Home 
Forward
6.6 Satellite Diversity
6.6.1 Background
Akturan and
Vogel [1997] and Vogel
[1997] describe a method by which they derive single and joint probability
distributions and diversity gains associated with communications employing
multiple satellites. The method consists of: (1) video recording hemispherical
images of the surrounding environment through a fisheye lens mounted atop
a mobile vehicle or photographing still images of the surrounding environment
through a fisheye lens held head-high, (2) performing image analysis of
sequences of the hemispherical scenes, (3) simulating a constellation of
"potentially visible satellite" locations for the particular region of
the world and different times of the day, (4) extracting "path-state" information
associated with the line-of-sight for each "potentially visible satellite"
(e.g., clear, shadowed, or blocked) for different times of the day for
each scene, (5) injecting the "path-state" information into an appropriate
density distribution model, and (6) computing single and joint cumulative
distributions associated with different satellite-look scenarios. Details
concerning the density function models for the different path states are
described in Chapter 10.
6.6.2 Cumulative
Distributions
Figure 6-9 depicts a series of L-Band distributions
(f 
1.6
GHz) for different diversity scenarios to the satellite for urban Japan,
assuming a simulated "Globalstar"
constellation of 48 satellites [Schindall,
1995]. In deriving the distributions given in Figure
6-9, 236 images were combined with approximately 1000 independent constellation
snapshots encompassing a 24 hour period (for each image). Hence, an equivalence
of 236,000 sets of path states went into the database, where approximately
50% of the time three satellites were potentially visible. The distribution
labeled "Highest Satellite" represents the distribution associated with
the satellite having the greatest elevation angle. This distribution was
derived under the condition that the mobile antenna transmits to or receives
radiation from a different satellite position every time a new satellite
achieves the highest elevation angle, independent of azimuth. The highest
elevation path may not necessarily have a "clear" path state. That is,
depending upon the scene at the time, it may be representative of a "blocked"
path state. The distribution labeled "Best Satellite" is also derived from
multiple satellites where the antenna is pointed to the satellite giving
the smallest fade. In calculating this distribution, a decision for "best
satellite" was made approximately every 20 seconds before "hand-over" was
potentially executed. The distribution labeled "2 Best Satellites" represents
the joint distribution associated with the two satellites giving jointly
the "smallest fades". At any instant of time, different pairs of satellites
may fall under the "2 Best Satellite" category. The distributions labeled
"3 Best Satellites" and "4 Best Satellites" are analogously defined. The
above joint distributions were derived assuming "combining diversity" where
the signals received are "added," as opposed to "hand-off" where the satellite
with the "highest" signal is processed. It is apparent that each of the
above distributions are calculated from many different satellites at variable
elevation and azimuth angles. Using the "Highest Satellite" distribution
as the reference, the fade is considerably reduced by switching to the
other diversity modes. For example, at the 20% probability, a 17 dB fade
for the highest satellite may be compared to 6 dB for the "Best" satellite
scenario, giving rise to an 11 dB diversity gain at that probability. We
note that the higher diversity combinations (e.g., 2, 3, and 4 Best Satellites)
do not significantly contribute to an increased diversity gain at percentages
greater than 20%. Figure 6-9 shows that using
the "3 Best Satellite" diversity mode, 1% probability gives rise to a 20
dB fade margin for an urban environment. This substantially high fade may
preclude voice communications for urban environments at small probabilities
even with satellite diversity.
Figure 6-9: Cumulative fade distributions at L Band (f
1.6 GHz) for the simulated Globalstar constellation with combining diversity
for Tokyo, Japan [Vogel,
1997, Akturan
and Vogel, 1997].
6.6.3 Satellite
Diversity Gain
Figure 6-10 depicts the diversity gains for
the distributions given in Figure 6-9 associated
with the "combining diversity" mode. Also shown are the diversity gain
results associated with the "hand-off" mode as defined above. For any fixed
probability, these two processing modes differ by approximately 1 dB (at
most), but also note that any implementation losses have been neglected.
All diversity gains are shown to have a peak at approximately 20% probability.
Figure 6-10: Path diversity gains at f
1.6 GHz derived from distributions for the simulated Globalstar constellation
with combining diversity and handoff for Tokyo, Japan [Vogel,
1997; Akturan
and Vogel, 1997].
6.6.4
Satellite Diversity Measurements at S-Band Employing TDRSS
Direct orbital satellite mobile measurements were conducted by Vogel
[1997] employing a portable satellite receiver with a prototype personal
antenna system using NASA’s Tracking
and Data Relay Satellite System (TDRSS). During three occasions, NASA
made available two TDRSS satellites for diversity measurements at ~2 GHz.
The receiver recorded simultaneously the signals from the two satellites
located at different azimuth and elevation angles. The mobile earth station
(MES) was hand-carried to several typical environments in Austin, Texas
during the spring-summer period when the deciduous trees were in full bloom.
Figure 6-11 shows single and joint distributions
for three path states which are characterized as follows for the specific
environments considered: (1) "clear" denotes an open field with unobstructed
line-of-sight paths to both satellites, (2) "shadowed" denotes walking
in the vicinity of a grove of tall and thin deciduous trees, and (3) "blocked"
denotes walking on the grounds of an apartment complex with multiple three-story
buildings set at various angles. The azimuths and elevation pairs to the
satellites denoted as T7 and T1 were respectively, 248°, 24° and
146°, 49°. Hence, these satellites were separated in azimuth and
elevation by 102° and 25°, respectively.
Figure 6-11: Single and joint probabilities for "clear", "shadowed" and
"blocked" scenarios derived employing S-Band (f = 2 GHz) TDRSS
measurements [Vogel, 1997].
The joint distributions in Figure 6-11 correspond
to the "hand-off" mode (maximum of two satellite signals). Fading for the
"clear" case at the smaller percentages is a result of ground specular
reflections and shadowing by the head. At the 1% level, the joint distribution
is shown to reduce the fade margin from a maximum of approximately 6 dB
to 2 dB. The "blocked" distributions show smaller fading than the "shadowed"
case for the following reasons. The "blocked" environment distribution
is in part representative of an "on" or "off" switch where the line of
sight is either "not blocked" or "blocked," respectively. In addition,
this distribution also includes diffraction and multipath from the buildings,
and some shadowing from a few trees The distribution is therefore strongly
dependent upon the density and height of the structures. On the other hand,
the shadowed distribution represents an almost continuum of tree fading
measured inside a grove; especially relative to the T7 satellite. The diversity
gains at 1% from the "shadowed" and "blocked" environments are noted to
range from 8 to 13 dB and 8 to 10 dB, respectively. These diversity gains
have similar magnitudes to those calculated employing the optical measurements
described in the previous section.
Home
Forward