Luini, L. Tropospheric Attenuation in GeoSurf Satellite Constellations. Remote Sens. 2021, 13, 5180.
[CrossRef]
3. Matricciani, E.; Riva, C. Outage Probability versus Carrier Frequency in GeoSurf Satellite Constellations with Radio–Links Faded
by Rain. Telecom 2022, 3, 504–513. [CrossRef]
4. Matricciani, E.; Riva, C. Transfer–functions and Linear Distortions in Ultra–Wideband Channels Faded by Rain in GeoSurf
Satellite Constellations. Future Internet 2023, 15, 27. [CrossRef]
5. Kennedy, R.S. Fading Dispersive Communication Channels; Wiley: New York, NY, USA, 1969.
6. Turin, G.L. Introduction to spread spectrum antimultipath techniques and their application to urban digital radio. Proc. IEEE
1980, 68, 328–353.
7. Pickholtz, R.; Schilling, D.; Milstein, L. Theory of spread–spectrum communications—A tutorial. IEEE Trans. Commun. 1982, 30,
855–884.
8. Viterbi, A.J. CDMA Principles of Spread Spectrum Communications; Addinson–Wesly: Reading, MA, USA, 1995.
9. Dinan, E.H.; Jabbari, B. Spreading codes fir direct sequence CDMA and wideband CDMA celluar networks. IEEE Commun. Mag.
1998, 36, 48–54. [CrossRef]
10. Veeravalli, V.V.; Mantravadi, A. The coding–spreading trade–off in CDMA systems. IEEE Trans. Sel. Areas Commun. 2002, 20,
396–408. [CrossRef]
11. Matsusaki, Y.; Masafumi, N.; Suzuki, Y.; Susumu, N.; Kamei, M.; Hashimoto, A.; Kimura, T.; Tanaka, S.; Ikeda, T. Development of
a Wide–Band Modem for a 21–GHz Band Satellite Broadcasting System. In Proceedings of the 2014 IEEE Radio and Wireless
Symposium (RWS), Newport Beach, CA, USA, 19–23 January 2014.
12. Parsons, D. The Mobile Radio Propagation Channel; Wiley: New York, NY, USA, 1994.
13. Simon, M.K.; Alouini, M.S. Digital Communication over Fading Channels: A Unified Approach to Performance Analysis; Wiley: New
York, NY, USA, 2000.
14. Greenstein, L.G.; Andersen, J.B.; Bertoni, H.L.; Kozono, S.; Michelson, D.G. (Eds.) Special Issue on Channel and Propagation
Modeling for Wireless Systems Design. J. Sel. Areas Commun. 2002, 20. [CrossRef]
15. Goldsmith, A. Wireless Communications; Cambridge University Press: Cambridge, NY, USA, 2005.
16. Pérez–Fontan, F.; Pastoriza–Santos, V.; Machado, F.; Poza, F.; Witternig, N.; Lesjak, R. A Wideband Satellite Maritime Channel
Model Simulator. IEEE Trans. Antennas Propag. 2022, 70, 214–2127. [CrossRef]
17. Matricciani, E. A Relationship between Phase Delay and Attenuation Due to Rain and Its Applications to Satellite and Deep–Space
Tracking. IEEE Trans. Antennas Propag. 2009, 57, 3602–3611. [CrossRef]
18. Matricciani, E. Physical–mathematical model of the dynamics of rain attenuation based on rain rate time series and a two–layer
vertical structure of precipitation. Radio Sci. 1996, 31, 281–295. [CrossRef]
19. Schwartz, M. Information, Transimission, Modulation and Noise, 4th ed.; McGraw–Hill Int. Ed.: New York, NY, USA, 1990.
20. Carassa, F. Comunicazioni Elettriche; Bollati Boringhieri: Turin, Italy, 1978.
21. Matricciani, E. Rain attenuation predicted with a two–layer rain model. Eur. Tranactions Telecommun. 1991, 2, 715–727. [CrossRef]
Telecom 2023, 4 745

22. Recommendation ITU–R P.839–4. Rain Height Model for Prediction Methods; ITU: Geneva, Switzerland, 2013.
23. Maggiori, D. Computed transmission through rain in the 1–400 GHz frequency range for spherical and elliptical drops and any
polarization. Alta Frequenza. 1981, 50, 262–273.
24. Papoulis, A. Probability & Statistics; Prentice Hall: Hoboken, NJ, USA, 1990.
25. Lindgren, B.W. Statistical Theory, 2nd ed.; MacMillan Company: New York, NY, USA, 1968.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.