31-10-2016, 11:17 AM
1462351366-UnderseaLaserCommunicationwith.pdf (Size: 236.55 KB / Downloads: 7)
ABSTRACT
Laser sources enable highly efficient optical
communications links due to their ability to be
focused into very directive beam profiles.
Recent atmospheric and space optical links
have demonstrated robust laser communications
links at high rate with techniques that are
applicable to the undersea environment. These
techniques contrast to the broad-angle beams
utilized in most reported demonstrations of
undersea optical communications, which have
employed LED-based transmitters. While the
scattering in natural waters will cause the beam
to broaden, a narrowly directive transmitter can
still significantly increase the optical power
delivered to a remote undersea terminal. Using
Monte Carlo analysis of the undersea scattering
environment, we show the two main advantages
of narrow-beam optical communication:
increased power throughput and decreased
temporal spread. Based on information theoretic
arguments, gigabit-per-second class links can
be achieved at 20 extinction lengths by utilizing
pulse position modulation, single-photon-sensitive
receivers, and modern forward error correction
techniques.
INTRODUCTION
Undersea wireless communications is a significant
challenge due to the highly attenuating
nature of seawater for most electromagnetic frequencies.
While acoustic communications has
been demonstrated over long propagation distances
(e.g., 1–10 km), it is limited to submegabit-per-second
data rates, can suffer severe
multi-path, and introduces orders of magnitude
greater latency than optical signaling. By contrast,
with blue-green optical communications,
gigabit-per-second rates can be achieved, over
distances potentially up to hundreds of meters in
the clearest waters, enabling a host of new applications.
Laser light can be collimated into
extremely narrow beams, with sub-milliradianclass
diffraction-limited divergence angles. Even
in seawater, despite significant scattering, narrow
transmitted beams yield an advantage by maximizing
the power delivered to a remote terminal, provided the two terminals can point to each
other with sufficient accuracy. In this article, we
explore the potential benefits as well as the challenges
for undersea wireless communication with
narrow optical beams.
The undersea systems analysis herein is
inspired by lessons learned in the last few
decades developing high-performance laser
communications (lasercom) for atmospheric
and space applications. The recent Lunar
Laser Communications Demonstration
(LLCD) [1] achieved error-free communications
from a moon-orbiting satellite to the
Earth’s surface at rates up to 622 Mb/s. Both
the space terminal (10 cm aperture transmitting
0.5 W of optical power) and ground terminal
(array of four 40 cm receive apertures)
were modest in terms of aperture and optical
power. Another highly successful atmospheric
lasercom demonstration was the Free-Space
Optical Communication Airborne Link
(FOCAL), achieving error-free 100 GB file
transfers over 25+ km from an aircraft to a
ground terminal at a rate of 2500 Mb/s, with
robust tracking out to 60 km [2]. Again, the
optical power (0.5 W) and aperture sizes (2.5
cm on the airplane, four 1.2 cm apertures on
the ground) were very modest. Neither system
required the complexity of adaptive optics.
LLCD and FOCAL provide lessons in optical
communications applicable to undersea. Both
systems used diffraction-limited beams to maximize
power delivery. Each terminal tracked light
transmitted from the remote terminal through a
cooperative means for accomplishing pointing,
acquisition, and tracking (PAT). The PAT systems
were robust amid platform vibrations and
through the turbulent atmosphere. Due to the
extremely long range, LLCD approached lasercom
from the perspective of fundamental performance
bounds. Two vital ingredients were
careful channel characterization and the information
capacity analysis of the modulator/receiver
pairs. In addition, LLCD demonstrated the
operational utility of single-photon-sensitive
detectors.
In this work, we explore the similarities and
differences of the atmospheric and undersea
channels, the technologies available to undersea, and the applicability of nuanced atmospheric
PAT techniques. The article is organized as follows.
We begin with a description of undersea
channel modeling, follow with a discussion of the
benefits of narrow-beam optical systems, and
conclude with implementation considerations.
UNDERSEA CHANNEL
CHARACTERIZATION
Successful undersea lasercom will require a system
design informed by robust and accurate
characterization of the propagation channel.
This will especially be true with narrow-beam
communications that seeks to push performance
to the frontiers of what is physically realizable.
Fortunately, ocean engineers have extensively
worked to characterize the propagation of light
through various seawater conditions. We rely on
these characterization efforts and interpret them
in the context of narrow-beam optical communication.
Signal attenuation due to absorption and
scattering is by far the dominant loss term in
any undersea optical communication link. While
the scale of attenuation varies dramatically
depending on the water characteristics, all
undersea propagation is characterized by a loss
exponential with propagation distance. The
standard method of describing this loss is in
terms of an absorption coefficient (typically
given as a) and a scattering coefficient (given as
b), both in units of m–1. A beam attenuation
length, or extinction length, of (a + b)–1 m
refers to the propagation distance that results in
a power reduced by a factor of e–1 0.37 due to
absorption and scattering. Alternatively, a scattering
length of b–1 m refers to a reduction of
e–1 due to scattering alone. Thus, small values of
a and b denote clear water, and allow light to
propagate for longer distances. Typically referenced
values are listed in Table 1. Higher scattering
coefficients correspond to waters with
higher concentrations of biomaterial, such as
phytoplankton or chromophoric dissolved organic
matter (CDOM), or in some cases suspended
sediment. We can see a variation in extinction
greater than a factor of 10; we also see that
even in clear ocean conditions the exponential
extinction is significant.
Atmospheric and free-space link losses are
typically dominated by a beam spreading term
proportional to R–2, where R is the propagation
range. For collimated beams of light undersea the extinction loss e–(a+b)R dominates, and the
R–2 loss becomes nearly negligible in comparison.
Note that this is only true for nearly collimated
light; if the light source has a broad initial
divergence angle, as with typical LED-based
sources, the R–2 loss plays a much more important
role in the link budget calculations. In
direct contrast to atmospheric lasercom, diffraction
is almost irrelevant in terms of calculating
link losses for undersea lasercom.
The most straightforward approximation of
scattering effects, referenced above, is to treat
all scattered photons as a link loss term. This
“scattering as loss” approximation is highly
appropriate for atmospheric lasercom links.
However, in seawater, light is strongly forwardscattered,
and some non-negligible fraction of
the scattered light will in fact be collected by the
receiver; a “scattering as loss” interpretation of
the undersea channel is especially pessimistic for
scenarios such as the turbid harbor, where the
scattering is substantially stronger than absorption.
Inclusion of scattered light is vital to a correct
comparison of wide-beam vs. narrow-beam
optical systems.
Scattering also has a significant temporal
effect on the optical waveform. Photons that
scatter one or more times but still reach the
receive terminal have traveled a longer distance
than “ballistic” photons that propagate
without being scattered, resulting in a temporal
spreading of the received waveform. Temporal
spreading is a function of several system
parameters, including the scattering coefficient,
the propagation distance, the transmit beam
size, the receive aperture size, and the receiver
field of view (FOV). Intuition regarding the
receiver FOV is particularly important for our
narrow beam analysis: a narrow FOV receiver
will limit the received photons to scattered
light with very small scattering angles, while a
wide FOV will accept light that scatters significantly
off axis. The former will have propagation
distances very close to the ballistic
photons, thus minimizing the temporal spread.
Wide angle photons will have traveled longer
distances, and their inclusion makes the temporal
spread more severe.
Modeling the channel response due to scattering
is best performed by means of ray-tracing
Monte Carlo simulation to compute the random
paths of an ensemble of photons. Each photon is
subject to a series of independent scattering
events, with the frequency of such events characterized
by the scattering coefficient b, and the
resulting angle randomly drawn from a scattering
phase function with a strong forward scattering
emphasis. Given an initial distribution of
photons constituting the lasercom beam exiting
the transmit aperture, we compute the independent
random walk for each photon and approximate
the distribution of photon density, arrival
angle, and time of arrival at the receive aperture,
using the invertible analytic volume scattering
function (VSF) in [4, Appendix B]. An
example photon density distribution is given in
Fig. 1. Despite the approximation’s dismissal of
coherence effects, such Monte Carlo ray tracing
simulations are widely valued computational
methods for modeling the undersea channel.
BENEFITS OF
NARROW-BEAM LASERCOM
Narrow-beam undersea lasercom has three significant
advantages over wide beam optical communication:
an increase in the light transmitted
across the channel, a reduction of the temporal
spread of the signal, and enhanced spatial and
spectral filtering options to reduce background
light. To illustrate these impacts on communication
performance, we begin with a discussion of
modulation and information capacity in photonstarved
channels. We follow with example scenarios
in clear and turbid waters.
MODULATION AND INFORMATION CAPACITY
FOR THE PHOTON-STARVED CHANNEL
The recent optical communication demonstration
from the moon’s orbit to an Earth ground
station (the aforementioned LLCD) demonstrated
high-rate optical communication in what is
sometimes referred to as the “photon-starved
channel.” Such a classification is given to a system
where the total signal flux relative to the
data rate is very limited. Deep space links (due
to their astronomical distances) and undersea
links (due to sea water’s exponential extinction)
can both exhibit photon-starved channels. While
large apertures and high optical powers can partially
compensate, practical size and power limitations
encourage maximizing photon efficiency.
In the case of LLCD, photon efficiency was
increased by utilizing optical bandwidth (a plentiful
resource) and detectors sensitive to single
photons.
For the high-rate LLCD downlink, multiple
bits of information were communicated for every
received photon. LLCD used a high-bandwidth
(5 GHz slot rate) signaling scheme utilizing 16-
ary pulse position modulation (PPM) and halfrate
forward error correction (FEC). For each
16-slot symbol, exactly one contained an optical
pulse; the temporal location of the pulse-containing
slot indicated which of 16 symbols was
transmitted. The receiver deployed an array of
single-photon detectors with precise time of arrival resolution. By detecting the arrival of
even a single photon per symbol, multiple bits of
information were transmitted.
PPM signaling and single-photon receivers
are directly applicable to the undersea environment.
Information theory allows us to compute
the best achievable efficiency with PPM and single-photon-sensitive
receivers. Modeling the
photon arrivals as Poisson distributed (characteristic
of laser light), we calculate the channel
capacity for PPM signaling in the ideal case of
no background light. (For a detailed derivation
of channel capacity for optical receivers, see [5].)
Figure 2a plots the achievable sensitivity vs. the
bandwidth expansion. A 16-ary PPM system with
1/2-rate FEC has a bandwidth expansion of 8
and can achieve –4.6 dB photons/bit, or 2.9
b/photon. Figure 2b shows the capacity of 16-ary
PPM when Nb photons of background light,
detector dark counts, and temporally spread signal
photons are included. In a low-noise case
such as a deep-water or night scenario, the sensitivity
is close to the noiseless result. Higher
background levels (e.g., from upward-looking,
near-surface, daytime scenarios) impact the
achievable sensitivity. Spatially and spectrally
narrow filters increase the information capacity
by reducing stray light to the detector.
CLEAR OCEAN SCENARIO:
INCREASE PHOTON DELIVERY
Consider a clear ocean scenario with absorption
and scattering coefficients of a = 0.114 m–1 and
b = 0.037 m–1, yielding an extinction length of
6.6 m. A low-size low-power narrow-beam lasercom
system can close the link over 20 extinction
lengths (132 m). Consider transmit and receive
terminals with a 2 cm diameter and a collimated
transmit beam at a wavelength of 515 nm. Even
with such small terminals, 132 m still represents
near-field transmission, so in vacuum an aligned
system would couple all of the light to the receiver.
(We assume that the terminals are properly
pointed; discussion of methods for pointing and
tracking follows in a later section.) The channel
loss due to absorption and scattering is 87 dB,
where we assume all scattered light is lost.